Voretigene Neparvovec for Biallelic RPE65- Mediated Retinal Disease: Effectiveness and Value Final Evidence Report February 14, 2018 Prepared for ©Institute for Clinical and Economic Review, 2018 ICER Staff/Consultants University of Washington School of Pharmacy Modeling Group* Reiner Banken, MD, MSc Josh Carlson, PhD, MPH Senior Fellow Pharmaceutical Outcomes Research and Policy Institute for Clinical and Economic Review Program, Department of Pharmacy University of Washington David Rind, MD, MSc Chief Medical Officer Institute for Clinical and Economic Review Marita Zimmerman, MPH, PhD Pharmaceutical Outcomes Research and Policy Geri Cramer, BSN, MBA Program, Department of Pharmacy Research Associate University of Washington Institute for Clinical and Economic Review Patricia G. Synnott, MALD, MS Solomon J. Lubinga, BPharm, MSc, PhD Senior Research Associate Pharmaceutical Outcomes Research and Policy Institute for Clinical and Economic Review Program, Department of Pharmacy University of Washington Rick Chapman, PhD, MS Director of Health Economics Institute for Clinical and Economic Review Sonya Khan, MPH Program Director, Midwest CEPAC Institute for Clinical and Economic Review *The role of the University of Washington School of Pharmacy Modeling Group is limited to the Steven D. Pearson, MD, MSc development of the cost-effectiveness model, and the President resulting ICER reports do not necessarily represent Institute for Clinical and Economic Review the views of UW. DATE OF PUBLICATION: February 14, 2018 We would also like to thank Erin Lawler, Maggie Webb, Molly Morgan, and Beverly Andrade for their contributions to this report. ©Institute for Clinical and Economic Review, 2018 Page ii Final Evidence Report – Voretigene Neparvovec About ICER The Institute for Clinical and Economic Review (ICER) is an independent non-profit research organization that evaluates medical evidence and convenes public deliberative bodies to help stakeholders interpret and apply evidence to improve patient outcomes and control costs. Through all its work, ICER seeks to help create a future in which collaborative efforts to move evidence into action provide the foundation for a more effective, efficient, and just health care system. More information about ICER is available at http://www.icer-review.org. The funding for this report comes from government grants and non-profit foundations, with the largest single funder being the Laura and John Arnold Foundation. No funding for this work comes from health insurers, pharmacy benefit managers, or life science companies. ICER receives approximately 20% of its overall revenue from these health industry organizations to run a separate Policy Summit program, with funding approximately equally split between insurers/PBMs and life science companies. For a complete list of funders and for more information on ICER's support, please visit http://www.icerreview.org/about/support/. About Midwest CEPAC The Midwest Comparative Effectiveness Public Advisory Council (Midwest CEPAC) – a core program of ICER – provides a public venue in which the evidence on the effectiveness and value of health care services can be discussed with the input of all stakeholders. Midwest CEPAC seeks to help patients, clinicians, insurers, and policymakers interpret and use evidence to improve the quality and value of health care. The Midwest CEPAC is an independent committee of medical evidence experts from across the Midwest, with a mix of practicing clinicians, methodologists, and leaders in patient engagement and advocacy. All Council members meet strict conflict of interest guidelines and are convened to discuss the evidence summarized in ICER reports and vote on the comparative clinical effectiveness and value of medical interventions. More information about Midwest CEPAC is available at https://icer-review.org/programs/midwest-cepac/. The findings contained within this report are current as of the date of publication. Readers should be cognizant that new evidence may emerge following the publication of this report that could potentially influence the results. ICER may revisit its analyses in a formal update to this report in the future. ©Institute for Clinical and Economic Review, 2018 Page iii Final Evidence Report – Voretigene Neparvovec In the development of this report, ICER’s researchers consulted with several clinical experts, patients, manufacturers, and other stakeholders. The following clinical experts provided input that helped guide the ICER team as we shaped our scope and report. None of these individuals is responsible for the final contents of this report or should be assumed to support any part of this report, which is solely the work of the ICER team and its affiliated researchers. For a complete list of stakeholders from whom we requested input, please visit: https://icer-review.org/material/voretigene-stakeholder-list/. Expert Reviewers Byron L. Lam, MD Professor of Ophthalmology Bascom Palmer Eye Institute Stephen R. Russell, MD Professor of Ophthalmology and Visual Sciences Carver College of Medicine, University of Iowa ©Institute for Clinical and Economic Review, 2018 Page iv Final Evidence Report – Voretigene Neparvovec Table of Contents Executive Summary ................................................................................................................................ 1 1. Background ........................................................................................................................................ 1 1.1 Introduction ................................................................................................................................. 1 1.2 Scope of the Assessment ............................................................................................................. 7 1.3 Definitions .................................................................................................................................... 9 1.4 Insights Gained from Discussions with Impacted Individuals and Advocacy Groups .................. 9 2. Summary of Coverage Policies and Clinical Guidelines ................................................................... 12 2.1 Coverage Policies ....................................................................................................................... 12 2.2 Clinical Guidelines ...................................................................................................................... 12 3. Clinical Effectiveness ........................................................................................................................ 13 3.1 Overview .................................................................................................................................... 13 3.2 Methods ..................................................................................................................................... 13 3.3 Results ........................................................................................................................................ 16 4. Comparative Value ........................................................................................................................... 33 4.1 Long-Term Cost Effectiveness .................................................................................................... 33 4.2 Value-Based Price Benchmarks .................................................................................................. 56 4.3 Potential Budget Impact ............................................................................................................ 56 4.4 Summary and Comment ............................................................................................................ 59 5. Additional Considerations ................................................................................................................ 61 5.1 Other Benefits and Contextual Considerations ......................................................................... 61 5.2 Identification of Low-Value Services .......................................................................................... 63 6. Summary of the Votes and Considerations for Policy ..................................................................... 64 6.1 About the Midwest CEPAC Process ........................................................................................... 64 6.2 Voting Results ............................................................................................................................ 67 6.3 Roundtable Discussion and Key Policy Implications .................................................................. 69 References ........................................................................................................................................... 74 Appendix A. Search Strategies and Results.......................................................................................... 80 Appendix B. Previous Systematic Reviews and Technology Assessments .......................................... 83 Appendix C. Ongoing Studies ............................................................................................................... 84 ©Institute for Clinical and Economic Review, 2018 Page v Final Evidence Report – Voretigene Neparvovec Appendix D. Clinical Effectiveness Supplemental Information............................................................ 85 Appendix E. Comparative Value Supplemental Information ............................................................... 89 Appendix F. Data Extraction Summary Table ...................................................................................... 90 Appendix G. Public Comments........................................................................................................... 102 Appendix H. Conflict of Interest Disclosures ..................................................................................... 103 ©Institute for Clinical and Economic Review, 2018 Page vi Final Evidence Report – Voretigene Neparvovec List of Acronyms Used in this Report AAV2 Adeno-Associated Viral Serotype 2 AE Adverse Event CS Contrast Sensitivity CβA Chicken Beta Actin EOSRD Early Onset Severe Retinal Dystrophy FST Full-Field Light Sensitivity Threshold ICO International Council of Ophthalmology IRD Inherited Retinal Dystrophy ITT Intent to Treat (includes all subjects enrolled and randomized) LCA Leber Congenital Amaurosis LogMAR LOGarithm of the Minimum Angle of Resolution LUX SI unit of illumination; one lumen per square meter mITT Modified Intent-To-Treat (includes all subjects exposed to investigational agent) MLMT Multi-Luminance Mobility Testing OCT Optical Coherence Tomography PDUFA Prescription Drug User Fee Act PLR Pupillary Light Reflex PP Per Protocol RP Retinitis Pigmentosa RPE Retinal Pigment Epithelium SAE Serious Adverse Event SECORD Severe Early Childhood Onset Retinal Dystrophy TEAE Treatment Emergent Adverse Event VA Visual Acuity VF Visual Field VN Voretigene Neparvovec vg Vector Genomes, units in VN product ©Institute for Clinical and Economic Review, 2018 Page vii Final Evidence Report – Voretigene Neparvovec Executive Summary Background Inherited retinal diseases (IRDs) are an important cause of childhood blindness and affect approximately 1 in 2,300 people worldwide.1,2 A number of IRDs are caused by recessive mutations in the gene RPE65 that codes for the protein RPE65. RPE65 (retinal pigment epithelium-specific 65 kDa protein; retinoid isomerohydrolase) is found in the retinal pigment epithelium (RPE) where it plays a critical role in the regeneration of light-reacting proteins in the retina.3 Mutations that affect both copies of the gene RPE65 (biallelic mutations) cause Leber Congenital Amaurosis, type 2 (LCA2), Early Onset Severe Retinal Dystrophy (EOSRD) and Severe Early Childhood-onset Retinal Dystrophy (SECORD), Retinitis Pigmentosa type 20 (RP20), and other phenotypes.4-7 All of these different disorders are rare, and their exact prevalence is unknown. Distinctions among these disorders may reflect the amount of remaining RPE65 activity, but these may also reflect clinical difficulties in assigning correct phenotypic diagnoses.8 Preliminary estimates from the manufacturer suggest that there are between 1,000 and 3,000 persons in the US with RPE65-mediated IRDs.9 Individuals with these disorders have progressive vision loss, which varies depending on the type of mutation and other factors. Individuals may become severely visually impaired during childhood, adolescence, or early adulthood; however, nearly all become fully blind in adulthood.3,4,7 Until now, there have been no therapies that alter the natural history of RPE65-mediated retinal disease. Clinical diagnosis for biallelic RPE65-mediated retinal disease is difficult and, when compared with genetic testing, has been found to incorrectly distinguish among individuals who have heterogenous clinical presentations and progression of disease.4,7,10 Leber Congenital Amaurosis (LCA), also known as congenital or early-infantile blindness, is one of the most severe IRDs. It accounts for around 5% of all inherited retinopathies and is present in approximately 20% of children attending schools for the blind.10 The diagnosis is based on blindness or severe visual impairment presenting in infancy (frequently before age six months), the oculo-digital sign (poking, rubbing, and/or pressing of the eyes), nystagmus, and changes on the electroretinogram.10,11 However, universally agreed-upon diagnostic criteria are lacking.11 The natural history of LCA varies with the genes and mutations involved. Overall, individuals with RPE65 mutations (LCA2), which account for about 6% of gene mutations causing LCA, tend to have better visual function than typically seen in other persons with LCA, with visual acuity often of 20/50 or better early in life.12,13 These individuals may show temporary mild improvements in visual acuity, but inexorably decline after a time of stability, usually reaching a level of inability to see hand motion (20/20,000) in adulthood.9,10,14,15 Even if visual acuity often remains relatively ©Institute for Clinical and Economic Review, 2018 Page ES1 Final Evidence Report – Voretigene Neparvovec preserved up to adolescence, declines in visual field are observed from infancy.4 Regardless of different levels of decline in visual acuity or peripheral vision among persons with biallelic RPE65- mediated inherited retinal disease, all individuals with this type of mutation are visually impaired at low levels of lighting from very young ages.16 Voretigene Neparvovec Procedure Voretigene neparvovec (VN; LUXTURNA™, Spark Therapeutics) was approved by the FDA on December 19, 2017 for the treatment of vision loss due to confirmed biallelic RPE65 mediated-IRD.9 VN is the first gene therapy entering the market in the US that targets a disease caused by mutations in a specific gene.9 Therapy with VN involves using a viral vector (adeno-associated virus serotype 2 [AAV2]) to transfect cells in the RPE with a functioning copy of RPE65. This does not repair or eliminate the defective gene, but rather introduces (i.e. adds) a normal copy of the gene into the cell. Over the last decade, AAV has been used as a vector of choice in gene therapies, having been used in well over 100 clinical trials.17 Adeno-associated virus vector is believed to be safe for many different types of gene therapy as it is not known to cause any disease, cannot reproduce without a helper virus, is less immunogenic than other viruses, and can be manufactured to only include the genetic information of the gene being transferred for therapy.18-20 The retina-brain barrier limits the distribution of the vector into other organs and creates an immune-privileged space limiting classical immune response, diminishing immune-related safety concerns.18,20 The AAV2 vector must be delivered in close proximity to the RPE, the region of the target cells for gene therapy. In order to access the retina during the procedure, it is necessary to completely remove the vitreous gel that fills the eye, a process called vitrectomy.21 Vitrectomy involves an incision into the eyeball and is a standard procedure used for various interventions on the retina and eye. Cataracts are the most common complication, but infections and tears of the retina may also occur.22 After vitrectomy, the liquid containing the vector is injected into the space between the retina and the RPE, a “subretinal” injection. The subretinal injection is administered into or near the macula, the area of the retina needed for visual acuity, and can lead to macular holes and tears, and to infection. Subretinal injection is not a common procedure, although it is performed for some other conditions, and the briefing document submitted by the manufacturer of VN mentions plans for intensive hands-on training of eye surgeons in the small number of centers that are expected to be authorized to administer VN.9 VN is a bilateral treatment; however, the second eye is treated at least six days after the first eye.23 ©Institute for Clinical and Economic Review, 2018 Page ES2 Final Evidence Report – Voretigene Neparvovec Insights Gained from Discussions with Impacted Individuals and Advocacy Groups Impacted individuals and advocacy organizations emphasized the challenges of growing up with low vision for both affected children and their parents/families. Individuals with RPE65-mediated retinal disease described the significant time and energy they have had to dedicate towards adapting to constantly deteriorating vision. Substantial adjustments are necessary for children to perform at the same level as their peers in school, and their academic and career success may directly depend on the quality of assistive services and resources at their disposal. Additionally, such adaptations are often not sufficient to ‘level the playing field.’ Certain career tracks remain out of reach for the visually-impaired population. One individual publicly stated, “I knew I could adapt to being a blind person but that my passions for math and science may not be realized and that was devastating.”24 Poor access to disability services, as well as society’s orientation around the need for sight, puts many individuals with visual impairment at a disadvantage. According to the 2015 American Community Survey, only 15% of individuals with a visual disability earn a bachelor’s degree or higher and just 28% find full-time/full-year employment.25 Nearly 30% of blind Americans live below the poverty line.25 Individuals with low vision often contend with feelings of social isolation. They may be perceived by others as “less intelligent” and may face bullying.24 The inability of individuals with RPE65-mediated retinal disease to navigate independently in dimly-lit settings limits their ability to participate in social activities. Both impacted individuals and IRD clinical experts highlighted the inability to see in dark settings as among the most limiting features of conditions such as LCA. Several of the participants in the Phase III trial of VN noted that prior to treatment, their condition did not permit them to participate in sports, go to the movies, leave the house without assistance on a cloudy day or after dusk, dine without special lighting accommodations, or even to see the facial features of friends and loved ones. Thus, the potential for a therapy to increase light sensitivity and mobility in dim lighting raises great hope among the inherited retinal disease community. Summary of Coverage Policies and Clinical Guidelines There are no clinical guidelines that discuss treatment options for individuals with IRD.26 ©Institute for Clinical and Economic Review, 2018 Page ES3 Final Evidence Report – Voretigene Neparvovec Comparative Clinical Effectiveness We reviewed 14 references related to four individual studies of voretigene neparvovec. The best quality data were from a phase III trial of 31 participants. Twenty-one subjects were randomized to treatment and ten were randomized to no treatment with the option for treatment after one year (delayed intervention arm). Treatment was bilateral with the second eye being treated 6-18 days after the first eye. A sham procedure was not employed. There was one dropout in each arm of the study. The average age of subjects enrolled in the study was 15.1 years (SD 10.9); however, ages ranged from four to 44 years. All participants had a confirmed biallelic RPE65 genetic mutation as well as visual acuity worse than or equal to 20/60 and/or visual field less than 20 degrees in any meridian. As is typical of biallelic RPE65-mediated retinal disease, the individuals enrolled in this study were heterogeneous across visual acuity, visual field, and multi-luminance mobility test (MLMT) at baseline. It is suggested that young children may have better outcomes from VN. However, data were not adequate to assess potential differences in outcomes by age. The primary efficacy endpoint for the Phase III trial was change in bilateral MLMT performance.21 The MLMT assesses the ability to navigate an obstacle course at varying light levels and was designed to be a functional measure that would best capture the impact of treatment. The MLMT is a 5 ft. by 10 ft. obstacle course with 12 distinct but standardized layouts, each with the same number of arrows, turns, and hazards (designed for a visual acuity of 20/200 on the Snellen chart).21,27 Participants were started at the lowest light levels (lux), moving higher until they passed.27,28 Individuals without vision impairment pass this test at the lowest light level (1 lux) 100% of the time. Descriptions of each lux level and a visual model of the MLMT can be seen in Figure 3.3. The walk test was performed on each eye individually, as well as bilaterally (i.e., binocularly). The primary endpoint was reported as the bilateral change in MLMT score; however, change in scores for the first eye were reported as a secondary endpoint.21 A change of one light level in passing was considered clinically meaningful by the sponsor.9 The results show that participants treated with VN saw a difference of 1.6 (95% CI, 0.72 to 2.41) in their bilateral MLMT change score at one year compared to no treatment (original intervention arm score improvement of 1.8, control arm score improvement of 0.2).9,21 Two-year data were presented at the Association for Research in Vision and Ophthalmology (ARVO) meeting in May 2017.29 Mean bilateral MLMT score change in the original intervention group was 1.9 (SD, 1.1) showing that benefits were sustained after the first year. Results from the cross-over control group (delayed intervention arm) were also presented. At one-year after treatment, the delayed intervention arm showed a mean bilateral MLMT score change of 2.1 (SD, 1.6) (see Figure ©Institute for Clinical and Economic Review, 2018 Page ES4 Final Evidence Report – Voretigene Neparvovec ES1).29 Three-year data presented at the ARVO meeting in November 2017 showed sustained benefits at three years for the intervention group and at two years for the delayed intervention arm.30 Figure ES1. Observed Mean Bilateral MLMT Lux Score in Modified Intent-to-Treat Participants in Phase III Study9 30 The phase III study also evaluated other important clinical measures such as full-field light sensitivity (FST), visual acuity, visual field, and quality of life.9 VN was found to increase sensitivity to white light (FST) in treated individuals compared to controls within thirty days of treatment and remained stable out to three-years.21,29 It was noted that 90% of subjects with improvements on the MLMT also had clinically meaningful improvement in light sensitivity (FST).21 A study to assess the relationship between outcomes on the MLMT and FST found a strong correlation between these two measures (-0.74; p<0.001).27 While visual acuity was not statistically different when averaged over both eyes at any timepoint during the study, visual acuity of the better seeing eye did show improvement. Visual field improvements were seen between those who received treatment and those who did not, with slight declines in original results at three years.21,29 Quality of life measures were collected during the study; however, the results have not been published or presented. ©Institute for Clinical and Economic Review, 2018 Page ES5 Final Evidence Report – Voretigene Neparvovec Harms The risks of VN are most often related to the surgical aspects of the procedure.31 At three-years, the most frequently reported adverse events include increases in intraocular pressure (7 events in 5 subjects), cataract (10 events in 5 subjects), retinal tear (3 events in 3 subjects) and retinal deposits (3 events in 3 subjects).30 Secondary safety studies looked at the immune response to VN treatment and showed no cytotoxic responses to either the vector or the gene. Neutralizing antibodies remained near baseline after injection.32 Vector was found in the tears and blood of some participants, but no systemic immune responses were reported.21,33 Two treatment related serious adverse events following VN treatment occurred in the study population. In the phase I follow-on trial, one participant developed bacterial endophthalmitis after surgery which led to irreversible optic atrophy.34,35 In the phase III trial, one participant developed foveal thinning causing a serious decrease in visual acuity at one year.21 Controversies and Uncertainties There are many limitations of the evidence base including interpretation of the measured outcomes and duration of clinical effect. Interpretation of Measured Outcomes The endpoints used in the VN trials are novel. The primary endpoint, the multi-luminance mobility test (MLMT), was designed to capture a critical aspect of the disease process (i.e. being unable to navigate in low light); however, the test itself has not been correlated to outcomes measured in a real-world setting. As such, there remains uncertainty regarding what a one to two-unit improvement in MLMT score means for individuals as they go about their day-to-day activities (e.g. descending stairs in a darkened hallway or using public transportation after dark). Data on quality of life were reported as part of the FDA panel materials but have not been released in a conference presentation or publication, making it difficult to interpret the value to participants. ICER requested quality of life data to use as part of our review but the manufacturer stated that they were unable to provide the data. Duration of Effect Long-term efficacy remains a question for this treatment. Individuals with an RPE65 mutation have significant retinal degeneration leading to worse functional vision over time.36 The therapeutic effects of gene therapy may not be permanent. Other AAV gene therapies have been studied, and treatment of RPE65-mediated IRD using AAV2 gene therapies other than VN showed limited ©Institute for Clinical and Economic Review, 2018 Page ES6 Final Evidence Report – Voretigene Neparvovec duration of benefit.37 In one study, the improvement in visual sensitivity peaked at one to three years after treatment and then declined with degeneration continuing at much the same rate as in the untreated retina.38 In other studies, the treatment effect also declined a few years after injection.39,40 This deterioration may be related to production of abnormal proteins, and VN does not stop such abnormal production. Additionally, much of the retina remains untreated with the current VN procedure. Whether VN has the potential to reduce or eliminate retinal degeneration is currently unknown. A clinical expert involved in the phase I trial presented a testimonial to the FDA that the effects in navigating the MLMT did not diminish in two participants (single-eye treatment) after seven years. Summary VN was shown to provide a significant improvement in mobility under dim light conditions in the treatment group as compared to the control group in the phase III trial. Harms, although present, were related to the surgical aspects of administration. No systemic immune responses from the vector or gene were seen following treatment. While visual improvement past three years was described by clinical experts, no published data exist. Even if improvements persist in treated cells, it remains unclear whether long-term retinal degeneration is impacted by gene therapy. The clinical studies of VN for the treatment of biallelic RPE65-mediated inherited retinal diseases show promise; however, fewer than 50 individuals have received treatment worldwide, and published follow-up data are less than three years in any participant. As a treatment for an ultra- rare disease, methodological limitations are anticipated. The manufacturer of VN has stated that they will follow all subjects out to 15 years per regulatory requirements. Thus, long-term safety and efficacy data should be forthcoming. Despite some uncertainties about the duration of effect and the real-world value of changes in MLMT score, we have high certainty that VN provides at least a small net health benefit (“B+”) for persons with biallelic RPE65-mediated inherited retinal. Long-Term Cost Effectiveness Long-Term Cost Effectiveness Analysis We developed a model to estimate the cost-effectiveness of VN for vision loss associated with biallelic RPE65-mediated inherited retinal disease compared to the standard of care (SoC). The modeled population reflected the VN clinical trial population, with an assumed mean age of 15 ©Institute for Clinical and Economic Review, 2018 Page ES7 Final Evidence Report – Voretigene Neparvovec years and 43% male.21 We also modeled a population with a mean age of three years. The models used one-year cycles over a lifetime time horizon. RPE65-mediated inherited retinal disease is an ultra-rare condition where indirect and nonmedical costs comprise a substantial proportion of total costs, and these costs themselves are large. As such, we used both a US health care system perspective, which focused only on direct medical care costs, and a modified societal perspective which included direct medical costs as well as direct non- medical costs and indirect costs for education, productivity loss, informal care, and nursing home care. We used a 3% discount rate for costs and health outcomes. We modeled categories of visual impairment and blindness based on visual acuity and visual field. We assumed that individuals were considered visually impaired when they reached visual acuity <0.63 decimals or visual field <1200 degrees (as measured by Goldmann III4e). We assumed that individuals were considered blind when they reached visual acuity <0.015 decimals or visual field <48 degrees (as measured by Goldmann III4e).41,42 We created functions for visual acuity and visual field by age based on the natural history of disease (data was digitized from Reape et al. Figure 3).4 In order to model the effect of VN compared to SoC, we used the change in visual acuity and the change in visual field between VN 29 54 These changes were the same for both age groups. We assumed the VA and VF levels would be maintained for the duration of the treatment effect, which was assumed to be 10 years, with a 10- year waning period. A 10-year treatment effect duration was selected because visual outcomes do not appear to be declining in three-year VN data, and anecdotal evidence suggests sustained response up to seven years, but effects in later years cannot be ensured. We applied utility values based on visual ability. We used published values for utilities from a community-based sample that used the standard gamble preference elicitation method.43 Data are significantly limited in this area, as quality-of-life data specific to RPE65-mediated retinal disease do not exist. Therefore, we used utility values based on studies of other retinal disease populations, which are often older. This may lead to biased estimates of quality-of-life and hence overall health outcomes. Due to data limitations, we were not able to link MLMT with classifications of visual impairment. We included three adverse events associated with VN use, based on adverse events categorized as moderate to severe in clinical trials: eye irritation, eye pruritus, and macular hole. Retinal tears were not included as they were assumed to be repaired during the surgery. The cost of VN treatment was $850,000, plus a cost for the surgery. We included cost for direct medical care (health system perspective), direct non-medical costs (modified societal perspective), and indirect costs (modified societal perspective). These sources were selected due to the relative ©Institute for Clinical and Economic Review, 2018 Page ES8 Final Evidence Report – Voretigene Neparvovec importance of visual acuity severity or age for each cost category. All costs were adjusted to 2017 US dollars. We calculated the incremental cost per QALY of VN treatment versus SoC, as well as the cost per additional blindness-free year for VN treatment versus SoC. ©Institute for Clinical and Economic Review, 2018 Page ES9 Final Evidence Report – Voretigene Neparvovec Cost-Effectiveness Model: Results Base Case Results Table ES1 shows results for populations receiving VN at age 15 or age 3 both from a health system perspective and a modified societal perspective. Table ES1. Base Case Results for Voretigene Neparvovec Compared to SoC Treatment SoC Voretigene Incremental Treatment Age: 15 Total Costs, US Health Care System Perspective $213,399 $1,039,019 $825,621 Total Costs, Modified Societal Perspective $1,899,605 $2,515,320 $615,715 Voretigene Costs $0 $854,876 $854,876 AE Costs $0 $222 $222 Direct Ophthalmic Medical Costs $138,833 $144,793 $5,960 Direct Medical Costs, Depression $6,834 $7,171 $336 Direct Medical Costs, Trauma $67,731 $31,957 -$35,774 Direct Non-Medical Costs, Caregiver $892,528 $791,951 -$100,577 Direct Non-Medical Costs, Transport $288,997 $257,132 -$31,865 Direct Non-Medical Costs, Nursing home $21,783 $21,783 $0 Indirect Costs, Productivity $437,043 $359,579 -$77,464 Indirect Costs, Education $45,856 $45,856 $0 Total QALYs 16.0 17.3 1.3 Blindness-Free Years 11.6 22.2 10.6 ICER, US Health Care System Perspective -- -- $643,813/QALY ICER, Modified Societal Perspective -- -- $480,130/QALY $/Additional Blindness Free Year, US Health Care System Perspective -- -- $77,937/Year $/Additional Blindness Free Year, Modified Societal Perspective -- -- $58,123/Year ©Institute for Clinical and Economic Review, 2018 Page ES10 Final Evidence Report – Voretigene Neparvovec Treatment Age: 3 Total Costs, US Health Care System Perspective $193,249 $962,240 $768,991 Total Costs, Modified Societal Perspective $1,782,630 $2,144,086 $361,456 Voretigene Costs $0 $854,876 $854,876 AE Costs $0 $222 $222 Direct Ophthalmic Medical Costs $135,618 $78,329 -$57,290 Direct Medical Costs, Depression $6,682 $3,814 -$2,868 Direct Medical Costs, Trauma $50,948 $24,999 -$25,950 Direct Non-Medical Costs, Caregiver $834,242 $543,647 -$290,595 Direct Non-Medical Costs, Transport $278,964 $220,336 -$58,628 Direct Non-Medical Costs, Nursing home $15,252 $15,252 $0 Indirect Costs, Productivity $306,021 $247,710 -$58,312 Indirect Costs, Education $154,901 $154,901 $0 Total QALYs 18.0 20.6 2.7 Blindness-Free Years 18.4 26.4 8.1 ICER, US Health Care System Perspective -- -- $287,915/QALY ICER, Modified Societal Perspective -- -- $135,331/QALY $/Additional Blindness Free Year, US Health Care System Perspective -- -- $95,175/Year $/Additional Blindness Free Year, Modified Societal Perspective -- -- $44,736/Year AE: Adverse Event, SoC: Standard of Care, ©Institute for Clinical and Economic Review, 2018 Page ES11 Final Evidence Report – Voretigene Neparvovec Sensitivity & Scenario Analysis Results To demonstrate effects of uncertainty on both costs and health outcomes, we varied input parameters using available measures of parameter uncertainty (i.e. standard errors) or plausible ranges to evaluate changes in cost per additional QALY for all model input parameters. We found that key drivers of the model were the utility function, baseline best eye visual acuity, and cost of VN (Figure 4.4). We modeled a scenario in which the duration of benefit is maintained over the lifetime of the individual. In this scenario, we found higher health gains and lower costs for VN persons relative to the base case. This led to lower ICERs of $385,000/QALY for the older age group and $161,000/QALY for the younger age group from the US health care system perspective, and $228,000/QALY for the older age group and $16,000/QALY for the younger age group from the modified societal perspective. Threshold Analysis Results Prices necessary to reach cost-effectiveness thresholds of $50,000, $100,000, $150,000, $250,000, and $500,000 per QALY are listed in Table ES2, for both age groups and both perspectives. Table ES2. Threshold Analysis Results Unit Price Unit Price Unit Price Unit Price Unit Price Discount to to to to to WAC Per from WAC Achieve Achieve Achieve Achieve Achieve Unit To Reach $50,000 $100,000 $150,000 $250,000 $500,000 Thresholds /QALY /QALY /QALY /QALY /QALY Age 15, US Health Care Perspective $850,000 $88,499 $152,619 $216,738 $344,978 $665,576 22% - 90% Age 3, US Health Care Perspective $850,000 $214,553 $348,098 $481,643 $748,733 -- 0% - 75% Age 15, Modified Societal Perspective $850,000 $298,405 $362,524 $426,644 $554,883 -- 0% - 65% Age 3, Modified Societal Perspective $850,000 $622,089 $755,633 -- -- -- 0% - 27% ©Institute for Clinical and Economic Review, 2018 Page ES12 Final Evidence Report – Voretigene Neparvovec Value-Based Price Benchmarks Our value-based benchmark prices for VN treatment at age 15, from both US health care system and modified societal perspectives, are presented in Table ES3. Results for age 3 are presented in the full report. As noted in the document, “Modifications to the ICER value assessment framework for treatments for ultra-rare diseases” (https://icer-review.org/wp-content/uploads/2017/11/ICER- Adaptations-of-Value-Framework-for-Rare-Diseases.pdf), the value-based benchmark price for a drug is defined as the price range that would achieve incremental cost-effectiveness ratios between $100,000 and $150,000 per QALY gained. However, it should be noted that for ultra-rare diseases such as this, decision-makers in the US and in international settings often give special weighting to other benefits and contextual considerations (see Chapter 5) that lead to coverage and funding decisions at higher prices, and thus higher cost-effectiveness ratios, than may be applied to decisions about other treatments. Table ES3. Value-Based Benchmark Prices for VN for the Treatment of Biallelic RPE65-Mediated Inherited Retinal Disease WAC Price to Achieve Price to Achieve Discount from WAC To $100,000 Per QALY $150,000 Per QALY Reach Thresholds Age 15, US Health Care $850,000 $152,619 $216,738 75% - 82% Perspective Age 15, Modified $850,000 $362,524 $426,644 50% - 57% Societal Perspective Potential Budget Impact We used results from the same model employed for the cost-effectiveness analyses to estimate total potential budget impact. Potential budget impact was defined as the total differential cost of using the new therapy rather than relevant existing therapy for the treated population, calculated as differential health care costs (including drug costs) minus any offsets in these costs from averted health care events. For VN treatment of individuals with biallelic RPE65-mediated inherited retinal disease, the annual potential budgetary impact of treating the entire eligible population across all prices (the $850,000 list price and the three cost-effectiveness threshold prices for $50,000, $100,000, and $150,000 per QALY) did not exceed the $915 million threshold. The greatest potential annual budget impact of treating the described population with VN was at the list price of $850,000, reaching 33% of the $915 million threshold. This was largely due to the relatively small number of individuals assumed to be treated per year (350) and the relatively low health care costs incurred following initial treatment with VN. ©Institute for Clinical and Economic Review, 2018 Page ES13 Evidence Report – Voretigene Neparvovec Summary and Comment This study had several limitations. First, the natural history of RPE65-mediated inherited retinal disease has not been thoroughly studied, therefore our underlying disease models have limited data. Second, we were limited in measures of effectiveness for VN to those measures that were captured in the clinical trials as outcomes, as well as in what measures could be linked to quality of life. Because most of the existing quality-of-life literature for blindness has used visual acuity, we were unable to thoroughly utilize all meaningful outcome measures from the clinical trials. Additionally, costs and quality of life measures have not, to our knowledge, been studied for this specific population; therefore, we assumed similarities between this population and people with other types of blindness or visual impairment. Because these populations are likely very different, and of older age groups, than the RPE65-mediated retinal disease population, this may have led to biased estimates, particularly for quality-of-life. In conclusion, we found that VN improves health outcomes compared to standard of care. The high cost makes this unlikely to be a cost-effective intervention at commonly used cost-effectiveness thresholds, except for the younger population from a modified societal perspective. However, for ultra-rare diseases, decision-makers in the US and in international settings often give special weighting to other benefits and to contextual considerations that lead to coverage and funding decisions at higher prices, and thus higher cost-effectiveness ratios, than applied to decisions about other treatments. We found that VN provided more health benefits when given to a younger population, and was therefore more likely to be cost-effective for younger persons. We also found that inclusion of indirect and non-medical costs decreased the total incremental costs for VN, and therefore decreased cost-effectiveness ratios. ©Institute for Clinical and Economic Review, 2018 Page ES14 Evidence Report – Voretigene Neparvovec Other Benefits or Disadvantages Potential Other Benefits This intervention provides significant direct patient Several individuals who received VN appreciated the health benefits that are not adequately captured by improvements in self-confidence, mobility, and the QALY. independence that they felt following treatment. These benefits may not be adequately captured in the QALY. This intervention offers reduced complexity that will N/A significantly improve patient outcomes. This intervention will reduce important health N/A disparities across racial, ethnic, gender, socio- economic, or regional categories. This intervention will significantly reduce caregiver or For individuals with RPE65-mediated inherited retinal broader family burden. disease, treatment with VN may improve independence and thus reduce caregiver and family burden although no data are available at this time. This intervention offers a novel mechanism of action VN is the first gene therapy entering the market in the US or approach that will allow successful treatment of that targets a disease caused by mutations in a specific many patients who have failed other available gene and will offer the first treatment option for treatments. individuals with RPE65-mediated retinal disease. This intervention will have a significant impact on Trial participants and their families stated that the improving the patient’s ability to return to work or treatment enabled them to continue schooling and enter school and/or their overall productivity. the workforce. This intervention will have a significant positive Schools must provide additional support to students with impact outside the family, including on schools visual impairment. If VN improves a student’s visual and/or communities. capability in a school environment, there is a potential for the treatment to also have a significant impact on the resources in the school system and the experience of teachers and other students. This intervention will have a significant impact on the Although availability of VN will likely lead to screening and entire “infrastructure” of care, including effects on earlier detection, this will primarily affect care by providing screening for affected patients, on the sensitization access to treatment with VN. of clinicians, and on the dissemination of understanding about the condition, that may revolutionize how patients are cared for in many ways that extend beyond the treatment itself. Other important benefits or disadvantages that N/A should have an important role in judgments of the value of this intervention. ©Institute for Clinical and Economic Review, 2018 Page ES15 Evidence Report – Voretigene Neparvovec Potential Other Contextual Considerations This intervention is intended for the care of Loss of visual function has been shown to diminish the individuals with a condition of particularly high quality of life significantly. severity in terms of impact on length of life and/or quality of life. This intervention is intended for the care of Visual loss from RPE65-mediated inherited retinal disease individuals with a condition that represents a starts at a young age and is generally progressive to particularly high lifetime burden of illness. blindness. This intervention is the first to offer any VN is the first treatment for RPE65-mediated inherited improvement for patients with this condition. retinal disease Compared to best supportive treatment, there is Given the lack of long term follow-up, there is uncertainty significant uncertainty about the long-term risk of about the long-term risk of VN. serious side effects of this intervention. Compared to best supportive treatment, there is Given the lack of long term follow-up, there is uncertainty significant uncertainty about the magnitude or about the long-term durability of the benefits of VN. durability of the long-term benefits of this intervention. There are additional contextual considerations that N/A should have an important role in judgments of the value of this intervention. ©Institute for Clinical and Economic Review, 2018 Page ES16 Evidence Report – Voretigene Neparvovec 1. Background 1.1 Introduction Inherited retinal diseases (IRDs) are an important cause of childhood blindness and affect approximately 1 in 2,300 people worldwide.1,2 A number of IRDs are caused by recessive mutations in the gene RPE65 that codes for the protein RPE65. RPE65 (retinal pigment epithelium-specific 65 kDa protein; retinoid isomerohydrolase) is found in the retinal pigment epithelium (RPE) where it plays a critical role in the regeneration of light-reacting proteins in the retina.3 Mutations that affect both copies of the gene RPE65 (biallelic mutations) cause Leber Congenital Amaurosis, type 2 (LCA2), Early Onset Severe Retinal Dystrophy (EOSRD) and Severe Early Childhood-onset Retinal Dystrophy (SECORD), Retinitis Pigmentosa type 20 (RP20), and other IRD phenotypes.4-7 All of these disorders are rare, and their exact prevalence is unknown. Distinctions among these disorders may reflect the amount of remaining RPE65 activity, but these may also reflect clinical difficulties in assigning correct phenotypic diagnoses.8 Preliminary estimates from the manufacturer suggest that there are between 1,000 and 3,000 persons in the US with RPE65- mediated IRDs.9 Leber Congenital Amaurosis (LCA), also known as congenital or early-infantile blindness, is one of the most severe IRDs. It accounts for around 5% of all inherited retinopathies and is present in approximately 20% of children attending schools for the blind.10 The diagnosis is based on blindness or severe visual impairment presenting in infancy (frequently before age six months), the oculo-digital sign (poking, rubbing, and/or pressing of the eyes), nystagmus, and changes on the electroretinogram.10,11 However, universally agreed-upon diagnostic criteria are lacking.11 While some congenital retinal visual impairments are accompanied by other neurological features, LCA is limited to dystrophy of the retina.13 Mutations in many different genes result in LCA (Figure 1.1).7,44 Each of these genes can exhibit a great number of mutations. LCA type 2 (LCA2), the LCA due to biallelic RPE65 mutations, has been associated with about 125 different identified gene mutations to date.9 Considering the number of genes and mutations involved, IRDs are one of the most genetically diverse groups of inherited disorders.31 The natural history of LCA varies with the genes and mutations involved. Overall, individuals with RPE65 mutations (LCA2), which account for about 6% of gene mutations causing LCA, tend to have better visual function than typically seen in other LCA patients with visual acuity often of 20/50 or better early in life.12,13 These individuals may show temporary mild improvements in visual acuity, but inexorably decline after a time of stability, usually reaching a level of inability to see hand motion (20/20,000) in adulthood.9,10,14,15 Even if visual acuity often remains relatively preserved up ©Institute for Clinical and Economic Review, 2018 Page 1 Final Evidence Report – Voretigene Neparvovec to adolescence, declines in visual field are observed from infancy.15 Regardless of different levels of decline in visual acuity or peripheral vision among individuals with biallelic RPE65-mediated inherited retinal disease, all individuals with this type of mutation are visually impaired at low levels of lighting from very young ages and nearly all become fully blind in adulthood.3,4,7 Conversely, while LCA comprises various gene mutations (including RPE65), a number of disorders due to RPE65 mutations have different names. The diagnoses of Early Onset Severe Retinal Dystrophy (EOSRD), Severe Early Childhood-onset Retinal Dystrophy (SECORD), and early-onset RP are increasingly considered milder forms of LCA2.7 In a retrospective chart review of 70 patients with biallelic mutations in RPE65, patients initially received 20 distinct clinical diagnoses, 44% had more than one diagnosis over the course of their care, and of those the average number of diagnoses was three. Clinical diagnoses of both RP and LCA were made in 13% of the patients (Figure 1.2). Figure 1.1. Cumulative Number of Identified LCA Genes44 ©Institute for Clinical and Economic Review, 2018 Page 2 Final Evidence Report – Voretigene Neparvovec Figure 1.2. Initial Clinical Diagnosis in 70 Individuals with Biallelic Mutations in RPE654 Effective treatments to reverse IRDs or slow their progression have generally been unavailable; however, voretigene neparvovec (VN; LUXTURNA™, Spark Therapeutics) was approved by the FDA on December 19, 2017 for the treatment of vision loss due to confirmed biallelic RPE65 mediated- IRD across all phenotypes.9 The disease process starts with the rod photoceptors, necessary for peripheral vision and night vision, and later progresses to the cone photoceptors in the macula that are necessary for visual acuity and color vision and involves distinct processes including a biochemical blockade leading to malfunctioning rods and degenerative processes leading to the death of cells of the retinal pigment epithelium (RPE) (a single layer of cells that form the blood-retina barrier and nourish photoreceptors).9,16 The mechanisms responsible for the degenerative processes are not well understood. The activity of the gene RPE65 is essential for the chain of chemical reactions transforming light into electrical signals called the visual cycle (Figure 1.3). RPE65 leads to the production of the protein RPE65, found in the retinal pigment epithelium (RPE) with a molecular weight of 65 kDa. RPE65 catalyzes a chemical reaction needed to produce 11-cis retinal, which is essential for the functioning of the rod photoceptors. Without 11-cis retinal peripheral and night vision would not be possible. ©Institute for Clinical and Economic Review, 2018 Page 3 Final Evidence Report – Voretigene Neparvovec Depending on the RPE65 mutation, the gene produces proteins which lack varying degrees of function. These proteins are often misfolded.45 Abnormal proteins are believed to contribute to retinal degeneration via direct cytotoxic effects, increasing the absorption of photons and leading to cell death.10,46,47 A lack of functioning RPE65 in the visual cycle also leads to the accumulation of cytotoxic retinal esters that contribute to cell death in the RPE.9 Figure 1.3. Visual Cycle10 Gene Therapy Gene therapies modify the expression of genes to treat disease. These can involve strategies that repair genes or that introduce new genes into cells. The process of deliberately introducing functioning genes into cells is called transfection and is accomplished using a vector, usually a virus.48 The new gene can function either as an integral part of the genome, which means the new properties will still be present when the cells divide, or it can function physically separated from the chromosomal DNA of the transfected cell and is usually not transmitted during the division of cells.49 Delivering the gene to the right place and switching it on, avoiding immune responses that can either render the gene therapy ineffective or harm the patient, and making sure that the new gene doesn’t disrupt the function of other genes, are some of central challenges of gene therapy.48 Therapy with VN involves using a viral vector (adeno-associated virus serotype 2 [AAV2]) to transfect cells in the RPE with a functioning copy of RPE65. This does not repair or eliminate the defective gene, but rather introduces a normal copy of the gene into the cell. Over the last decade, AAV has been used as a vector of choice in gene therapies, having been used in well over 100 clinical trials.17 Adeno-associated virus vector is believed to be safe for many different types of gene therapy as it does not cause any disease, cannot reproduce without a helper virus, is less ©Institute for Clinical and Economic Review, 2018 Page 4 Final Evidence Report – Voretigene Neparvovec immunogenic than other viruses, and can be manufactured to only include the genetic information of the gene being transferred for therapy.18-20 AAV2 has a specific affinity with retinal cells. The retina-brain barrier limits the distribution of the vector into other organs and creates an immune- privileged space limiting classical immune response, diminishing safety concerns about immune responses.18,20 The AAV2 vector must be delivered in close proximity to the retinal pigment epithelium (RPE), the region of the target cells for gene therapy. In order to access the retina during the procedure, it is necessary to completely remove the vitreous gel that fills the eye, a process called vitrectomy.21 Vitrectomy involves an incision into the eyeball and is a standard procedure used for various interventions on the retina and eye. Cataracts are the most common complication, but infections and tears of the retina may also occur.22 After vitrectomy, the liquid containing the vector is injected into the space between the retina and the RPE, a “subretinal” injection. The subretinal injection is administered into or near the macula, the area of the retina needed for visual acuity, and can lead to macular holes and tears and to infection. Subretinal injection is not a common procedure, although it is performed for some other conditions, and the briefing document submitted by the manufacturer of VN mentions plans for intensive hands-on training of eye surgeons in the small number of centers that are expected to be authorized to administer VN.9 The recommended treatment regimen for VN is bilateral subretinal injections of 1.5E11 vector genomes (vg) in 0.3 mL per eye. Each eye is treated on separate days but at a close interval, no fewer than 6 days apart.23 This spacing was designed to monitor for complications and reduce potential immune response from two administrations in a short timeframe.9 The therapeutic effects of gene therapy may not be permanent. Other AAV gene therapies have been studied, and treatment of RPE65-mediated IRD using AAV2 gene therapies other than VN showed limited duration of benefit.37 In one study, the improvement in visual sensitivity peaked at one to three years after treatment and then declined with degeneration continuing at much the same rate as in the untreated retina.38 In other studies, the treatment effect also declined a few years after injection.39,40 The decline in visual function in different gene therapies for biallelic RPE65-mediated retinal disease can be either due to a limited effect of the gene therapy being used or due to degenerative processes that continue despite an improvement in photoreceptor functioning. The gene therapies used to date have differed. The AAV viruses themselves have different structures and there are potentially important differences in the genetic material accompanying the normal RPE65 gene, such as promoters and enhancers.18 Furthermore, the preparation of the solution used for administering the treatment may be different. For example, VN is administered in ©Institute for Clinical and Economic Review, 2018 Page 5 Final Evidence Report – Voretigene Neparvovec a solution containing surfactant to help prevent loss of vector on surfaces of the materials used in delivering the vector.9 Due to these important differences, the efficacy and safety results of the different trials cannot be compared. As such, the systematic review on the comparative clinical effectiveness will be limited to the intervention VN. With VN, the duration of effect is uncertain. It is important to consider the possibility of degenerative processes continuing despite an improvement in photoreceptor functioning. Indeed, the mechanisms responsible for these degenerative processes that are leading to the final stage of the disease, the absence of light perception, are not well understood. The mutated genes continue to produce mutant proteins in parallel to the functioning RPE65 produced by the normal gene delivered by transfection. As these mutant proteins are believed to contribute to retinal degeneration by different cytotoxic effects,10,46,47 it is possible that degeneration continues, in spite of continuing therapeutic effect of the gene therapy. Until there is long-term follow-up of individuals treated with VN, the duration of benefit will remain uncertain. Concurrent Innovations • A series of phase I/II trials have examined the efficacy of the synthetic prodrug QLT091001, for replacing enzymatic activity in patients deficient in 11-cis-retinal both in patients with RPE65 mutations and in those with LRAT mutations.50 These trials have had promising results;51-53 a phase III trial is due to commence in the near future.7 • As noted, retinal degeneration was seen to progress in some gene therapy trials. Neuroprotective therapies are currently being studied, and it has been proposed to eventually use them synergistically with gene therapy.54 • In animal models of gene therapy, new AAV vectors have been effective with intravitreal injections, potentially obviating the need for subretinal injections.54 Intravitreal injections are easier and safer than subretinal injections. • The Argus™ II retinal prosthesis system55 received an FDA Humanitarian Device Exemption in 2013 for “use in patients with severe to profound retinitis pigmentosa who meet the following criteria: 1) adults, age 25 years or older; 2) bare light or no light perception in both eyes.”56 We did not find any reports on its use specifically in LCA or in biallelic RPE65- mediated retinal disease. A trial of stem cell treatments for various retinal degenerative diseases was started in 2016 and is expected to run until 2021.57 ©Institute for Clinical and Economic Review, 2018 Page 6 Final Evidence Report – Voretigene Neparvovec 1.2 Scope of the Assessment The scope for this assessment is described on the following pages using the PICOTS (Population, Intervention, Comparators, Outcomes, Timing, and Settings) framework. Evidence was collected from available randomized controlled trials. Observational studies and case series were considered for inclusion as well, given the limited evidence base for VN. Our evidence review included input from individuals and advocacy organizations, data from regulatory documents, information submitted by manufacturers, and other grey literature when the evidence met ICER standards (for more information, see https://icer- review.org/methodology/icers-methods/icer-value-assessment-framework/grey-literature-policy/). Analytic Framework The analytic framework for this assessment is depicted in Figure 1.4. Figure 1.4. Analytic Framework The diagram begins with the population of interest on the left. Actions, such as treatment, are depicted with solid arrows which link the population to outcomes. For example, a treatment may be associated with specific health outcomes. Outcomes are listed in the shaded boxes; those within the rounded boxes are intermediate outcomes (e.g., structural/functional tests of the retina), and those within the squared-off boxes are key measures of benefit (e.g., quality of life). The key ©Institute for Clinical and Economic Review, 2018 Page 7 Final Evidence Report – Voretigene Neparvovec measures of benefit are linked to intermediate outcomes via a dashed line, as the relationship between these two types of outcomes may not always be validated. Curved arrows lead to the adverse events of treatment which are listed within the blue ellipsis.58 Populations The population of focus for this review was all persons with vision loss due to biallelic RPE65- mediated retinal disease. Interventions The intervention of interest was subretinal injections of VN. Comparators The comparator was best supportive treatment. This included correction of refractive error, low- vision aids, and optimal access to educational and work-related opportunities. Outcomes Outcome measures included tests of retinal function, such as visual acuity, full-field sensitivity testing (FST), and tests of functional vision, such as the multi-luminance mobility test (MLMT) used in the phase III trial of VN.59 All outcomes are described in the clinical effectiveness section. Discussions with advocacy groups highlighted certain outcomes that we assessed as the evidence allowed. These included improvements in visual acuity, improvements in night vision, and a halting or slowing of disease progression. The ability to navigate obstacles in lower light settings, for example, can translate into increased mobility and independence, which can have a significant impact on a visually-impaired individual’s quality of life and productivity. We also looked for evidence on patient-reported quality of life. Timing Evidence on intervention effectiveness and harms were derived from studies of any duration. Settings All relevant settings were considered, including inpatient, clinic, and office settings. ©Institute for Clinical and Economic Review, 2018 Page 8 Final Evidence Report – Voretigene Neparvovec Value Framework Considerations ICER has modified its value assessment framework for treatments of certain ultra-rare conditions (http://icer-review.org/material/final-ultra-rare-adaptations/). Biallelic RPE65-mediated retinal disease meets ICER’s criteria for an “ultra-rare” condition. Although modifications to the framework were made after the initiation of this report, the present document takes into account the adaptations and accords VN all the appropriate advantages that would have been expected had the framework been in place prior to the beginning of the assessment process. 1.3 Definitions Biallelic: Affecting both copies of a specific gene (on the paternal and maternal chromosomes)60 Cone cells: Photoreceptor cells in the central retina needed for color vision and visual acuity61 Fovea: Small central area in the macula with the highest concentration of cones providing sharp vision61 Intravitreal: Inside the vitreous humour, a transparent gel-like substance that fills the eye61 Macula: Area of the retina where the cones are located, used in seeing fine detail61 Rod cells: Photoreceptor cells in the outer regions of the retina used for peripheral and night vision61 Vector: Vehicle, often a virus, to carry the new DNA into the cells of a patient with a genetic disease48 1.4 Insights Gained from Discussions with Impacted Individuals and Advocacy Groups Educating a child with low vision takes significant amounts of time, energy, and money beyond that already required for any child. A treatment that delays the onset of visual impairment needs to be considered not simply in the number of years that visual loss is delayed, but also in terms of life stages. For example, being able to navigate on one’s own in school can be very important, even if visual acuity is insufficient to learn without adaptive devices. Completing education and entering the work force is another important life stage where delaying visual loss can potentially have important affects on easing transitions that are required when first becoming employed. ©Institute for Clinical and Economic Review, 2018 Page 9 Final Evidence Report – Voretigene Neparvovec Progressive vision loss can both create uncertainty for impacted individuals and require repeated re-adaptation of skills. We heard from impacted individuals that a therapy that stops decline in vision would be very important even if it did not improve vision. Such a therapy would provide greater certainty in decision making for the future base on stable skills and abilities. Individuals and advocacy organizations emphasized the challenges of growing up with low vision for both affected children and their parents/families. Individuals with RPE65-mediated retinal disease described the significant time and energy they have had to dedicate towards adapting to constantly deteriorating vision. Substantial adjustments are necessary for children to be able to perform at the same level as their peers in school, and their academic and career success may directly depend on the quality of assistive services and resources at their disposal. One mother who testified at the FDA’s Cellular, Tissue, and Gene Therapies Advisory Committee meeting on VN, left her career to stay home and help “level the educational playing field” for her two visually-impaired sons.24 This mother, along with other caregivers who participated in the meeting, described the significant investments their families made in early intervention teachers, as well as the hundreds of hours spent learning braille, practicing how to navigate with a white cane, memorizing emergency escape routes, learning to cross the road safely, mapping out dark hallways at school, and completing homework assignments with assistance from special teachers and family members.24 While such adaptations require considerable investments of time and resources, they are not sufficient in and of themselves to ‘level the playing field.’ Certain career tracks may remain out of reach for the visually-impaired population. One individual publicly stated, “I knew I could adapt to being a blind person but that my passions for math and science may not be realized and that was devastating.”24 Insufficient access to disability services, as well as society’s orientation around the need for sight, puts many individuals with visual impairment at a disadvantage. According to the 2015 American Community Survey, only 15% of individuals with a visual disability earn a bachelor’s degree or higher and just 28% find full-time/full-year employment. Nearly 30% of blind Americans live below the poverty line.25 Moreover, affected individuals often contend with feelings of social isolation. They may be perceived by others as “less intelligent” and may face bullying.24 The inability of individuals with RPE65-mediated retinal disease to navigate independently in dimly-lit settings limits their ability to participate in social activities. Both impacted individuals and IRD clinical experts highlighted the inability to see in dark settings as among the most limiting features of conditions such as LCA. Several of the participants in the phase III trial of VN noted that their condition did not permit them to participate in sports, go to the movies, leave the house without assistance on a cloudy day or after dusk, dine without special lighting accommodations, or even to see the facial features of friends and loved ones. One participant described her condition as follows: “it was like sunglasses over your eyes while looking through this little tunnel.”24 The potential for a therapy to increase light sensitivity and mobility in dim lighting, therefore raises great hope among the inherited retinal disease community. With VN’s FDA approval, impacted individuals and families are hoping the ©Institute for Clinical and Economic Review, 2018 Page 10 Final Evidence Report – Voretigene Neparvovec treatment will become rapidly available for all individuals with biallelic RPE65-mediated retinal disease. ©Institute for Clinical and Economic Review, 2018 Page 11 Final Evidence Report – Voretigene Neparvovec 2. Summary of Coverage Policies and Clinical Guidelines 2.1 Coverage Policies Besides supportive treatment, no other treatment is available to stop the progression or to improve vision for LCA and similar IRDs. Instead, individuals with IRD must turn to supportive services and rehabilitation services to address their needs as their vision wanes. Because these services are, many times, provided outside of the health care system, traditional coverage policies do not address the needs of visually impaired individuals. For example, while Medicare does cover some rehabilitative services, other supportive services, like white canes or guide dogs, are not covered by private or public payers. In 2002, the Centers for Medicare and Medicaid Services (CMS) issued a program memorandum indicating that individuals who are blind or visually impaired are eligible to receive rehabilitation services from covered providers as prescribed by a physician.62 This includes therapeutic services relating to mobility, daily living activities, and other medically necessary rehabilitation goals. In order to address these gaps in services, state-level agencies, national organizations, and local organizations have created different programming designed to support individuals with low vision or blindness, so as to provide services and supports not provided by the healthcare system. For example, the state of Missouri, through their Department of Social Services, provides rehabilitation services, such as job training, mobility, independent living training, children’s services, and screening and treatment programs for the blind.62 Other organizations, including American Council of the Blind, American Foundation for the Blind, Guide Dogs for the Blind, National Federation of the Blind, Foundation Fighting Blindness, and many others, provide services, resources, and funding to allow individuals to seek out the support they need. 2.2 Clinical Guidelines The American Academy of Ophthalmology (AAO) issued eye care guidelines for patients with inherited retinal disease. These guidelines focus on the diagnosis and screening of IRDs. Genetic testing and screening were particularly emphasized. These guidelines do not discuss treatment options as, until now, there have been few or no treatments available to individuals with IRDs.26 We found no guidelines that discussed gene therapy as a treatment for IRDs. ©Institute for Clinical and Economic Review, 2018 Page 12 Final Evidence Report – Voretigene Neparvovec 3. Clinical Effectiveness 3.1 Overview To inform our analysis of the comparative clinical effectiveness of VN for confirmed biallelic RPE65- mediated inherited retinal diseases, we abstracted evidence from available clinical studies, whether in published or unpublished form (e.g., conference abstracts or presentations, FDA review documents). Our population included individuals with RPE65-mediated inherited retinal diseases, particularly Leber Congenital Amaurosis type 2 (LCA2), Retinitis Pigmentosa type 20, Early Onset Severe Retinal Dystrophy (EOSRD), Rod Cone Dystrophy, and Severe Early Childhood Onset Retinal Dystrophy (SECORD). Our primary intervention of interest was VN; studies of related RPE65 gene therapies will be discussed in a contextual fashion. Because there is no current treatment for blindness caused by RPE65-mediated retinal diseases, we did not limit our search to any specific comparator, outcome, timing, or setting to capture the full universe of available data. When reviewing clinical evidence in ultra-rare populations, ICER acknowledges the challenges of study design, recruitment, and availability of data on long-term outcomes. As such, we aim to add specific context to our findings, when possible. 3.2 Methods Data Sources and Searches Procedures for the systematic literature review on VN for RPE65-mediated retinal diseases followed established methods in systematic review research.63 We conducted the review in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.64 The PRISMA guidelines include a checklist of 27 items, further detail of which is available in Appendix Table A1. We searched MEDLINE, Cochrane Central Register of Controlled Trials, and EMBASE for relevant studies. We limited each search to English-language studies of human subjects and excluded articles indexed as guidelines, letters, editorials, narrative reviews, case reports, or news items. To supplement the above searches and ensure optimal and complete literature retrieval, we performed a manual check of the references of recent peer-reviewed publications and public ©Institute for Clinical and Economic Review, 2018 Page 13 Final Evidence Report – Voretigene Neparvovec reports. Further details of the search algorithms, methods for study selection, and data extraction are available in Appendix A and Appendix F. Study Selection We included evidence from all relevant clinical studies, irrespective of whether they used a comparative study design. We did not include studies that used a product other than VN. In recognition of the evolving science for gene therapy, we supplemented our review of published studies with data from conference proceedings, regulatory documents, information submitted by manufacturers, and other grey literature that met ICER standards for review (for more information, see https://icer-review.org/methodology/icers-methods/icer-value-assessment-framework/grey- literature-policy/). We excluded abstracts that reported data also available in peer-reviewed publications. Where data was only available from a press release, we did not include the information in our review. Data Extraction and Quality Assessment Data were extracted by one member of the research team and validated by two others. Because only one study was a randomized controlled trial, the overall quality of the supporting evidence was moderate. Attempts were made to negotiate with the sponsor to gain insight into endpoints that were missing or unclear. Where supplementary evidence was provided, it was incorporated into Section 3, as appropriate. Assessment of Level of Certainty in Evidence We used the ICER Evidence Rating Matrix (see Figure 3.1) to evaluate the evidence for a variety of outcomes. ICER does not change its approach to rating evidence for ultra-rare conditions. The evidence rating reflects a joint judgment of two critical components: a) The magnitude of the difference between a therapeutic agent and its comparator in “net health benefit” – the balance between clinical benefits and risks and/or adverse effects AND b) The level of certainty in the best point estimate of net health benefit.65 ©Institute for Clinical and Economic Review, 2018 Page 14 Final Evidence Report – Voretigene Neparvovec Figure 3.1. ICER Evidence Rating Matrix Assessment of Bias As part of our quality assessment, we evaluated the evidence base for the presence of potential publication bias. Given the emerging nature of the evidence base for newer treatments, we performed an assessment of publication bias using the clinicaltrials.gov database of trials. We scanned the site to identify studies completed more than two years ago that would have met our inclusion criteria and for which no findings have been published. Any such studies may have provided qualitative evidence for use in ascertaining whether there was a biased representation of ©Institute for Clinical and Economic Review, 2018 Page 15 Final Evidence Report – Voretigene Neparvovec study results in the published literature. For this review, we did not find evidence of any completed studies that have not been published. Data Synthesis and Statistical Analyses Data on relevant outcomes were summarized in evidence tables (see Appendix F) and are synthesized qualitatively in the text of the report. 3.3 Results Study Selection Our literature search identified 359 potentially relevant references (see Appendix Figure A1), of which 14 met our inclusion criteria; these citations related to four individual studies. Primary reasons for study exclusion included being the wrong type of study (non-interventional) or the wrong intervention (different vector). Details of all included studies are summarized in Appendix F and in the sections that follow. Key Studies Three of the four key studies had no control arm; however, two studies used the individual’s untreated eye as a control. A safety and proof of concept study enrolled three individuals. The phase I study (Study 101) was a dose escalation trial that treated 12 participants in their worse eye. Study 102, a phase I follow-on to Study 101, treated the same participants in their contralateral eye. The recently published phase III randomized control trial (Study 301, n=31) provides the best quality evidence on the clinical effectiveness of VN. Long-term data from each study were included in our review, whether from publications or from conference proceedings. Data available only through a press release or annual report were not included in our evidence summary. It is important to understand the challenges of completing large, randomized clinical trials in an ultra-rare disease population. ICER strives to consider feasibility constraints when reviewing evidence in these circumstances. ©Institute for Clinical and Economic Review, 2018 Page 16 Final Evidence Report – Voretigene Neparvovec Phase III: Study 301 Study 301, the phase III trial of VN, enrolled 31 subjects randomized in a 2:1 fashion. Twenty-one subjects were randomized to treatment and 10 were randomized to control (no treatment with the option for treatment after one year). There was no sham procedure in the trial. There was one dropout in each arm leaving 20 treated and nine control participants. Statistical analysis was performed on the intent-to-treat population and a modified intent-to-treat population (for a protocol deviation) (see Figure 3.2). Study participants and physicians were not blinded to allocation; however, those scoring the MLMT were blinded. Figure 3.2. Study 301 Protocol Design21,29 The average age of subjects enrolled in the study was 15.1 years (SD 10.9; range 4-44). Participants had a confirmed biallelic RPE65 genetic mutation, visual acuity worse than or equal to 20/60, and/or visual field less than 20 degrees in any meridian. See Table 3.1 for study demographics. Most characteristics were balanced between the two groups, although US versus non-US participants and passing levels on the multi-luminance mobility test (MLMT; described in detail below) were imbalanced. At baseline, 57% of the intervention group passed the test at the lux level of <125 (vs. 40% in the control arm) and 43% of the intervention group passed at ≥ 125 lux (vs. 60% ©Institute for Clinical and Economic Review, 2018 Page 17 Final Evidence Report – Voretigene Neparvovec in the control arm). It is unclear whether this imbalance affected the primary endpoint analysis, but imbalances like this are to be expected in trials with very few participants. Baseline visual acuity, visual field and light sensitivity means between the two groups were not reported in Table 3.1 but were available in the supplementary appendix in raw form. Table 3.1. Study 301 Demographics21,35 Intervention (N=21) Control (N=10) Mean (SD) Age 14.7 (11.8) 15.9 (9.5) Male 9 (43%) 4 (40%) Race - - White 14 (67%) 7 (70%) Asian 3 (14%) 2 (20%) American Indian or Alaska Native 2 (10%) 1 (10%) Black or African American 2 (10%) 0 Ethnicity (Not Hispanic) 16 (76%) 9 (90%) US Resident 17 (81%) 6 (60%) Less than 10 Years Old 9 (43%) 4 (40%) MLMT Passing Level at Baseline - - < 125 lux 12 (57%) 4 (40%) ≥ 125 lux 9 (43%) 15 (60%) Quality of Individual Studies Using criteria published by the US Preventive Services Task Force (USPSTF; see Appendix D), we rated the phase III trial (Study 301) to be of fair quality. An imbalance in the randomized cohort’s ability to pass the MLMT, the inability to fully blind investigators or participants, and changes in endpoints from the original study design led to the fair rating. Studies that lacked a control group or were only available in grey literature sources were not assigned a quality rating. The limitations, uncertainties, and gaps in evidence are discussed in the Controversies and Uncertainties section. Clinical Benefits As with many ultra-rare conditions, the endpoints used in the clinical trials of VN are novel. This was necessary to account for the unique pathophysiology and distinct needs of those with biallelic RPE65-mediated retinal disease, which may differ from other retinal diseases such as macular degeneration. An in-depth explanation of each endpoint is provided prior to data presentation. Primary Endpoint Change in Bilateral Multi-luminance mobility test (MLMT) The primary efficacy endpoint for Study 301 was change in bilateral multi-luminance mobility test (MLMT) performance.21 The MLMT was created by the study sponsor in conjunction with the FDA ©Institute for Clinical and Economic Review, 2018 Page 18 Final Evidence Report – Voretigene Neparvovec to define a quantifiable measure of functional vision that incorporates aspects of visual acuity, visual field, and light sensitivity.27 The MLMT tests the ability to navigate an obstacle course at varying light levels and was designed to be a functional endpoint for biallelic RPE65-mediated retinal disease; however, the outcome could not directly assess real-world functional improvements.28 The MLMT is a 5ft by 10ft obstacle course with 12 unique but standardized layouts, each with the same number of arrows, turns, and hazards (designed for a visual acuity of 20/200 on the Snellen chart).21,27 Participants are tested under seven different lighting conditions or lux levels.28 Individuals without vision impairment pass this test at the lowest light level (1 lux) 100% of the time. Descriptions of each lux level and a visual model of the MLMT can be seen below in Figure 3.3. Figure 3.3. Multi-luminance Mobility Test (MLMT) Example and Light Levels28 ©Institute for Clinical and Economic Review, 2018 Page 19 Final Evidence Report – Voretigene Neparvovec Passing is defined as completing the course in three minutes or less with fewer than four errors (total obstacles=15; <0.25 accuracy).21 Each lux level is mapped to a number ranging from 0 to 6 with the lowest light level (1 lux) having the highest score (6). Passage at a 50-lux level corresponds to a score of 3 (see Table 3.2). Change in MLMT is calculated by taking the difference between the baseline and one-year score. Positive change numbers indicate passage at lower light levels (positive outcome). Table 3.2. Scoring for Multi-Luminance Mobility Test to Calculate Change Score21 Mapping of Passing Lux Level to Score Lux 1 4 10 50 125 250 400 (lowest) (highest) Score 6 5 4 3 2 1 0 Per the Study 301 protocol, participants were dark adapted for 40 minutes, randomly assigned a patched eye and were started at the lowest light levels, moving higher until they passed.27 Participants were randomized to a new configuration of the test each walk to reduce the chance of course memorization.21 The walk was performed on each eye individually, as well as bilaterally (i.e., binocularly). The primary endpoint was reported as the bilateral change in MLMT score; however, change in scores for the first eye were reported as a secondary endpoint.21 Walks were performed at baseline, 30, 90, 180 and 365 days after randomization.21 A change of one light level in passing was considered clinically meaningful by the sponsor.9 All walks were audio and videotaped and scored by two masked, specially trained evaluators at a separate location from the testing site.21 The results show that participants treated with VN saw a difference of 1.6 (95% CI, 0.72 to 2.41) in their bilateral MLMT change score at one year compared to placebo (intervention arm score improvement of 1.8, control arm score improvement of 0.2).9,21 This result indicates that participants treated with VN were able to see in lower light conditions. Zero participants in the intervention group had worsening MLMT scores at one year while three participants in the control group were unable to pass at their baseline lux level one year later.21 Additionally, 65% of the intent-to-treat intervention arm showed maximal improvement in MLMT (passing at 1 lux) as compared to zero participants in the control arm.21 As noted above, more participants in the intervention arm than the control arm were able to pass the test at low light levels (<125 lux) at baseline (57% vs 40%), suggesting that participants in the control arm may have had more advanced disease.21 ©Institute for Clinical and Economic Review, 2018 Page 20 Final Evidence Report – Voretigene Neparvovec Table 3.3. Change in MLMT Score at One Year21 MLMT Score Change from to 1 Year in Intent-To-Treat Population Intervention (n=21) Control (n=10) Difference P-value (from permutation (95% CI) test) Both Eyes Mean (SD) 1.8 (1.1) 0.2 (1.0) 1.6 (0.72 to 2.41) 0.0013 Range 0 to 4 −1 to 2 - - Median (IQR) 2 (1 to 3) 0 (-1 to 1) - - First Eye Mean (SD) 1.9 (1.2) 0.2 (0.6) 1.7 (0.89 to 2.52) 0.0005 Range 0 to 4 -1 to 1 - - Median (IQR) 2 (1 to 3) 0 (0 to 1) - - Second Eye Mean (SD) 2.1 (1.2) 0.1 (0.7) 2.0 (1.14 to 2.85) 0.0001 Range 0 to 5 -1 to 1 - - Median (IQR) 2 (1 to 3) 0 (0 to 1) - - Permutation test p-value represents the proportion of p-values that are smaller than the value observed in the actual dataset using Wilcoxon rank-sum and exact method. IQR: Interquartile Range Two-year data were recently presented at the Association for Research in Vision and Ophthalmology (ARVO) 2017 Meeting. Mean bilateral MLMT score change in the intervention cohort (original intervention) was 1.9 (SD, 1.1) showing that benefits were sustained after the first year.29 Results from the cross-over control group were also presented (delayed intervention arm). All nine subjects in the cross-over control arm went on to receive VN after completing the protocol- required control period. At one-year after treatment, these subjects showed a mean bilateral MLMT score change of 2.1 (SD, 1.6).29 It is unclear why the delayed intervention cohort received greater benefit than the original intervention cohort; however, familiarity with the MLMT course and expectations may have played a role. See Figure 3.4 for lux score and change score data. Recently released three-year data from the American Academy of Ophthalmology Annual Meeting (2017) showed sustained benefits.30 ©Institute for Clinical and Economic Review, 2018 Page 21 Final Evidence Report – Voretigene Neparvovec Figure 3.4. Observed Mean Bilateral MLMT Lux Score in Modified Intent-to-Treat Participants Out to Two Years in Phase III Study9 Secondary Endpoints Three secondary endpoints were evaluated in a hierarchical order: full-field light sensitivity (FST) using white light averaged over both eyes, monocular MLMT score change for the first treated eye, and visual acuity averaged over both eyes.9 Full-Field Light Sensitivity Threshold (FST) Full-field light sensitivity testing was performed using both white and chromatic stimuli and was reported as log10(cd.s/m2). Light sensitivity testing is performed to assess photoreceptor response and a subject’s perception of light sensitivity at different luminance levels.9 White and blue lights target the rod photoreceptors while red light targets cone photoreceptors.34 In Study 301, participant’s eyes were dilated, double patched and dark adapted for 40 minutes. Each eye was then tested (while the contralateral eye remained patched) using a Ganzfeld dome (a 40-cm dome shaped white screen).9 Lights flashed inside the dome alongside a sound (beep or buzz) to notify the participant to indicate whether they see the light flash or not. Light flashes vary in intensity from bright to dim. ©Institute for Clinical and Economic Review, 2018 Page 22 Final Evidence Report – Voretigene Neparvovec The study sponsor indicated that FST is a valuable outcome for the IRD population because it is not affected by nystagmus, allows for testing of those with high levels of visual disability, and does not incorporate sampling bias from tests where specific areas of vision are targeted.9 As a disease predominantly defined by night blindness, full-field light sensitivity threshold testing was thought to be one of the most relevant measures to show benefit from VN therapy.9 For this measure, a negative result indicates improved light sensitivity.9 Clinically meaningful improvement in FST was identified as a 1 log change (10 dB).21 In Study 301, participants in the intent-to-treat intervention group saw an improvement in white light FST of -2.08 log 10(cd.s/m2) between baseline and one-year with no improvement at one-year in the control group (difference between groups: -2.11, 95% CI, -3.19 to -1.04) (Figure 3.5).21 Results were seen immediately after treatment and continued out to one-year. Data qualify issues in FST measurements included missing data from unreliable testing (protocol-related deviations).21 A quantitative description of missing data was not provided. The sponsor acknowledged that missing data was not imputed.21 It is unclear how missing data may have affected the results. Figure 3.5. Observed Mean FST White Light Averaged Over Both Eyes in the MITT Population in the Phase III Study9 The modeled treatment group difference between the intervention arm and control arm in Study 301 at one-year was -2.11 (95% CI, -3.19 to -1.04) log10(cd.s/m2).9 In the original intervention cohort, this benefit remained durable out to three years (-2.27 log10(cd.s/m2 and -2.04 log10(cd.s/m2 at two and three years respectively).29,30 The cross-over control arm experienced a - 2.86 log10(cd.s/m2) difference at one-year and -2.69 log10(cd.s/m2) difference at two years.29,30 An exploratory analysis modeled light sensitivity in red and blue light using a repeated measures model. The change in estimates of full-field light sensitivity with blue light averaged over both eyes ©Institute for Clinical and Economic Review, 2018 Page 23 Final Evidence Report – Voretigene Neparvovec was -1.96 log10(cd.s/m2) (SD 0.37) in the intervention arm versus a change of 0.13 log10(cd.s/m2) (SD 0.49) in the control arm (p-value=0.002).21 The change in estimates for full-field light sensitivity with red light averaged over both eyes was similar (intervention arm mean change -1.29 log10[cd.s/m2), SD 0.17; control arm mean change 0.16 log10[cd.s/m2), SD 0.24; p-value=<0.001).21 It was noted that 90% of subjects with improvements on the MLMT also had clinically meaningful improvement in light sensitivity (FST).21 A study to assess the relationship between found a strong correlation between these two measures (-0.74; p<0.001).27 Change in First Treated Eye Multi-Luminance Mobility Test (MLMT) Change in monocular MLMT scores for the first treated eye also showed a difference in score of 1.7 from baseline to one-year between the intervention and control arm (see Table 3.4).21,35 Table 3.4. Change in First Eye MLMT Score at One Year in Study 30121 First Eye Intervention Control Difference Permutation N=21 N=10 (95% CI) p-value Score Change Mean (SD) 1.9 (1.2) 0.2 (0.6) 1.7 0.001 (0.89 to 2.52) Range (min, max) 0 to 4 -1 to 1 -- -- Quartile 25th 1 0 -- -- Median 2 0 -- -- 75th 3 1 -- -- Visual Acuity Best corrected visual acuity (BCVA), or best vision an individual can achieve with the assistance of corrective lenses, averaged over both eyes, was evaluated as a secondary endpoint in Study 301. Using a scale adapted from Holladay,66 which can calculate a visual acuity score for individuals who are unable to read conventional charts through the use of hand motion and counting fingers, investigators averaged together the BCVA of each individual eye. Over one year of follow-up, the mean treatment group difference (intervention – control) was -0.16 LogMAR (95% CI, -0.41 to 0.08; 0.029 decimals) which corresponded to a gain of 8.1 letters on the eye chart.9 A post-hoc analysis using a different scale from Lange and colleagues67 found comparable results, although differences reached statistical significance (9.0 letters in the intervention group vs. 1.6 letters in the control group; difference of 7.4 letters; 95% CI 0.1 to 14.6; post-hoc p=0.0469).21 Study 301 investigators considered a meaningful change in visual acuity to be a gain of at least 15 letters (≤0.3 LogMAR) on the Early Treatment Diabetic Retinopathy Study eye chart. While no ©Institute for Clinical and Economic Review, 2018 Page 24 Final Evidence Report – Voretigene Neparvovec control participants achieved a meaningful change in visual acuity over the first year of the trial, six of 20 participants in the intervention group gained 15 or more letters in the first eye and four of 20 participants achieved such a change in the second eye.21 We heard from experts in the field that there are flaws with averaging visual acuity across both eyes, especially if one eye has extreme values. As part of our review, we assessed the individually- reported first and second eye visual acuity data published in the Study 301 supplement using a modified impairment method ((best eye*4+worst eye*1)/5 at baseline and one-year) to calculate a “best eye visual acuity”.21,68 We found that individuals in the treatment arm saw an average improvement in visual acuity of -0.17 LogMAR at one-year compared to a smaller improvement of -0.03 LogMAR in the control group. We did not perform statistical analysis. Additionally, the FDA reviewed the data on monocular eye improvements in the first year and found no statistically significant differences between the treatment and control groups.35 Lack of a statistically significant difference was consistent two and three-years post-treatment.30 Exploratory Endpoints Visual Field Visual field is an important outcome for biallelic RPE65-mediated retinal diseases. Unlike some other visual impairments, rods play a primary role in these diseases and degenerative loss of visual field is documented in the natural history.8 Nearly 100% of participants with RPE65-mediated retinal diseases were found to have peripheral retinal abnormality.4 Multiple visual field measures were used in Study 301 as exploratory endpoints. The Goldmann visual field (GVF) perimetry test was used to measure kinetic fields, while the Humphrey computerized test was used to measure static fields in the macula and fovea. Within the GVF, both III4e and V4e stimuli were used; however, the smaller III4e was used whenever possible.a Participants in Study 301 were tested in each eye individually. The Goldmann visual field test requires manual movement of a stimulus from non-seeing to seeing areas in the participant’s visual field. Participants were instructed to press a button (or similar device) when the light became visible. Contour lines, also called isopters, were drawn to outline the visual fields. Scotomas, or areas of decreased light sensitivity, were mapped within these fields. Goldmann visual field (GVF) was reported as sum total degrees.21 The Humphrey visual field (HVF) test utilizes a machine (computer) to assess visual field and has become more common in clinical practice than Goldmann perimetry. In a Humphrey examination, the participant hits a button when a light is seen in the periphery while eyes are focused centrally. a The “III" defines the stimulus size III while the"4e" identifies the intensity of the stimulus used. ©Institute for Clinical and Economic Review, 2018 Page 25 Final Evidence Report – Voretigene Neparvovec In Study 301, the Humphrey analyzer focused on the central areas of the retina, namely the macula and fovea. Humphrey visual field data are reported in decibels (dB). For Humphrey macular VF, a Fastpac strategy with size V test stimulus was used.21 Goldmann visual field measurement showed a statistically significant difference in total sum degrees in the intervention group compared with the control group following treatment with VN (Table 3.5).21 However, differences in the median and mean baseline Goldmann visual field (sum total degrees) between the intervention and control arms were identified (median intervention = 153; median control = 372).21 It is unclear how these differences may have influenced the statistically significant finding in this secondary endpoint. A statistically significant difference in macular threshold visual field was also reported; however, foveal sensitivity showed no differences between the arms (Table 3.5).21 Authors of the phase III study indicated that participants in the intervention group had foveal sensitivities closer to normal levels at baseline compared with those in the control group and therefore hypothesize that lack of significant findings may be due to the limited potential for improvement.21 Table 3.5. Visual Field Outcomes21 Study 301 Visual Field Summary Intervention (n=21) Control (n=10) 1 year Baseline 1 year Change Baseline 1 year Change Difference in P-value Arms (post- (95% CI) hoc) N 20 20 19 10 9 9 - - Goldmann Visual Field III4e (Sum Total Degrees) Mean 332.9 673.9 302.1 427.7 397.8 -76.7 378.7 (145.5 .0059 (SD) (413.3) (423.7) (289.6) (372.3) (367.3) (258.7) to 612.0) Median 153 592 (287 257 372 (109 349 (105 -4 - - (IRQ) (53 to to 1045) (19 to to 686) to 474) (-186 to 469) 520) 31) Humphrey Visual Field, Foveal Sensitivity (dB) Mean 22.4 (6.8) 25.8 (9.1) 2.4 17.6 (8.9) 21.5 (8.9) 2.3 0.04 .18 (SD) (9.7) (5.3) (-7.1 to 7.2) Median 24 30 5 17 26 2 - - (IQR) (19 to (21 to 32) (-1 to 7) (11 to 27) (17 to (-1 to 3) 27) 28) Humphrey Visual Field, Macula Threshold (mean dB) Mean 16.1 (5.5) 24.0 (8.0) 7.7 14.4 (8.0) 15.8 (7.4) -0.2 7.9 0.0005 (SD) (6.2) (1.7) (3.5 to 12.2) Median 15 28 8 16 16 -1 - - (12 to (19 to 29) (4 to 13) (10 to 22) (13 to (-1 to 1) 21) 21) dB: Decibels, IQR: Interquartile Range ©Institute for Clinical and Economic Review, 2018 Page 26 Final Evidence Report – Voretigene Neparvovec Goldmann visual field and Humphrey macular threshold improvements were stable out to two- years in the original intervention group.29 Cross-over controls (delayed intervention) showed a mean change in sum total degrees in Goldmann visual field of 194.3 (244.7) and mean change in Humphrey visual field macula threshold of 5.23 (SD, 9.92) at one year.29 Recently presented data showed a mean (SD) change in GVF III4e sum total degrees averaged over both eyes of 282.2 (256.5) at three years in the original intervention group and 182.6 (309.9) in the delayed intervention group at two years.30 Figure 3.6. Observed Mean Goldmann Visual Field Sum Total Degrees, Both Eyes, in Modified Intent-To-Treat Participants in Phase III Study9 Mean ± standard error. BL, baseline; GVF, Goldmann visual field; III4e, size and intensity of stimuli Quality of Life Study 301 included two quality of life assessments, a modified visual function questionnaire (VFQ- 25), and an in-home orientation and mobility assessment. Data presented to the FDA on the modified VFQ-25 can be seen in Table 3.6. Both subject-score and parent-score averages were significantly higher in the treated arm compared to the control arm at one year.9 The sponsor communicated that the modified VFQ-25 tool is not validated, presenting a major limitation in our assessment of the effect of VN on quality of life. The in-home mobility assessment included independent specialists watching subjects in their personal environment and documenting aspects of functional visual abilities.21 A qualitative description of the findings provided to the FDA stated a correlation was found between improved ©Institute for Clinical and Economic Review, 2018 Page 27 Final Evidence Report – Voretigene Neparvovec MLMT scores and better in-home mobility testing outcomes; however, no data was available for this review.9 Table 3.6. Visual Function Questionnaire Average Scores (ITT)9 Intervention Control Intervention Control Difference N=21 N=10 N=21 N=10 (Intervention-Control) Observed Change from Baseline n Mean (SD) n Mean (SD) n Mean (SD) n Mean (SD) Difference p- value 95% CI Subject Scores Baseline 21 4.4 (1.4) 9 4.9 (1.5) - - - - - - Year 1 20 7.0 (1.9) 9 5.1 (1.8) 20 2.6 (1.8) 9 0.1 (1.4) 2.4 (1.0 to 3.8) 0.001 Parent Scores Baseline 15 3.6 (1.3) 5 3.3 (1.7) - - - - - - Year 1 15 7.5 (1.5) 5 3.1 (1.8) 15 3.9 (1.9) 5 -0.2 (1.3) 4.0 (2.1 to 6.0) 0.002 Harms In total, 41 participants and 81 eyes are part of ongoing safety monitoring.9 At this time, more than ten individuals have been followed for safety issues for the past seven years.9 The risks of VN are most frequently related to the surgical component of the procedure (see Table 3.7).31,35 A full summary of treatment emergent adverse events on the full safety cohort can be found in Table 3.7. Table 3.7. Summary of Treatment Emergent Adverse Events from Phase I and Phase III Studies9 Phase I (N=12) Phase III (N=29) Total Population (N=41) At least 1 TEAE 12 (100%) 29 (100%) 41 (100%) Serious TEAE* 5 (42%) 4 (14%) 9 (22%) TEAE Severity Mild 4 (33%) 10 (34%) 14 (34%) Moderate 6 (50%) 15 (52%) 21 (51%) Severe 2 (17%) 4 (14%) 6 (15%) Vector-related TEAEs 0 3 (10%) 3 (7%) Procedure-related TEAEs 10 (83%) 19 (66%) 29 (71%) Ocular TEAEs 11 (92%) 19 (66%) 30 (73%) *Phase I Serious TEAS: increased intraocular pressure, anal fistula, cryptorchidism, paresthesia, lower limb fracture; Phase III Serious TEAS: retinal disorder, convulsion, adverse drug reaction (2), menorrhagia, pneumonia; TEAEs: Treatment Emergent Adverse Events Secondary safety studies looked at the immune response to VN treatment and showed no cytotoxic responses to either the vector or the gene. Neutralizing antibodies remained near baseline after injection.32 Vector was found in the tears and blood of some participants, but no systemic immune responses were reported.21,33 One subject who underwent contralateral eye treatment in the phase ©Institute for Clinical and Economic Review, 2018 Page 28 Final Evidence Report – Voretigene Neparvovec I follow-on study had consistently increased antibody titers following their first treatment that did not materialize into clinical symptoms.34 There are no data on whether reinjection in the same-eye would cause an immune response, as same-eye retreatment has only been evaluated in animal studies. Two treatment related serious adverse events following VN treatment occurred in the study population; one in the phase I follow-on trial and one in the phase III trial.35 One participant in the phase I follow-on study developed a bacterial (Staphylococcus epidermidis) endophthalmitis after surgery. This individual was treated with intravitreal antibiotics and periocular steroids; however, due to increased intraocular pressure from these treatments, experienced irreversible optic atrophy.34,35 Data from this subject were not used in statistical analyses of efficacy.34 In the phase III study, a participant with an average (both eye) visual acuity of 1.95 LogMAR at baseline experienced decreased central vision and foveal thinning leading to worse visual acuity at one-year (4.0 LogMAR).21 Decreases in retinal thickness were also reported in RPE65 gene therapy trials using a different vector following injection in the foveal area.69 Most of the adverse events reported in the safety population were ocular in nature (63% subjects).35 A summary of ocular-specific adverse events is reported in Table 3.8. Table 3.8. Summary of Ocular Adverse Events35 Ocular AEs Subjects (N=41) Treated Eyes (N=81) Any ocular AE 30 (73%) 51 (63%) Conjunctival Hyperemia 9 (22%) 9 (11%) Increased Intraocular Pressure (IOP) 8 (20%) 10 (12%) Cataract 7 (17%) 11 (14%) Retinal Tear 4 (10%) 4 (5%) Eye Pain 4 (10%) 4 (5%) Corneal Dellen 3 (7%) 3 (4%) Eye Inflammation 3 (7%) 5 (6%) Subretinal Deposits 3 (7%) 3 (4%) Endophthalmitis 1 (2%) 1 (1%) Eye Irritation 3 (7%) 3 (4%) Macular Hole 3 (7%) 3 (4%) Maculopathy 2 (5%) 3 (4%) Foveal Thinning 1 (2%) 2 (2%) Retinal Hemorrhage 1 (2%) 1 (1%) Fovea Dehiscence 1 (2%) 1 (1%) AE: Adverse Event ©Institute for Clinical and Economic Review, 2018 Page 29 Final Evidence Report – Voretigene Neparvovec Controversies and Uncertainties There are limitations of the evidence base leading to many uncertainties. These include interpretation of the measured outcomes, duration of effect, variation of effect with age, and procedure technique. Interpretation of Measured Outcomes The endpoints used in the VN trials are novel. The primary endpoint, the multi-luminance mobility test (MLMT), was designed to capture a critical aspect of the disease process (i.e. being unable to navigate in low light); however, the test itself has not been correlated to outcomes measured in a real-world setting. As such, there remains uncertainty regarding what a one to two-unit improvement in MLMT score means for individuals as they go about their day-to-day activities (e.g. descending stairs in a darkened hallway or using public transportation after dark). Duration of Effect Long-term efficacy remains a question for this treatment. While four-year data are available in a select number of treated individuals, whether the benefits of VN last five years, 10 years, or a lifetime are unknown. A clinical expert involved in the phase I trial presented a testimonial to the FDA that the effects in navigating the MLMT did not diminish in two participants (single-eye treatment) after seven years. Even if treated retinal cells receive unlimited benefit, how that benefit may be offset by worsening vision from ongoing degeneration remains uncertain. Individuals with an RPE65 mutation have significant retinal degeneration leading to worse functional vision over time.36 Whether VN has the potential to reduce or eliminate retinal degeneration is currently unknown; however, at least one researcher has published evidence that, in humans, RPE65 gene therapy does not affect the progressive nature of retinal degeneration.38 These studies used gene therapies other than VN, however. Multiple differences existed between these therapies and VN including the vector, manufacturing process, surgical procedure, and participants enrolled in the trials. This makes comparing outcomes across trials difficult. We are uncertain whether the deterioration seen in other therapies will occur in individuals who received VN. One challenge to assessing ongoing degeneration is the pace of deterioration in functional outcomes. It may take up to a decade to observe worsening in visual outcome measures in this population.31 Variability of Treatment Effect Statements have been made by study investigators regarding improved efficacy in younger individuals with a healthier retinal structure.31,33 Data to support this are scant, although the youngest participants in the phase I study did show substantial improvements in the multi- luminance mobility test (MLMT) while older participants did not show such a pronounced benefit.33 ©Institute for Clinical and Economic Review, 2018 Page 30 Final Evidence Report – Voretigene Neparvovec The phase III study included a greater number of younger individuals in the treatment arm as compared to the control arm.9 Given the few candidates that have received treatment, an adequately-powered subgroup analysis of this question was not feasible. Testimonials provided to ICER and the FDA do point towards younger participants experiencing greater results after treatment. The location of retinal injection plays a role in efficacy and safety of VN treatment. Whether efficacy improves when larger retinal areas are treated (at one time or over time with sequential treatment), has not yet been evaluated. The pathophysiology of rod versus cone contributions to visual function in this population and the effect of RPE65 gene replacement on these types of photoreceptors is still unfolding.69 When cells near the macula are targeted, there is a greater potential for improvements in visual acuity due to the larger number of cone photoreceptors in that region; however, due to the risk of macular holes, the phase III study avoided injection in this area.21 Similarly, visual field improvements were reported to be correlated to the area of retina covered by the injected vector.33 Functional MRI studies confirmed that cortical activation is related to the area of injection. Dose escalation studies determined that a single injection of 1.5x1011 vg in 0.3 ml was optimal for the desired outcome although a clear dose response was not found.33 The sponsor has indicated that increased dosages and volumes have the potential to increase risk without associated benefits.9 Summary Voretigene neparvovec was shown to provide a significant improvement in mobility under dim light conditions in the treatment group as compared to the control group in the phase III trial. Harms, although present, were related to the surgical aspects of administration. No systemic immune responses from the vector or gene were seen following treatment. While visual improvement past three years was described by clinical experts, no published data exist. Even if improvements persist in treated cells, it remains unclear whether long-term retinal degeneration is impacted by gene therapy. Table 3.9. ICER Evidence Ratings ICER Evidence Rating Voretigene neparvovec B+ The clinical studies of VN for the treatment of biallelic RPE65-mediated inherited retinal diseases show promise; however, fewer than 50 individuals have received treatment worldwide, and published follow-up data is less than three years in any participant. As a treatment for an ultra-rare disease, methodological limitations are anticipated. The manufacturer of VN has stated that they ©Institute for Clinical and Economic Review, 2018 Page 31 Final Evidence Report – Voretigene Neparvovec will follow all subjects out to 15 years per regulatory requirements. Thus, long-term safety and efficacy data should be forthcoming. The wide-ranging phenotypes and lack of clear improvement in some individuals lead to difficulty in identifying who will most benefit from treatment. Currently, signs point to highest efficacy in those individuals with healthy retinal cells (prior to degeneration or early degeneration) but no systematic subgroup analysis was reported using OCT evidence of retinal structure. Voretigene neparvovec has a relatively good safety profile, although compared to not receiving treatment, there are harms that are not insignificant, including retinal damage and worsening vision. Similarly, no valid quality of life data have been reported, making it more difficult to understand the value of this therapy to the individual. Participant testimonials provided at the FDA panel were overwhelmingly positive, however. While many uncertainties remain, VN provided a small to substantial improvement over standard care. Thus, we consider the evidence on VN in biallelic RPE65-mediated inherited retinal diseases to be “incremental or better” (B+). ©Institute for Clinical and Economic Review, 2018 Page 32 Final Evidence Report – Voretigene Neparvovec 4. Comparative Value 4.1 Long-Term Cost Effectiveness Overview The primary aim of this analysis was to estimate the cost-effectiveness of VN for vision loss associated with biallelic RPE65-mediated inherited retinal disease compared to the standard of care. The model structure for this assessment is described below. The model was developed in Microsoft Excel. For this section of the report, any data inputs or sources that presented visual acuity in logMAR scale were converted to the decimal scale, using: VAdecimal = 10^(-VAlogMAR). Cost-Effectiveness Model: Methods Model Structure We developed a de novo Markov model with two general health states: “alive with biallelic RPE65- mediated retinal disease” and “dead”. Within the alive individuals, we tracked vision-related clinical measures and quality of life. These measures were tracked over time for individuals who received VN and standard of care. This model structure was selected due to the limited availability of natural history data for biallelic RPE65-mediated inherited retinal disease, which precluded use of a more complex model structure. Among those alive in the model, we tracked age, visual acuity, visual field, categorical visual impairment/blindness, and quality-of-life (Figure 4.1). Age was used to model life-expectancy. Quality-of-life was modeled as a function of visual ability (see utilities methods section below for more detail). Visual acuity (best eye) and visual field (average of both eyes) were used to categorize individuals as not visually impaired, visually impaired, or blind (see clinical inputs methods section below for more detail). Ideally, we would use additional measures, such as the MLMT, to categorize individuals’ visual impairment. However, data for these outcomes from VN trials were not available in metrics that could be linked and validated to these categories. We used a US health care system perspective (i.e., focus on direct medical care costs only). However, RPE65-mediated inherited retinal disease is an ultra-rare condition where indirect and nonmedical costs comprise a substantial proportion of total costs, and these costs themselves are large. Therefore, we also included an analysis using a modified societal perspective which included direct medical costs as well as indirect costs for education and productivity loss, and direct non- medical costs for informal care, transportation, and nursing home care. For impact inventory see Appendix E. We used a 3% discount rate for costs and health outcomes. The model used one-year cycles over a lifetime time horizon. ©Institute for Clinical and Economic Review, 2018 Page 33 Final Evidence Report – Voretigene Neparvovec Figure 4.1 Model Framework Alive with RPE65-Mediated Death Retinal Disease Track: • Age • Visual Acuity (best eye) • Visual Field (average of both eyes) • Visual Impairment/Blindness • Quality of Life Target Population The population for this analysis was individuals in the United States with biallelic RPE65-mediated inherited retinal disease. The modeled population reflected the VN clinical trial population, with a mean age of 15 years, mean baseline VA of 0.096 (best eye), mean baseline VF of 363.8 (average for both eyes) sum total degrees, and 43% male.21 We also modeled a population with a mean age of three years. Treatment effect did not differ by age group, but remaining life expectancy and baseline visual ability were adjusted according to life tables and natural history based visual impairment progression. Treatment Strategies The interventions assessed in this model were: • Voretigene neparvovec (Spark Therapeutics) • Standard of care (SoC) SoC treatment for individuals with vision loss associated with biallelic RPE65-mediated inherited retinal disease does not generally include major vision-related interventions aside from regular physician visits and supportive care. The model estimated the average amount of time individuals live and their quality of life over time with VN or SoC. Utility-adjusted time spent in each health state was summed to provide estimates of expected quality-adjusted life years (QALYs) for each treatment arm. ©Institute for Clinical and Economic Review, 2018 Page 34 Final Evidence Report – Voretigene Neparvovec Model outcomes of interest included: • By intervention: o Total health care costs (discounted) o Life-years (discounted) o QALYs (discounted) • Pairwise comparisons: o Incremental cost-effectiveness ratio (ICER) of VN treatment versus SoC (cost per QALY) o Incremental cost per additional blindness-free year for VN treatment versus SoC Key Model Characteristics and Assumptions We made several assumptions for this model (Table 4.1). Table 4.1. Key Model Assumptions Assumption Rationale Biallelic RPE65-mediated inherited retinal disease There is limited and variable evidence that this and VN treatment do not affect mortality. disease or treatment affect mortality. Treatment effect is maintained for 10 years, followed Trial data for VN are limited to three years, with by a 10-year waning of effect, after which the rate of anecdotal evidence up to seven years and duration of decline in vision is the same as with SoC. treatment effect after that time is not known. As treatment effects do not appear to be changing at the five-year time point, we assumed a 10-year effect, as well as a waning period in which the rate of change is slower than SoC. We additionally modeled a lifetime treatment effect duration as a scenario analysis. Impacted individuals are considered visually As defined in clinical categorizations. impaired when VA<0.63 decimals or VF<1200 degrees, and blind when VA<0.015 decimals or VF<48 degrees based on the average of both eyes. VA: Visual Acuity, VF: Visual Field ©Institute for Clinical and Economic Review, 2018 Page 35 Final Evidence Report – Voretigene Neparvovec Model Inputs Clinical Inputs Clinical Probabilities/Response to Treatment We modeled categories of visual impairment and blindness based on visual acuity and visual field. We assumed that individuals were considered visually impaired when they reached visual acuity <0.63 decimals or visual field <1200 degrees (as measured by Goldmann III4e). We assumed that individuals were considered blind when they reached visual acuity <0.015 decimals or visual field <48 degrees (as measured by Goldmann III4e).41,42 We created a function for visual acuity by age based on the natural history of disease (data was digitized from Reape et al. Figure 3), assuming an exponential functional form (based on visual fit) in the figure then converting to the decimal scale.4 The resulting model form coefficients are shown in Table 4.2. To calculate the function for best eye visual acuity, we used the mean best eye visual acuity from the VN trial, 0.095 decimals, at mean age 15, to recalculate the intercept, and assumed the age coefficient and functional form were the same as for average visual acuity (Table 4.2).29 For visual field, we used equivalent methods to create a function for the average visual field, using digitized natural history data. We assumed a linear functional form based on visual fit, and did not calculate best eye and worst eye separately.4 In order to model the effect of VN compared to SoC, we used the change in visual acuity between VN and SoC for average and best eye (Table 4.2).29 This change was the same for both age groups. We assumed this visual acuity would be maintained for the duration of the treatment effect, which was assumed to be 10 years. A 10-year treatment effect duration was selected because visual outcomes do not appear to be declining in three-year VN data, and anecdotal reports of sustained benefits out to seven years, but effects in later years cannot be ensured. As discussed above, there are theoretical reasons to be concerned that the benefit may wane over time. After the treatment effect duration ends, individuals entered a waning period, assumed to be 10 years. During the waning period, visual outcomes changed at a percentage of the SoC rate of change, based on the percentage of the waning period that has passed. For example, in year one of the waning period, the change in visual outcome would be 1/10 the SoC rate, and in year two would be 2/10 the SoC rate. After the waning period ends, we assumed that visual acuity for those treated with VN changed at the same rate as that of those treated with SoC. ©Institute for Clinical and Economic Review, 2018 Page 36 Final Evidence Report – Voretigene Neparvovec Table 4.2. Visual Acuity (Decimal) Model and Visual Field (Sum Total Degrees from Goldmann III4e) Inputs Clinical Category Value Source Average Eye Visual Acuity Function Functional Form 10^-(function) Assumed Intercept -0.55 Digitized data4 Age Coefficient 0.04360 Digitized data4 Best Eye Visual Acuity Function Functional Form 10^-(function) Assumed Intercept -0.63 Calculated using the function above and Russell, 201721 Age Coefficient 0.04360 Assumed same as average eye Change in Visual Acuity with Voretigene Neparvovec* Best eye, age 15 0.039* Russell, 201721 Best eye, age 3 0.104* Russell, 201721 Average Eye Visual Field Function Functional Form linear Assumed Baseline at Age 15 363.81 Russell, 201721 Age Coefficient -24.27 Digitized data4 Change in Visual Field with Voretigene Neparvovec, 281.56 Russell, 201721 Average (not age specific) Duration of Treatment Effect 10 years Assumed Duration of Waning Period 10 years Assumed *The change in VA with VN does not explicitly differ by age group. The difference seen here is due to conversion from logMAR to decimals. For both age groups, the change is based on a value of -0.15 logMAR change for VN individuals compared to SoC. The baseline VA differs for the age groups for SoC (1.02 logMAR for age 15 and 0.60 logMAR age 3). When converted to decimals this leads to values of 0.905 for SoC and 0.135 for VN for age 15 (difference of 0.039), and 0.251 for SoC and 0.355 for VN for age 3 (difference of 0.104). Mortality We modeled mortality based on gender-specific United States life tables.70 Biallelic RPE65- mediated inherited retinal disease and VN treatment were assumed to have no effect on mortality. Utilities We applied utility values based on visual ability. Data are significantly limited in this area, as quality-of-life data specific to RPE65-mediated retinal disease do not exist. Therefore, we used utility values from other retinal disease populations, which are often older. This may lead to biased estimates of quality-of-life and hence overall health outcomes. ©Institute for Clinical and Economic Review, 2018 Page 37 Final Evidence Report – Voretigene Neparvovec We calculated utility based on visual acuity in the best eye and visual field average of both eyes, and used whichever showed lower utility at a given time. We used published values for utilities over a range of visual acuity from 0.01 (counting fingers-hand motion) to 0.8 (no disutility) decimals.43 Using these utility values, we created a linear function for utility by visual acuity (Table 4.3), which was supported based on visual fit to the data. We then applied the same utility values over a range of visual field measures from 48 to 1440 degrees, and created a linear function (Table 4.3). This source was selected because it is a community-based sample, which is aligned with reference case methods, and used the standard gamble, a gold- standard method. Because this source did not include utility values for those with the lowest level of visual ability (no light perception), the utility associated with no light perception was derived from the linear function. Due to data limitations, we were not able to link MLMT with classifications of visual impairment. We completed a scenario analysis using a different source for utility values and incorporated a piecewise linear function to account for the steep utility drop at the low end of the visual impairment seen in this data source. We used published values for utilities over a range of visual acuity from 0 to 1.6 decimals.71 We created a piecewise function for utility with visual acuity with an inflection point at 0.063 decimals. We then applied the same utility values over range of visual acuity from 0 to 1440 degrees, and created a piecewise function for utility with visual field with an inflection point at 240 degrees (Table 4.3). Table 4.3. Utility Values for Health States Scenario Intercept Slope R2 Source Visual Acuity Base Case 0.5695 0.4865 0.96 Lloyd, 200843 Scenario, 0.063 decimals 0.2745 5.5459 0.99 Brown, 200171 Scenario, >0.063 decimals 0.6266 0.3287 0.96 Brown, 200171 Visual Field Base Case 0.5410 0.0003 0.89 Lloyd, 200843 Scenario, 240 degrees 0.2649 0.0017 1 Brown, 200171 Scenario, 240-719 degrees 0.6375 0.00009375 1 Brown, 200171 ©Institute for Clinical and Economic Review, 2018 Page 38 Final Evidence Report – Voretigene Neparvovec Adverse Events We included three adverse events associated with VN use, based on adverse events categorized as moderate to severe in clinical trials, as shown in Table 4.4. Retinal tears were not included as they were assumed to be repaired during the surgery. Table 4.4. Included Adverse Events Adverse Event Rate21 Cost Disutility Eye Irritation 5% $80 0 Source: CPT Assumed 99214 Eye Pruritus, Ongoing 5% $80 0 Source: CPT Assumed 99214 Macular Hole/ Degeneration 5% $4,447 0.0533 for 6 months, based on difference in Peel group Source: DRG from baseline to 6 months72 124 Economic Inputs The cost of VN treatment was $850,000, plus a cost for the surgery (Table 4.5). We applied costs for direct medical care (physicians/other providers, fundus photography, fluorescein angiography, optical coherence tomography, indocyanine green angiography, laser photocoagulation, intravitreal drug injections, photodynamic therapy), direct medical costs for ophthalmic-related depression, direct medical costs for ophthalmic-related trauma, caregiver costs, and transportation costs based on visual acuity (Table 4.5).73 We applied costs for people categorized as visual impaired or blind for education if they were under age 18, nursing home care if they were over age 65, and productivity loss based on age group (Table 4.5).74 These sources were selected due to the relative importance of visual acuity severity or age for each cost category. For those cost categories using the Brown et al. data, we subtracted costs of the control cohort from each sub-cohort cost. All costs were adjusted to 2017 US dollars. ©Institute for Clinical and Economic Review, 2018 Page 39 Final Evidence Report – Voretigene Neparvovec Table 4.5. Drug Cost Inputs Cost Category Value Source Voretigene Neparvovec $850,000 Manufacturer Surgery $4,876 DRG 117, Intraocular procedures without CC/MCC Direct Cost of Medical Care, Annual Including physicians/other providers, fundus photography, fluorescein angiography, optical coherence tomography, indocyanine green angiography, laser photocoagulation, intravitreal drug injections, photodynamic therapy VA 0 to <0.05 $4,778 Brown, 201673 VA 0.05 to <0.2 $5,204 Brown, 201673 VA 0.2 to <0.4 $1,308 Brown, 201673 VA 0.4 to <0.8 $1,994 Brown, 201673 VA 0.8 $0 Brown, 201673 Direct Medical Costs For Ophthalmic-Related Depression, Annual VA 0 to <0.05 $235 Brown, 201673 VA 0.05 to <0.2 $259 Brown, 201673 VA 0.2 to <0.4 $62 Brown, 201673 VA 0.4 to <0.8 $257 Brown, 201673 VA 0.8 $0 Brown, 201673 Direct Medical Costs For Ophthalmic-Related Trauma, Annual VA 0 to <0.05 $2,870 Brown, 201673 VA 0.05 to <0.2 $315 Brown, 201673 VA 0.2 to <0.4 $393** Brown, 201673 VA 0.4 to <0.8 $1,690** Brown, 201673 VA 0.8 $0** Brown, 201673 Indirect Costs of Education, Annual* Additional Costs of Education for Visually Impaired or Blind Child Compared to Normal Sighted Child Visually Impaired Age 0-17 $11,984 Wittenborn 201374 Blind Age 0-17 $11,984 Wittenborn 201374 ©Institute for Clinical and Economic Review, 2018 Page 40 Final Evidence Report – Voretigene Neparvovec Indirect Costs for Productivity Loss, Annual* Visually Impaired Age 18-39 $9,930 Wittenborn 201374 Age 40-64 $21,074 Wittenborn 201374 Age 65+ $7,316 Wittenborn 201374 Blind Age 18-39 $18,068 Wittenborn 201374 Age 40-64 $27,221 Wittenborn 201374 Age 65+ $7,315 Wittenborn 201374 Direct Non-Medical Costs for Caregivers, Annual* VA 0 to <0.05 $32,652 Brown, 201673 VA 0.05 to <0.2 $25,468 Brown, 201673 VA 0.2 to <0.4 $11,972 Brown, 201673 VA 0.4 to <0.8 $4,860 Brown, 201673 VA 0.8 $0 Brown, 201673 Direct Non-Medical Costs for Transportation, Annual* VA 0 to <0.05 $10,563 Brown, 201673 VA 0.05 to <0.2 $8,287 Brown, 201673 VA 0.2 to <0.4 $6,118 Brown, 201673 VA 0.4 to <0.8 $2,764** Brown, 201673 VA 0.8 $0 Brown, 201673 Direct Non-Medical Costs for Nursing Home Care, Annual* Visually Impaired Age 65+ $3,829 Wittenborn 201374 Blind Age 65+ $7,988 Wittenborn 201374 *Used in modified societal perspective only **These values were corrected from those printed in the Brown et al. 2016 manuscript based on communication with the author, including a change to $0 for trauma costs for the control cohort. Sensitivity Analyses We ran one-way sensitivity analyses to identify the key drivers of model outcomes, using available measures of parameter uncertainty (i.e. standard errors) or plausible ranges for each input described in the model inputs section above. Probabilistic sensitivity analyses were also performed by jointly varying all model parameters over 5,000 simulations, then calculating 95% credible range estimates for each model outcome based on the results. We used normal distributions for age, gender, vision-related health outcomes and costs, and adverse event rates; beta distributions for utilities and disutilities. Additionally, we performed a threshold analysis by systematically altering the price of VN to estimate the maximum prices that would correspond to given willingness-to-pay (WTP) thresholds. ©Institute for Clinical and Economic Review, 2018 Page 41 Final Evidence Report – Voretigene Neparvovec In order to address the potential for outcomes based pricing, we performed a threshold analysis using a rebate for any person whose treatment failed. We assumed a 30% failure rate, based on the percentage of people who did not meet the clinically meaningful change of 1 log unit of change21, and completed a threshold analysis for rebate percent. Scenario Analyses We performed a scenario analysis in which we modeled a lifetime treatment effect duration for VN, and a scenario using and alternative source for the utility values. Model Validation We used several approaches to validate the model. First, we provided preliminary methods and results to manufacturers, advocacy groups, and clinical experts. Based on feedback from these groups, we refined data inputs used in the model. Second, we varied model input parameters to evaluate face validity of changes in results. We performed model verification for model calculations using internal reviewers. Finally, we compared results to other cost-effectiveness models in other vision-related areas. Cost-Effectiveness Model: Results Base Case Results We tracked best eye visual acuity and best eye visual field over time (Figure 4.2). Using these values, we tracked disutility, and calculated overall expected utility over time (Figure 4.3) ©Institute for Clinical and Economic Review, 2018 Page 42 Final Evidence Report – Voretigene Neparvovec Figure 4.2. Best Eye Visual Acuity and Average Visual Field over Time for Patients at Age 15 (top) and Age 3 (bottom) Figure 4.3. Visual Disutility and Overall Utility Over Time for Patients at Age 15 (top) and Age 3 (bottom) ©Institute for Clinical and Economic Review, 2018 Page 43 Final Evidence Report – Voretigene Neparvovec In the population receiving VN at age 15, the average total lifetime cost for individuals treated with VN was approximately $1,039,000 from a US health care system perspective. This included VN costs of $855,000. Patients treated with VN also accumulated a total of approximately $222 in adverse event costs and $184,000 in other direct medical costs (Table 4.6). The average total lifetime cost for individuals treated with SoC was $213,000 of direct medical costs from a US health care perspective (Table 4.6). Direct non-medical costs were approximately $1,071,000 for VN and $1,203,000 for SoC, and indirect costs were approximately $403,000 for VN and $483,000 for SoC. Voretigene neparvovec provided an additional 1.3 QALYs over the remaining lifetime of an individual, leading to an incremental cost-effectiveness ratio of approximately $644,000 per additional QALY gained from the US health care system perspective, and $480,000 per additional QALY gained from the modified societal perspective. Voretigene neparvovec provided an additional 10.6 blindness-free years over the remaining lifetime of an individual, leading to a cost of approximately $78,000 per additional blindness-free year from the US health care system perspective, and $58,000 per additional blindness-free year from the modified societal perspective. In the population receiving VN at age 3, the average total lifetime cost for individuals treated with VN was approximately $962,000 from a US health care system perspective. This included VN costs of $855,000. Patients treated with VN also accumulated a total of approximately $222 in adverse event costs and $107,000 in other direct medical costs (Table 4.6). The average total lifetime cost for individuals treated with SoC was $193,000 of direct medical costs from a US health care perspective (Table 4.6). Direct non-medical costs were approximately $779,000 for VN and $1,128,000 for SoC, and indirect costs were approximately $403,000 for VN and $461,000 for SoC. Voretigene neparvovec provided an additional 2.7 QALYs over the remaining lifetime of an individual, leading to an incremental cost-effectiveness ratio of approximately $288,000 per additional QALY gained from the US health care system perspective, and $135,000 per additional QALY gained from the modified societal perspective. Voretigene neparvovec provided an additional 8.1 blindness-free years over the remaining lifetime of an individual, leading to a cost of approximately $95,000 per additional blindness-free year from the US health care system perspective, and $52,000 per additional blindness-free year from the modified societal perspective. ©Institute for Clinical and Economic Review, 2018 Page 44 Final Evidence Report – Voretigene Neparvovec Table 4.6. Base Case Results for Voretigene Neparvovec Compared to SoC Treatment SoC Voretigene Incremental Treatment Age: 15 Total Costs, US Health Care System $213,399 $1,039,019 $825,621 Perspective Total Costs, Modified Societal $1,899,605 $2,515,320 $615,715 Perspective Voretigene Costs $0 $854,876 $854,876 AE Costs $0 $222 $222 Direct Ophthalmic Medical Costs $138,833 $144,793 $5,960 Direct Medical Costs, Depression $6,834 $7,171 $336 Direct Medical Costs, Trauma $67,731 $31,957 -$35,774 Direct Non-Medical Costs, Caregiver $892,528 $791,951 -$100,577 Direct Non-Medical Costs, Transport $288,997 $257,132 -$31,865 Direct Non-Medical Costs, Nursing home $21,783 $21,783 $0 Indirect Costs, Productivity $437,043 $359,579 -$77,464 Indirect Costs, Education $45,856 $45,856 $0 Total QALYs 16.0 17.3 1.3 Blindness-Free Years 11.6 22.2 10.6 ICER, US Health Care System Perspective -- -- $643,813/QALY ICER, Modified Societal Perspective -- -- $480,130/QALY $/Additional Blindness Free Year, US -- -- $77,937/Year Health Care System Perspective $/Additional Blindness Free Year, -- -- $58,123/Year Modified Societal Perspective ©Institute for Clinical and Economic Review, 2018 Page 45 Final Evidence Report – Voretigene Neparvovec Treatment Age: 3 Total Costs, US Health Care System $193,249 $962,240 $768,991 Perspective Total Costs, Modified Societal $1,782,630 $2,144,086 $361,456 Perspective Voretigene Costs $0 $854,876 $854,876 AE Costs $0 $222 $222 Direct Ophthalmic Medical Costs $135,618 $78,329 -$57,290 Direct Medical Costs, Depression $6,682 $3,814 -$2,868 Direct Medical Costs, Trauma $50,948 $24,999 -$25,950 Direct Non-Medical Costs, Caregiver $834,242 $543,647 -$290,595 Direct Non-Medical Costs, Transport $278,964 $220,336 -$58,628 Direct Non-Medical Costs, Nursing home $15,252 $15,252 $0 Indirect Costs, Productivity $306,021 $247,710 -$58,312 Indirect Costs, Education $154,901 $154,901 $0 Total QALYs 18.0 20.6 2.7 Blindness-Free Years 18.4 26.4 8.1 ICER, US Health Care System Perspective -- -- $287,915/QALY ICER, Modified Societal Perspective -- -- $135,331/QALY $/Additional Blindness Free Year, US -- -- $95,175/Year Health Care System Perspective $/Additional Blindness Free Year, -- -- $44,736/Year Modified Societal Perspective AE: Adverse Event ©Institute for Clinical and Economic Review, 2018 Page 46 Final Evidence Report – Voretigene Neparvovec Sensitivity Analysis Results To demonstrate effects of uncertainty on both costs and health outcomes, we varied input parameters using available measures of parameter uncertainty (i.e. standard errors) or plausible ranges to evaluate changes in cost per additional QALY for all model input parameters. We found that key drivers of the model were the utility function, baseline best eye visual acuity, and cost of VN (Figure 4.4). For the younger population from a modified societal perspective, the annual change in VA for SoC individuals was also a driver. ©Institute for Clinical and Economic Review, 2018 Page 47 Final Evidence Report – Voretigene Neparvovec Figure 4.4. Tornado Diagram(s) for One-Way Sensitivity Analyses of Incremental Cost- Effectiveness Ratio for Voretigene Neparvovec versus Standard of Care for Individuals Who Receive Voretigene Neparvovec at Age 15 (top) and Age 3 (bottom) Age 15, US Health Care System Perspective Age 15, Modified Societal Perspective Age 3, US Health Care System Perspective Age 3, Modified Societal Perspective We ran 5,000 Monte Carlo simulations for a probabilistic sensitivity analysis (Table 4.7). We found that VN had an almost 0% probability of being cost-effective compared to SoC at a threshold of $150,000/QALY for both the age 15 cohort and the age 3 cohort from the health care system perspective. However, the VN had a 74.9% probability of being cost-effective at the $150,000 threshold for the age 3 cohort from the modified societal perspective. ©Institute for Clinical and Economic Review, 2018 Page 48 Final Evidence Report – Voretigene Neparvovec Table 4.7. Probabilistic Sensitivity Analysis Results: Voretigene Neparvovec Versus Standard of Care Cost-Effective at Cost-Effective at Cost-Effective at $50,000 per QALY $100,000 per QALY $150,000 per QALY Age 15, US Health Care Perspective 0.0% 0.0% 0.0% Age 3, US Health Care Perspective 0.0% 0.0% 0.2% Age 15, Modified Societal Perspective 0.1% 0.7% 2.5% Age 3, Modified Societal Perspective 5.1% 31.5% 68.9% Scenario Analysis Results We modeled a scenario in which the duration of benefit is maintained over the lifetime of the impacted individual. In this scenario, we found higher health gains and lower costs for VN patients relative to the base case. This led to lower ICERs of $385,000/QALY for the older age group and $161,000/QALY for the younger age group from the US health care system perspective, and $228,000/QALY for the older age group and $16,000/QALY for the younger age group from the modified societal perspective (Table 4.8). ©Institute for Clinical and Economic Review, 2018 Page 49 Final Evidence Report – Voretigene Neparvovec Table 4.8. Scenario Results for Voretigene Compared to SoC When Duration of Treatment Benefit is Lifetime Treatment SoC Voretigene Incremental Treatment Age: 15 Total Costs, US Health Care System Perspective $213,399 $1,020,093 $806,695 Total Costs, Modified Societal Perspective $1,899,605 $2,377,595 $477,990 Voretigene Costs $0 $854,876 $854,876 AE Costs $0 $222 $222 Direct Ophthalmic Medical Costs $138,833 $148,619 $9,786 Direct Medical Costs, Depression $6,834 $7,387 $552 Direct Medical Costs, Trauma $67,731 $8,988 -$58,742 Direct Non-Medical Costs, Caregiver $892,528 $727,375 -$165,152 Direct Non-Medical Costs, Transport $288,997 $236,673 -$52,324 Direct Non-Medical Costs, Nursing home $21,783 $10,442 -$11,341 Indirect Costs, Productivity $437,043 $337,155 -$99,888 Indirect Costs, Education $45,856 $45,856 $0 Total QALYs 16.0 18.1 2.1 Blindness-Free Years 11.6 28.6 17.0 ICER, US Health Care System Perspective -- -- $384,624/QALY ICER, Modified Societal Perspective -- -- $227,901/QALY $/Additional Blindness Free Year, US Health Care System Perspective -- -- $47,541/Year $/Additional Blindness Free Year, Modified Societal Perspective -- -- $28,170/Year Treatment Age: 3 Total Costs, US Health Care System Perspective $193,249 $908,401 $715,153 Total Costs, Modified Societal Perspective $1,782,630 $1,853,809 $71,180 Voretigene Costs $0 $854,876 $854,876 AE Costs $0 $222 $222 Direct Ophthalmic Medical Costs $135,618 $39,562 -$96,056 Direct Medical Costs, Depression $6,682 $1,869 -$4,813 Direct Medical Costs, Trauma $50,948 $11,872 -$39,076 Direct Non-Medical Costs, Caregiver $834,242 $362,079 -$472,163 Direct Non-Medical Costs, Transport $278,964 $185,037 -$93,926 Direct Non-Medical Costs, Nursing home $15,252 $7,311 -$7,941 Indirect Costs, Productivity $306,021 $236,079 -$69,943 Indirect Costs, Education $154,901 $154,901 $0 Total QALYs 18.0 22.4 4.4 Blindness-Free Years 18.4 30.2 11.9 ICER, US Health Care System Perspective -- -- $161,187/QALY ICER, Modified Societal Perspective -- -- $16,043/QALY $/Additional Blindness Free Year, US Health Care System Perspective -- -- $60,191/Year $/Additional Blindness Free Year, Modified Societal Perspective -- -- $5,991/Year AE: Adverse Events ©Institute for Clinical and Economic Review, 2018 Page 50 Final Evidence Report – Voretigene Neparvovec We also modeled a scenario in which we used alternative utility sources derived from individuals with visual impairment and a piecewise linear utility function. In this scenario, we found higher health gains for VN patients relative to the base case. This led to lower ICERs of $158,000/QALY for the older age group and $161,000/QALY for the younger age group from the US health care system perspective, and $118,000/QALY for the older age group and $75,000/QALY for the younger age group from the modified societal perspective (Table 4.9). ©Institute for Clinical and Economic Review, 2018 Page 51 Final Evidence Report – Voretigene Neparvovec Table 4.9. Scenario Results for Voretigene Compared to SoC Using Non-Linear Utility Function Treatment SoC Voretigene Incremental Treatment Age: 15 Total Costs, US Health Care System Perspective $213,399 $1,039,019 $825,621 Total Costs, Modified Societal Perspective $1,899,605 $2,515,320 $615,715 Voretigene Costs $0 $854,876 $854,876 AE Costs $0 $222 $222 Direct Ophthalmic Medical Costs $138,833 $144,793 $5,960 Direct Medical Costs, Depression $6,834 $7,171 $336 Direct Medical Costs, Trauma $67,731 $31,957 -$35,774 Direct Non-Medical Costs, Caregiver $892,528 $791,951 -$100,577 Direct Non-Medical Costs, Transport $288,997 $257,132 -$31,865 Direct Non-Medical Costs, Nursing home $21,783 $21,783 $0 Indirect Costs, Productivity $437,043 $359,579 -$77,464 Indirect Costs, Education $45,856 $45,856 $0 Total QALYs 10.7 16.0 5.2 Blindness-Free Years 11.6 22.2 10.6 ICER, US Health Care System Perspective -- -- $157,844/QALY ICER, Modified Societal Perspective -- -- $117,713/QALY $/Additional Blindness Free Year, US Health Care System Perspective -- -- $77,937/Year $/Additional Blindness Free Year, Modified Societal Perspective -- -- $58,123/Year Treatment Age: 3 Total Costs, US Health Care System Perspective $193,249 $962,240 $768,991 Total Costs, Modified Societal Perspective $1,782,630 $2,144,086 $361,456 Voretigene Costs $0 $854,876 $854,876 AE Costs $0 $222 $222 Direct Ophthalmic Medical Costs $135,618 $78,329 -$57,290 Direct Medical Costs, Depression $6,682 $3,814 -$2,868 Direct Medical Costs, Trauma $50,948 $24,999 -$25,950 Direct Non-Medical Costs, Caregiver $834,242 $543,647 -$290,595 Direct Non-Medical Costs, Transport $278,964 $220,336 -$58,628 Direct Non-Medical Costs, Nursing home $15,252 $15,252 $0 Indirect Costs, Productivity $306,021 $247,710 -$58,312 Indirect Costs, Education $154,901 $154,901 $0 Total QALYs 14.5 19.3 4.8 Blindness-Free Years 18.4 26.4 8.1 ICER, US Health Care System Perspective -- -- $160,593/QALY ICER, Modified Societal Perspective -- -- $75,485/QALY $/Additional Blindness Free Year, US Health Care System Perspective -- -- $95,175/Year $/Additional Blindness Free Year, Modified Societal Perspective -- -- $44,736/Year AE: Adverse Event ©Institute for Clinical and Economic Review, 2018 Page 52 Final Evidence Report – Voretigene Neparvovec Threshold Analysis Results Prices necessary to reach cost-effectiveness thresholds of $50,000, $100,000, $150,000, $250,000, and $500,000 per QALY are listed in Table 4.10, for both age groups and both perspectives. Threshold prices were higher for the younger age group, and higher from the modified societal perspective. A discount of up to 75%, would be necessary to reach a cost-effectiveness threshold of $150,000/QALY. Smaller discounts would be needed to achieve a cost-effectiveness threshold of $250,000 or $500,000 per QALY. No discount would be required for the younger age group to reach a $500,000/QALY threshold. Table 4.10. Threshold Analysis Results Unit Price Unit Unit Price Unit Price Unit Price Discount to Price to to to to WAC Per from WAC Achieve Achieve Achieve Achieve Achieve Unit To Reach $50,000 $100,000 $150,000 $250,000 $500,000 Thresholds /QALY /QALY /QALY /QALY /QALY Age 15, US Health Care $850,000 $88,499 $152,619 $216,738 $344,978 $665,576 22% - 90% Perspective Age 3, US Health Care $850,000 $214,553 $348,098 $481,643 $748,733 -- 0% - 75% Perspective Age 15, Modified $850,000 $298,405 $362,524 $426,644 $554,883 -- 0% - 65% Societal Perspective Age 3, Modified Societal $850,000 $622,089 $755,633 -- -- -- 0% - 27% Perspective We varied the rebate percentage that would be required for individuals who failed treatment in order for VN to reach relevant willingness-to-pay thresholds. For the older population from the US healthcare system perspective, a 72% rebate would be required to reach a $500,000/QALY threshold. For the older population from a modified societal perspective, a threshold of $250,000/QALY could not be reached because it would require more than a 100% rebate for the 30% of individuals who fail the treatment. For the younger population from a US health care system perspective, a 40% rebate would be required to reach the $250,000/QALY threshold. For the younger population from a modified societal perspective, a 37% rebate would be required to reach the $100,000/QALY threshold, and an 89% rebate would be required to reach the $50,000/QALY threshold. Model Validation Model validation followed standard practices in the field. We tested all mathematical functions in the model to ensure they were consistent with the report (and supplemental Appendix materials). We also conducted sensitivity analyses with null input values to ensure the model was producing ©Institute for Clinical and Economic Review, 2018 Page 53 Final Evidence Report – Voretigene Neparvovec findings consistent with expectations. Further, independent modelers tested the mathematical functions in the model as well as the specific inputs and corresponding outputs. Model validation was also conducted in terms of comparisons to other model findings. We searched the literature to identify models that were similar to our analysis, with comparable populations, settings, perspective, and treatments. We also compared our model to previously published models. We searched the literature to identify models that were similar to our own, with comparable populations, settings, perspective, and treatments. While there are no economic evaluations assessing the cost-effectiveness of VN, we reviewed other relevant models in vision-related diseases, with comparisons focusing primarily on modeling approach, and less on the results of these economic evaluations. The incremental cost-effectiveness results whenever stated are intended to serve only as illustrative of what different treatments offer for treating severe visual impairment stemming from different disorders, and not as a direct comparison to the ICER analysis on VN. Bennison et al. evaluated the cost-effectiveness of ocriplasmin relative to standard-of-care for the treatment of vitreomacular traction (VMT) and macular hole.75 This model comprised two parts, a short-term decision tree to simulate whether patients had a successful anatomic outcome, and a long-term Markov model simulating long-term clinical and cost outcomes. While blindness in the ICER model was defined as visual acuity <0.015 decimals, Bennison et al. defined it as 6/60, or 0.1 decimals. Visual acuity decline over time was also modeled differently, with Bennison et al. treating visual acuity decline in the non-study eye the same as that of those in the general population. In the ICER model, disutilities for vision-related impairment were linked to visual acuity in the best eye and visual field average of both eyes, whereas Bennison et al. awarded disutilities based on adverse events related to underlying cause of vision impairment. The ICERs of ocriplasmin relative to standard-of-care ranged from £18,056 for treating VMT without epiretinal membrane (ERM) or full thickness macular hole (FTMH), to £61,059 for treating VMT with ERM but no FTMH. Dunbar et al. evaluated the cost-utility of screening and laser treatment relative to no screening and treatment for retinopathy of prematurity (ROP), in infants with birthweight <1500g or gestational ages of 28 weeks or less, in a neonatal intensive care unit.76 Treatment effect with laser therapy was assumed to be permanent. Visual acuity was measured at the time of screening and at the 10-year post-therapy time point. Utilities were derived from visual acuity estimates in the better-seeing eye. Utility for non-treated eye was 0.59 and treated eye was 0.69, with treatment effect lasting 77.5 years. This study estimated a cost per QALY ratio of $1,565 for screening and laser therapy. Another economic evaluation by Rothschild et al. comparing screening and laser treatment relative to no screening and treatment for retinopathy of prematurity (ROP) in infants with birthweight <1500g found screening and treatment to be cost-saving from a US societal perspective.77 ©Institute for Clinical and Economic Review, 2018 Page 54 Final Evidence Report – Voretigene Neparvovec Mitchell et al. evaluated the cost-effectiveness of ranibizumab monotherapy or in combination with laser therapy relative to laser monotherapy for the treatment of diabetic macular edema (DME) in a UK population.78 The study used clinical efficacy estimates from the RESTORE trial. Health state utilities for the target patient population were derived from the EQ-5D, and then linked to the best corrected visual acuity (BCVA), with a BCVA score of 0-25 associated with a mean utility of 0.547 and a health state with BCVA score of 86-100 associated with a mean utility value of 0.860. At the end of 15 years, the mean cost per QALY gained with ranibizumab monotherapy was £24,028 and with ranibizumab combination therapy was £36,106. Another economic evaluation conducted from a Canadian health system perspective by Haig et al., using the same clinical efficacy estimates from the RESTORE trial as Mitchell et al. found cost per QALY results of CA$24,494 using ranibizumab monotherapy and CA$36,414 of ranibizumab combination therapy relative to laser monotherapy over a three-year time horizon.79 In the treatment of age-related macular degeneration (AMD), a leading cause of severe visual impairment in older adults, we found three economic evaluations conducted in different regional settings.80-82 Utility estimates in these models for severe visual impairment ranged between 0.534 to 0.55 across the three studies, while those for full vision (in at least the better-seeing eye) ranged from 0.653 to 0.89. Incremental cost-effectiveness results from these three studies are listed below. Table 4.11. Key Prior Economic Models Assessing Cost-Effectiveness of Age-Related Macular Degeneration (AMD) Treatments Study Setting Intervention Comparator Time Incremental Cost- (Perspective) Horizon effectiveness ratio (cost per QALY gained) Yanagi et al., Japan Intravitreal Ranibizumab/ 12 IAI dominates (more 201780 (Societal) aflibercept Pegaptanib sodium/ years effective at lower cost) injection Photodynamic therapy/ (IAI) Best Supportive Care Hopley et al., United Screening + No screening or treatment 5 years £22,722 200481* Kingdom treatment (Third Party with high Payer) dose zinc & antioxidants in 65+ year- olds Vottonen & Finland Aflibercept Bevacizumab/Ranibizumab 8 years €1,801,228/Dominates Kankaanpaa82* (Hospital) *These studies assessed cost-effectiveness of treatments for wet AMD ©Institute for Clinical and Economic Review, 2018 Page 55 Final Evidence Report – Voretigene Neparvovec 4.2 Value-Based Price Benchmarks Our value-based benchmark prices for VN treatment at ages 15 and 3, from both US health care system and modified societal perspectives, are presented in Table 4.12. As noted in the document, “Modifications to the ICER value assessment framework for treatments for ultra‐rare diseases” (https://icer-review.org/wp-content/uploads/2017/11/ICER-Adaptations-of-Value-Framework-for- Rare-Diseases.pdf), the value-based benchmark price for a drug is defined as the price range that would achieve incremental cost-effectiveness ratios between $100,000 and $150,000 per QALY gained. However, it should be noted that for ultra-rare diseases such as this, decision-makers in the US and in international settings often give special weighting to other benefits and contextual considerations (see Section 6.1) that lead to coverage and funding decisions at higher prices, and thus higher cost-effectiveness ratios, than may be applied to decisions about other treatments. From a health care system perspective, the discounts from WAC that would be required to meet both threshold prices are 75% to 82% for VN treatment at age 15 and 43% to 59% for treatment at age 3. These discounts were reduced when indirect costs were accounted for in the modified societal perspective, at 50% to 57% for treatment at age 15 and from no discount at $150,000 per QALY to 11% to achieve $100,000 per QALY for treatment at age 3. Table 4.12. Value-Based Benchmark Prices for VN for the Treatment of Biallelic RPE65-Mediated Inherited Retinal Disease WAC Price to Achieve Price to Achieve Discount from WAC To $100,000 Per QALY $150,000 Per QALY Reach Thresholds Age 15, US Health Care $850,000 $152,619 $216,738 75% - 82% Perspective Age 3, US Health Care $850,000 $348,098 $481,643 43% - 59% Perspective Age 15, Modified $850,000 $362,524 $426,644 50% - 57% Societal Perspective Age 3, Modified $850,000 $755,633 $889,178 ND – 11% Societal Perspective ND: No Discount Required 4.3 Potential Budget Impact We used the cost-effectiveness model to estimate the potential total budgetary impact of VN for the population with biallelic RPE65-mediated inherited retinal disease. We used the stated cost of VN treatment of $850,000, and the three threshold prices for each drug, in our estimates of budget impact. ©Institute for Clinical and Economic Review, 2018 Page 56 Final Evidence Report – Voretigene Neparvovec Potential Budget Impact Model: Methods We used results from the same model employed for the cost-effectiveness analyses to estimate total potential budget impact. Potential budget impact was defined as the total differential cost of using the new therapy rather than relevant existing therapy for the treated population, calculated as differential health care costs (including drug costs) minus any offsets in these costs from averted health care events. All costs were undiscounted and estimated over one- and five-year time horizons. The five-year timeframe was of primary interest, given the potential for cost offsets to accrue over time and to allow a more realistic impact on the number of participants treated with the new therapy. The potential budget impact analysis included the candidate population eligible for treatment: individuals in the US with biallelic RPE65-mediated inherited retinal disease. To estimate the size of the potential candidate population for treatment, we used inputs from a US Securities and Exchange Commission Form 10-K Annual Report by Spark Therapeutics, Inc., which estimated “that there are approximately 3,500 individuals with RPE65-mediated inherited retinal diseases in the United States and the five major European markets.”83 The US population represents approximately 50.06% of the total population in the US and the five major European markets (Germany, France, Italy, Spain and the United Kingdom). Applying that proportion to the total of 3,500 patients results in an estimate of approximately 1,750 eligible individuals in the US. Assuming equal distribution over five years, this resulted in an estimate of 350 patients eligible for VN in the US per year. ICER’s methods for estimating potential budget impact are described in detail at this link: https://icer-review.org/final-vaf-2017-2019/ and have recently been updated. The intent of our revised approach to budgetary impact is to document the percentage of participants that could be treated at selected prices without crossing a budget impact threshold that is aligned with overall growth in the US economy. Briefly, we evaluate a new drug that would take market share from one or more drugs, and calculate the blended budget impact associated with displacing use of existing therapies with the new intervention. In this analysis, we assumed that VN would not displace an active treatment for biallelic RPE65-mediated inherited retinal disease, as none were available for these individuals. Using this approach to estimate potential budget impact, we then compared our estimates to an updated budget impact threshold that represents a potential trigger for policy mechanisms to improve affordability, such as changes to pricing, payment, or individual eligibility. As described in ICER’s methods presentation (http://icer-review.org/wp-content/uploads/2016/02/ICER-Value- Assessment-Proposed-Updates-Webinar-021317.pdf), this threshold is based on an underlying assumption that health care costs should not grow much faster than growth in the overall national economy. From this foundational assumption, our potential budget impact threshold is derived ©Institute for Clinical and Economic Review, 2018 Page 57 Final Evidence Report – Voretigene Neparvovec using an estimate of growth in US gross domestic product (GDP) +1%, the average number of new drug approvals by the FDA over the most recent two-year period, and the contribution of spending on retail and facility-based drugs to total health care spending. Calculations are performed as shown in Table 4.13. For 2017-18, therefore, the five-year annualized potential budget impact threshold that should trigger policy actions to manage access and affordability is calculated to total approximately $915 million per year for new drugs. Table 4.13. Calculation of Potential Budget Impact Threshold Item Parameter Estimate Source 1 Growth in US GDP, 2017 (est.) +1% 3.20% World Bank, 2016 2 Total health care spending, 2016 ($) $2.71 trillion CMS NHE, 2014 3 Contribution of drug spending to total health 17.7% CMS National Health care spending (%) Expenditures (NHE), 2016; Altarum Institute, 2014 4 Contribution of drug spending to total health $479 billion Calculation care spending ($) (Row 2 x Row 3) 5 Annual threshold for net health care cost $15.3 billion Calculation growth for ALL new drugs (Row 1 x Row 4) 6 Average annual number of new molecular 33.5 FDA, 2016 entity approvals, 2013-2014 7 Annual threshold for average cost growth per $457.5 million Calculation individual new molecular entity (Row 5 ÷ Row 6) 8 Annual threshold for estimated potential $915 million Calculation budget impact for each individual new molecular entity (doubling of Row 7) Potential Budget Impact Model: Results Table 4.14 illustrates the per-patient budget impact calculations, based on the list price for VN ($850,000), and the prices for VN to reach $150,000, $100,000, and $50,000 per QALY ($216,738, $152,619, and $88,499, respectively) compared to standard of care. Note that we used the threshold prices assuming treatment at age 15 rather than age three as we assumed that the prevalent population would be treated initially. Treatment at age three would result in higher threshold prices (due to greater QALY gains) but would apply to the much smaller incident population. ©Institute for Clinical and Economic Review, 2018 Page 58 Final Evidence Report – Voretigene Neparvovec Table 4.14. Per-Patient Budget Impact Calculations Over a Five-year Time Horizon Average Annual Per Patient Budget Impact List Price $150,000/QALY $100,000/QALY $50,000/QALY Voretigene Neparvovec $396,270 $107,080 $77,799 $48,518 Standard of Care $5,775 Difference $390,495 $101,305 $72,024 $42,743 QALY: Quality-Adjusted Life Year The average annual potential budgetary impact when using the list price ($850,000) was an additional per-patient cost of approximately $390,500. Average annual potential budgetary impact at the three cost-effectiveness threshold prices for the drug ranged from approximately $101,300 per patient using the price ($216,738) to achieve $150,000 per QALY to approximately $42,700 the price ($88,499) to achieve a $50,000 per QALY cost-effectiveness threshold. For VN treatment of individuals with biallelic RPE65-mediated inherited retinal disease, the annual potential budgetary impact of treating the entire eligible population across all prices (the list price and the three cost-effectiveness threshold prices for $50,000, $100,000, and $150,000 per QALY) did not exceed the $915 million threshold. The greatest potential annual budget impact of treating the described population with VN was at the list price of $850,000, reaching 33% of the $915 million threshold. This was largely due to the relatively small number of patients assumed to be treated per year (350) and the relatively low health care costs incurred following initial treatment with VN. 4.4 Summary and Comment Limitations This study had several limitations. First, the natural history of RPE65-mediated inherited retinal disease has not been thoroughly studied, therefore our underlying disease models have limited data. Second, we were limited in measures of effectiveness for VN to those measures that were captured in the clinical trials as outcomes, as well as in what measures could be linked to quality of life. Because the majority of existing quality-of-life literature for blindness has used visual acuity, we were unable to thoroughly utilize all meaningful outcome measures from the clinical trials. Additionally, costs and quality of life measures have not, to our knowledge, been studied for this specific patient population; therefore, we assumed similarities between this population and people with other types of blindness or visual impairment. Because these populations are likely very different, and of older age groups, than the RPE65-mediated retinal disease population, this may have led to biased estimates, particularly for quality-of-life. ©Institute for Clinical and Economic Review, 2018 Page 59 Final Evidence Report – Voretigene Neparvovec Conclusions We found that VN improves patient health outcomes compared to standard of care. The high cost makes this unlikely to be a cost-effective intervention at commonly used cost-effectiveness thresholds. However, if a societal perspective is used for a younger population, VN is likely to be cost-effective compared to standard of care at a threshold of $150,000/QALY. In addition, for ultra- rare diseases, decision-makers in the US and in international settings often give special weighting to other benefits and to contextual considerations that lead to coverage and funding decisions at higher prices, and thus higher cost-effectiveness ratios, than applied to decisions about other treatments. We found that VN provided more health benefits when given to a younger population, and was therefore more likely to be cost-effective for younger individuals. We also found that inclusion of indirect and non-medical costs decreased the total incremental costs for VN, and therefore decreased cost-effectiveness ratios. ©Institute for Clinical and Economic Review, 2018 Page 60 Final Evidence Report – Voretigene Neparvovec 5. Additional Considerations 5.1 Other Benefits and Contextual Considerations Our reviews seek to provide information on other benefits offered by the intervention to the impacted individuals, caregivers, the delivery system, other patients, or the public that would not have been considered as part of the evidence on comparative clinical effectiveness. These elements are listed in the table below. Table 5.1. Potential Other Benefits or Contextual Considerations Potential Other Benefits This intervention provides significant direct patient health benefits that are not adequately captured by the QALY. This intervention offers reduced complexity that will significantly improve patient outcomes. This intervention will reduce important health disparities across racial, ethnic, gender, socio-economic, or regional categories. This intervention will significantly reduce caregiver or broader family burden. This intervention offers a novel mechanism of action or approach that will allow successful treatment of many patients who have failed other available treatments. This intervention will have a significant impact on improving the patient’s ability to return to work or school and/or their overall productivity. This intervention will have a significant positive impact outside the family, including on schools and/or communities. This intervention will have a significant impact on the entire “infrastructure” of care, including effects on screening for affected patients, on the sensitization of clinicians, and on the dissemination of understanding about the condition, that may revolutionize how patients are cared for in many ways that extend beyond the treatment itself. Other important benefits or disadvantages that should have an important role in judgments of the value of this intervention. Potential Other Contextual Considerations This intervention is intended for the care of individuals with a condition of particularly high severity in terms of impact on length of life and/or quality of life. This intervention is intended for the care of individuals with a condition that represents a particularly high lifetime burden of illness. This intervention is the first to offer any improvement for patients with this condition. Compared to best supportive treatment, there is significant uncertainty about the long-term risk of serious side effects of this intervention. Compared to best supportive treatment, there is significant uncertainty about the magnitude or durability of the long-term benefits of this intervention. There are additional contextual considerations that should have an important role in judgments of the value of this intervention. ©Institute for Clinical and Economic Review, 2018 Page 61 Final Evidence Report – Voretigene Neparvovec Individuals born with biallelic RPE65-mediated retinal diseases currently have no therapies that alter the progression of vision loss. As such, VN represents the first therapy that may stabilize, delay, halt, or reverse loss of vision. Additionally, the availability of treatment may change the paradigm of care by fostering improved screening processes, including genetic testing. Although biallelic RPE65-mediated inherited retinal disease represents a lifelong condition, its perceived severity is highly individual.84 Several individuals who received VN appreciated the improvements in self-confidence, mobility, and independence that they felt following treatment. These benefits may not be adequately captured in the QALY. A qualitative study of research priorities for people with visual impairments in the Netherlands showed that improving mobility in a visually-oriented society has the potential to improve quality of life through increased independence, decreased social isolation, and improved overall enjoyment.85 Although the overall impact of VN on productivity has not yet been studied, it can be postulated that improvements in independence, mobility, and overall visual function may expand the range of employment options open to individuals with biallelic RPE65-mediated inherited retinal disease and increase their ability to participate in social activities. One phase III participant provided insight into how treatment with VN directly allowed her to perform work that she would not be able perform had she not received treatment. Improvements in independence also have the potential to add value to the lives of parents, caregivers and other friends or family members who often make special accommodations to their homes, routines, and employment to ensure that the needs of those with visual impairment are met. In our discussion with impacted individuals and parents, some expressed the improved ability to navigate their school settings without assistance and shared stories of successful transitions between education and the workforce. As discussed in the Controversies and Uncertainties section, the degree to which VN may alter disease progression over the long term is unknown. Likewise, while the adverse events surrounding administration of the therapy appear mild to moderate in severity, the long-term risk of serious side effects remains unclear. Finally, as with many new therapies entering the market, the potential for VN to exacerbate health disparities cannot be ignored. Spark Therapeutics has publicly stated that VN will only be available in a limited number of Centers of Excellence that specialize in inherited retinal diseases.9 Individuals who do not live in close proximity to one of these centers may have difficulty accessing the treatment. Similarly, the high price tag of VN, compounded by the deductibles and copayments associated with the treatment’s surgical component, may make this therapy out of reach for those without adequate insurance coverage. ©Institute for Clinical and Economic Review, 2018 Page 62 Final Evidence Report – Voretigene Neparvovec 5.2 Identification of Low-Value Services As described in its Final Value Assessment Framework for 2017-2019, ICER will now include in its reports information on wasteful or lower-value services in the same clinical area that could be reduced or eliminated to create headroom in health care budgets for higher-value innovative services (for more information, see https://icer-review.org/final-vaf-2017-2019/). Throughout this review process, ICER encouraged all stakeholders to suggest services (including treatments and mechanisms of care) currently used for people with visual disorders that could be reduced, eliminated, or made more efficient. We were looking for information on low-value services used in the management of visual disorders beyond the potential offsets that arise from a new intervention. We received no such suggestions. ©Institute for Clinical and Economic Review, 2018 Page 63 Final Evidence Report – Voretigene Neparvovec 6. Summary of the Votes and Considerations for Policy 6.1 About the Midwest CEPAC Process During Midwest CEPAC public meetings, the Midwest CEPAC Panel deliberates and votes on key questions related to the systematic review of the clinical evidence, an economic analysis of the applications of treatments under examination, and the supplementary information presented. Panel members are not pre-selected based on the topic being addressed and are intentionally selected to represent a range of expertise and diverse perspectives. Acknowledging that any judgment of evidence is strengthened by real-life clinical and patient perspectives, subject matter experts are recruited for each meeting topic and provide input to Midwest CEPAC Panel members before the meeting to help clarify their understanding of the different interventions being analyzed in the evidence review. The same clinical experts serve as a resource to the Midwest CEPAC Panel during their deliberation, and help to shape recommendations on ways the evidence can apply to policy and practice. After the Midwest CEPAC Panel votes, a policy roundtable discussion is held with the Midwest CEPAC Panel, clinical experts, patient advocates, payers, and when feasible, manufacturers. The goal of this discussion is to bring stakeholders together to apply the evidence to guide patient education, clinical practice, and coverage and public policies. Participants on policy roundtables are selected for their expertise on the specific meeting topic, are different for each meeting, and do not vote on any questions. At the January 25, 2018 meeting, the Midwest CEPAC Panel discussed issues regarding the application of the available evidence to help patients, clinicians, and payers address important questions related to the use of voretigene neparvovec for biallelic RPE65-mediated retinal disease. Following the evidence presentation and public comments (public comments from the meeting can be accessed https://www.youtube.com/watch?v=c6y2uKBacfc&feature=youtu.be, starting at minute (1:43:30), the Midwest CEPAC Panel voted on key questions concerning the comparative clinical effectiveness, comparative value, and other benefits and contextual considerations related to voretigene neparvovec. These questions are developed by the ICER research team for each assessment to ensure that the questions are framed to address the issues that are most important in applying the evidence to support clinical practice, medical policy decisions, and patient decision- making. The voting results are presented below, along with specific considerations mentioned by Midwest CEPAC Panel members during the voting process. ©Institute for Clinical and Economic Review, 2018 Page 64 Final Evidence Report – Voretigene Neparvovec In its deliberations and votes related to value, the Midwest CEPAC Panel considered the individual patient benefits, and incremental costs to achieve such benefits, from a given intervention over the long term. There are four elements to consider when deliberating on long-term value for money (see Figure X below): 1. Comparative clinical effectiveness is a judgment of the overall difference in clinical outcomes between two interventions (or between an intervention and placebo), tempered by the level of certainty possible given the strengths and weaknesses of the body of evidence. Midwest CEPAC uses the ICER Evidence Rating Matrix as its conceptual framework for considering comparative clinical effectiveness. 2. Estimated incremental cost-effectiveness is the average incremental cost per patient of one intervention compared to another to achieve a desired “health gain,” such as an additional stroke prevented, case of cancer diagnosed, or gain of a year of life. Alternative interventions are compared in terms of cost per unit of effectiveness, and the resulting comparison is presented as a cost-effectiveness ratio. Relative certainty in the cost and outcome estimates continues to be a consideration. As a measure of cost-effectiveness, the Midwest CEPAC voting panel follows common academic and health technology assessment standards by using cost per quality-adjusted life year (QALY), with formal voting on “long- term value for money” when the base case incremental cost-effectiveness ratio is between $50,000 per QALY and $175,000 per QALY. 3. Other benefits refer to any significant benefits or disadvantages offered by the intervention to the individual patient, caregivers, the delivery system, other patients, or the public that would not have been considered as part of the evidence on comparative clinical effectiveness. Examples of other benefits include better access to treatment centers, mechanisms of treatment delivery that require fewer visits to the clinician’s office, treatments that reduce disparities across various patient groups, and new potential mechanisms of action for treating clinical conditions that have demonstrated low rates of response to currently available therapies. Other disadvantages could include increased burden of treatment on patients or their caregivers. For each intervention evaluated, it will be open to discussion whether other benefits or disadvantages such as these are important enough to factor into the overall judgment of long-term value for money. There is no quantitative measure for other benefits or disadvantages. 4. Contextual considerations include ethical, legal, or other issues (but not cost) that influence the relative priority of illnesses and interventions. Examples of contextual considerations include whether there are currently any existing treatments for the condition, whether the ©Institute for Clinical and Economic Review, 2018 Page 65 Final Evidence Report – Voretigene Neparvovec condition severely affects quality of life or not, and whether there is significant uncertainty about the magnitude of benefit or risk of an intervention over the long term. There is no quantitative measure for contextual considerations. Figure 6.1. Conceptual Structure of Long-term Value for Money ©Institute for Clinical and Economic Review, 2018 Page 66 Final Evidence Report – Voretigene Neparvovec 6.2 Voting Results 1) For patients with RPE-65 mediated inherited retinal disease, is the evidence adequate to demonstrate that the net health benefit of treatment with voretigene neparvovec is greater than that of supportive care? Yes: 12 votes No: 0 votes Comment: After a discussion of the potential harms of the surgery associated with the therapy, the panel voted unanimously in the affirmative. 2) When compared to best supportive care, does voretigene neparvovec offer one or more of the following “other benefits” for patients with RPE-65 mediated inherited retinal disease? (select all that apply) This intervention provides significant direct patient health benefits that are not 6/12 adequately captured by the QALY. This intervention offers reduced complexity that will significantly improve patient 1/12 outcomes. This intervention will reduce important health disparities across racial, ethnic, gender, 2/12 socioeconomic, or regional categories. This intervention will significantly reduce caregiver or broader family burden. 12/12 This intervention offers a novel mechanism of action or approach that will allow 11/12 successful treatment of many patients who have failed other available treatments. This intervention will have a significant impact on improving return to work and/or 12/12 overall productivity. This intervention will have a significant positive impact outside the family, including 7/12 on schools and/or communities. This intervention will have a significant impact on the entire “infrastructure” of care, 5/12 including effects on screening for affected patients, on the sensitization of clinicians, and on the dissemination of understanding about the condition, that may revolutionize how patients are cared for in many ways that extend beyond the treatment itself. ©Institute for Clinical and Economic Review, 2018 Page 67 Final Evidence Report – Voretigene Neparvovec Comments: The panel overwhelmingly agreed that this therapy would impact caregiver/family burden, impact an individual’s ability to work and his or her overall productivity, and that this intervention represented a novel mechanism of action that will allow individuals the opportunity to receive effective treatments where before there were none. Some members of the panel raised concerns that health disparities could be increased if disadvantaged patients were less likely to be diagnosed and receive treatment. One panel member raised the possibility of additional benefits related to the impact on developmental trajectory and achievement in school, building social relationships, and prevention of mental illness by intervening early in a child’s life. 3) Are any of the following contextual considerations important in assessing voretigene neparvovec’s long-term value for money in patients with RPE-65 mediated inherited retinal disease? (select all that apply) This intervention is intended for the care of individuals with a condition of particularly 9/12 high severity in terms of impact on length of life and/or quality of life. This intervention is intended for the care of individuals with a condition that 10/12 represents a particularly high lifetime burden of illness. This intervention is the first to offer any improvement for patients with this condition. 11/12 Compared to the comparator, there is significant uncertainty about the long-term risk 7/12 of serious side effects of this intervention. Compared to the comparator, there is significant uncertainty about the magnitude or 11/12 durability of the long-term benefits of this intervention. Comment: The panel had substantial concerns about the durability of the effect of VN, given only short-term data, and some concerns about whether there were adequate data on safety. As indicated by the votes, the panel felt that all the above contextual considerations were relevant in assessing VN. ©Institute for Clinical and Economic Review, 2018 Page 68 Final Evidence Report – Voretigene Neparvovec 4) Given the available evidence on comparative clinical effectiveness and incremental cost effectiveness, and considering other benefits and contextual considerations, what is the long-term value for money of voretigene neparvovec compared with supportive care for patients with RPE-65 mediated inherited retinal disease? Low: 3 votes Intermediate: 7 votes High: 2 votes Comment: Panel members indicated that they were weighing other benefits and contextual considerations in deciding on their votes, particularly mentioning uncertainty around long-term durability of effect but also the novel mechanism of action of this first gene therapy. Panelists voting Low and Intermediate also cited the very high cost of VN. 6.3 Roundtable Discussion and Key Policy Implications Following its deliberation on the evidence, the Midwest CEPAC Panel engaged in a moderated discussion with a policy roundtable about how best to apply the evidence on Voretigene Neparvovec for Biallelic RPE65-Mediated Retinal Disease to policy and practice. The policy roundtable members included two patient representatives; one clinical expert and two payers. The discussion reflected multiple perspectives and opinions, and therefore, none of the statements below should be taken as a consensus view held by all participants. The names of the Policy Roundtable participants are shown below, and conflict of interest disclosures for all meeting participants can be found in (Appendix H). Table 6.1. Policy Roundtable Members Name Title and Affiliation Katelyn Chandler Corey, MSc Participant in VN phase III trial Patrick Gleason, PharmD Senior Director Health Outcomes, Prime Therapeutics Christine Kay, MD Vitreoretinal Surgeon, Vitreo Retinal Associates Janet LaBreck Former Commissioner, Rehabilitation Services Administration Bill Martin Vice President/GM, Accredo Chief Commercial Officer, Express Scripts The roundtable discussion was facilitated by Dr. Steven Pearson, MD, MSc, President of ICER. The main themes and recommendations from the discussion are organized by audience and summarized below. ©Institute for Clinical and Economic Review, 2018 Page 69 Final Evidence Report – Voretigene Neparvovec 1. The price of VN exceeds usual thresholds for cost effectiveness, but this treatment is the first available for a serious ultra-rare disorder. Payers and other policymakers seeking to judge the value of VN should recognize the heightened responsibility they have in evaluating treatments like VN to consider the treatment’s broader benefits to patients and society while simultaneously working to maintain affordability of health insurance for all patients now and in the future. 2. Even at its current price, the small number of potential patients means that the cumulative costs for treatment with VN will cause no immediate shock to the affordability of health care. Nonetheless, all stakeholders in the health system should realize that a growing stream of treatments for rare and ultra-rare disorders cannot all be priced at levels far above traditional cost-effectiveness thresholds without seriously threatening the financial sustainability of the health system. All stakeholders must therefore collaborate to develop new approaches to pricing and payment for these treatments that can reward innovative therapies in proportion to their benefits for patients while ensuring the restraint necessary to preserve access to high-value care for all patients. The majority of the MW CEPAC voted that treatment with VN represents intermediate long- term value for money. There are unusually large other benefits and contextual considerations with VN, including its position as the first gene therapy in the United States that targets a disease caused by mutations in a specific gene. As additional gene therapies become available, payers will need to carefully assess the stresses caused by very high prices for therapies for ultra-rare conditions. 3. Manufacturers should take the lead responsibility to reach out to public and private payers ahead of FDA approval to negotiate innovative pricing and reimbursement strategies for high-cost therapies like VN that are delivered once or over a short term but that offer the potential for long-term patient benefits. The development of reimbursement strategies for VN by Spark Therapeutics should be considered as potential best practice by other manufacturers of high cost therapies for ultra-rare conditions. These include work on outcomes-based rebates/refunds, amortizing payments, and refunding out-of-pocket costs to patients from copayments, coinsurance, and deductibles. 4. The Centers for Medicare and Medicaid Services (CMS) should take steps to permit private payers to use innovative payment mechanisms without triggering Medicaid Best Price constraints. The manufacturer of VN has worked with payers to develop an outcomes-based rebate plan. Because only small numbers of patients will be treated with VN, the manufacturer is limited in the size of the rebate/refund that can be returned to payers on any one patient ©Institute for Clinical and Economic Review, 2018 Page 70 Final Evidence Report – Voretigene Neparvovec without risking lowering the overall price of VN that would be paid by CMS. This issue is discussed in ICER’s 2017 white paper on gene therapy. CMS should work with payers and the manufacturer to allow large outcome-based rebates/refunds to both public and private payers for treatments of ultra-rare disorders without triggering the Medicaid Best Price provision. 5. Clinical societies, patient groups, and the manufacturer of VN should work to educate all optometrists and ophthalmologists about RPE65-mediated retinal diseases and develop referral networks to facilitate rapid diagnosis. The availability of VN provides an important treatment option for a small subset of patients with IRDs. Finding patients while they still have the greatest preserved vision will increase the value of this therapy. We heard from clinical experts that diagnosis of IRDs is often delayed. Providers likely to see such patients, particularly optometrists and ophthalmologists (both pediatric and adult ophthalmologists), should be educated about presentations of RPE65-mediated retinal disease and the importance of early diagnosis, and provided with information on specialist referral and genetic diagnosis. 6. Payers and the manufacturer should collaborate with retinal specialists to develop policies that promote appropriate access to genetic testing for individuals at high-risk of treatable genetic retinal diseases. Currently the manufacturer of VN is offering free genetic testing. In the future, it is possible that the cost of such testing will be shifted to payers. Payers need to ensure that prior authorization for genetic testing will not serve as a barrier to access to treatment. Clinical societies need to develop guidelines about the proper use of genetic testing for RPE65- mediated retinal disease. 7. Medicaid should assure that its reimbursement policies and clinical networks can support appropriate identification and referral of patients for treatment with VN. Medicaid reimbursement can be so low that in some areas few specialists are willing to see patients with Medicaid insurance. Managed Medicaid programs may also offer extremely limited networks of specialists, making it difficult for Medicaid patients to find a specialist without extensive travel. Medicaid programs should address these issues in order to ensure that patients who need evaluation by a trained retinal specialist for a potentially treatable genetic retinal disorder can be seen without significant delay or travel burden. ©Institute for Clinical and Economic Review, 2018 Page 71 Final Evidence Report – Voretigene Neparvovec 8. Self-insured entities, especially smaller employers, and insurers, should purchase reinsurance or adopt other measures to help manage the potential for unanticipated costs of very expensive treatments such as gene therapy. VN is likely the first of many coming gene therapies for ultra-rare conditions. While the overall budget impact across the health system of VN is not large because there are so few patients with RPE65-mediated IRDs, treating one or two such patients (as might occur with siblings in a family) could create catastrophic losses for a small employer or payer. Reinsurance could help prevent rare catastrophic losses. However, this is likely only a short- term solution. In the future, reinsurance companies may exclude specific high cost treatments such as gene therapy, and other societal models of reimbursement will need to be considered. Arrangements between payers and manufacturers that spread payments out over time can help amortize the financial burden that might otherwise occur in a single year. 9. Payers should seek to negotiate payment terms that are as similar as possible across the limited number of Centers of Excellence that will provide VN. Payers should be transparent if financial considerations lead them to require patients to travel to more- distant Centers for treatment. The manufacture expects to name seven centers of excellence to administer treatment with VN. Each of these centers will likely be able to negotiate individually with payers, and payers will be unlikely to allow their covered patients to access treatment at all seven centers. Payers should be transparent with patients if they require longer travel distances for the purpose of achieving cost savings. 10. Manufacturers should work with researchers, as well as patients and families, to link novel outcome measures such as the navigation test used to evaluate VN to established functional and quality of life measures. A novel outcome measure, such as the Multi-luminance Mobility Test (MLMT), may be appropriate for novel therapeutic areas. While the MLMT integrates the functions of visual acuity, visual field, and light sensitivity into a test for functional vision, the test has not been validated in relation to existing functional and quality of life measures. The absence of such validation and linkage interferes with the assessment of the value and value-based price of a new therapy. 11. Researchers and the manufacturer of VN should work to identify clinical characteristics that better predict patients most likely to benefit from treatment. Biallelic RPE65-mediated retinal disease affects groups of patients of different ages and with variable clinical presentations. Future research is required to identify those patients who ©Institute for Clinical and Economic Review, 2018 Page 72 Final Evidence Report – Voretigene Neparvovec have a higher chance of an increased benefit. A better understanding of the relation between baseline measures of the retina (such as optical coherence tomography) and outcomes will improve the selection of patients for treatment, and thus the long-term value for money of VN. However, if in the future it is possible to give very large outcome-based rebates/refunds for VN, it may be appropriate to treat even patients in the later stages of RPE65-mediated IRDs with a goal of potentially delaying further declines in vision. **** This is the first ICER review of voretigene neparvovec. ©Institute for Clinical and Economic Review, 2018 Page 73 Final Evidence Report – Voretigene Neparvovec References 1. Steinkuller PG, Du L, Gilbert C, Foster A, Collins ML, Coats DK. Childhood blindness. J AAPOS. 1999;3(1):26-32. 2. Sharif W, Sharif Z. Leber's congenital amaurosis and the role of gene therapy in congenital retinal disorders. Int J Ophthalmol. 2017;10(3):480-484. 3. Samardzija M, Barben M, Geiger P, Grimm C. The Consequences of Hypomorphic RPE65 for Rod and Cone Photoreceptors. Adv Exp Med Biol. 2016;854:341-346. 4. Reape K, Chung D, Schaefer G, et al. Natural History of Individuals With Retinal Degeneration Due to Biallelic Mutations in the RPE65 Gene. Paper presented at: ARVO2017; Baltimore, MD. 5. Sparrow JR, Hicks D, Hamel CP. The retinal pigment epithelium in health and disease. Curr Mol Med. 2010;10(9):802-823. 6. Astuti GD, Bertelsen M, Preising MN, et al. Comprehensive genotyping reveals RPE65 as the most frequently mutated gene in Leber congenital amaurosis in Denmark. Eur J Hum Genet. 2016;24(7):1071-1079. 7. Kumaran N, Moore AT, Weleber RG, Michaelides M. Leber congenital amaurosis/early-onset severe retinal dystrophy: clinical features, molecular genetics and therapeutic interventions. Br J Ophthalmol. 2017;101(9):1147-1154. 8. Chung D, Wellman J, E. L, al. e. The natural history of disease progression in patients with RPE65- mediated inherited retinal dystrophy. Invest Ophthalmol Vis Sci. 2016;2017(Oct 15):6588. http://iovs.arvojournals.org/article.aspx?articleid=2564400. 9. Spark Therapeutics Inc. FDA ADVISORY COMMITTEE BRIEFING DOCUMENT Spark Therapeutics, IncLUXTURNATM (voretigene neparvovec). In: MEETING OF THE CELLULAR, TISSUE, AND GENE THERAPIES ADVISORY COMMITTEE MEETING DATE: October 12, 2017 - Available For Public Release.2017: https://www.fda.gov/downloads/AdvisoryCommittees/CommitteesMeetingMaterials/BloodVac cinesandOtherBiologics/CellularTissueandGeneTherapiesAdvisoryCommittee/UCM579300.pdf. Accessed 2017-10-12. 10. den Hollander AI, Roepman R, Koenekoop RK, Cremers FP. Leber congenital amaurosis: genes, proteins and disease mechanisms. Prog Retin Eye Res. 2008;27(4):391-419. 11. Weleber R, Francis P, Trzupek K, Beattie C. Leber Congenital Amaurosis. In: Pagon RA AM, Ardinger HH, et al., ed. GeneReviews. Seattle: University of Washington; 2013. 12. Paunescu K, Wabbels B, Preising MN, Lorenz B. Longitudinal and cross-sectional study of patients with early-onset severe retinal dystrophy associated with RPE65 mutations. Graefes Arch Clin Exp Ophthalmol. 2005;243(5):417-426. 13. Alkharashi M, Fulton AB. Available Evidence on Leber Congenital Amaurosis and Gene Therapy. Semin Ophthalmol. 2017;32(1):14-21. 14. Chung DC, Traboulsi EI. Leber congenital amaurosis: clinical correlations with genotypes, gene therapy trials update, and future directions. J AAPOS. 2009;13(6):587-592. 15. Walia S, Fishman GA, Jacobson SG, et al. Visual acuity in patients with Leber's congenital amaurosis and early childhood-onset retinitis pigmentosa. Ophthalmology. 2010;117(6):1190- 1198. 16. Jacobson SG, Aleman TS, Cideciyan AV, et al. Defining the residual vision in leber congenital amaurosis caused by RPE65 mutations. Invest Ophthalmol Vis Sci. 2009;50(5):2368-2375. ©Institute for Clinical and Economic Review, 2018 Page 74 Final Evidence Report – Voretigene Neparvovec 17. ClinicalTrials.gov. Search for trials using AAV. https://clinicaltrials.gov/ct2/results?cond=&term=aav&cntry1=&state1=&recrs=. Accessed Oct 19, 2017. 18. Bennett J. Gene Therapy for Leber’s 2 Congenital Amaurosis Due to RPE65 Mutations. In: Rakoczy PE, ed. Gene- and cell-based treatment strategies for the eye. Heidelberg: Springer; 2015:9-26. 19. Mingozzi F, High KA. Overcoming the Host Immune Response to Adeno-Associated Virus Gene Delivery Vectors: The Race Between Clearance, Tolerance, Neutralization, and Escape. Annu Rev Virol. 2017;4(1):511-534. 20. Naso MF, Tomkowicz B, Perry WL, 3rd, Strohl WR. Adeno-Associated Virus (AAV) as a Vector for Gene Therapy. BioDrugs. 2017. 21. Russell S, Bennett J, Wellman JA, et al. Efficacy and safety of voretigene neparvovec (AAV2- hRPE65v2) in patients with RPE65-mediated inherited retinal dystrophy: a randomised, controlled, open-label, phase 3 trial. Lancet. 2017. 22. American Academy of Ophthalmology. EyeWiki-Pars Plana Vitrectomy. 2015; http://eyewiki.aao.org/Pars_Plana_Vitrectomy. Accessed Nov 8, 2017. 23. Administration FaD. Luxterna package insert. In: Vaccine BaB, ed. https://www.fda.gov/BiologicsBloodVaccines/CellularGeneTherapyProducts/ApprovedProducts/ ucm589507.htmhttps://www.fda.gov/BiologicsBloodVaccines/CellularGeneTherapyProducts/Ap provedProducts/ucm589507.htm2017. 24. U.S. Food and Drug Administration (FDA). Cellular Tissue and Gene Therapies Advisory Committee Meeting. Open Public Hearing. October 12, 2017. 25. Cornell University. Disability Statistics. 2015 American Community Survey. 2016; http://www.disabilitystatistics.org/reports/acs.cfm?statistic=4. Accessed November 7, 2017. 26. Stone EM, Aldave AJ, Drack AV, et al. Recommendations for genetic testing of inherited eye diseases: report of the American Academy of Ophthalmology task force on genetic testing. Ophthalmology. 2012;119(11):2408-2410. 27. Chung DC, McCague S, Yu ZF, et al. Novel mobility test to assess functional vision in patients with inherited retinal dystrophies. Clin Exp Ophthalmol. 2017. 28. Annual report [press release]. 2016. 29. Stephen Russell MJB, MD, PhD; Jennifer A. Wellman, MS; Daniel C. Chung, DO, MA; Katherine A. High, MD; Zi-Fan Yu, ScD; Amy Tillman, MS; Albert M. Maguire, MD,. Two-Year Results for a Phase 3 Trial of Voretigene Neparvovec in Biallelic RPE65-mediated Inherited Retinal Disease. Paper presented at: ARVO2017; Baltimore, MD. 30. Russell S, Bennett J, Wellman J, et al. Phase 3 Trial Update of Voretigene Neparvovec in Biallelic RPE65 Mutation-Associated Inherited Retinal Disease. Paper presented at: Annual Meeting of Ophthalmology2017; New Orleans, LA. 31. Pierce EA, Bennett J. The Status of RPE65 Gene Therapy Trials: Safety and Efficacy. Cold Spring Harb Perspect Med. 2015;5(9):a017285. 32. Bennett J, Ashtari M, Wellman J, et al. AAV2 gene therapy readministration in three adults with congenital blindness. Sci Transl Med. 2012;4(120):120ra115. 33. Maguire AM, High KA, Auricchio A, et al. Age-dependent effects of RPE65 gene therapy for Leber's congenital amaurosis: a phase 1 dose-escalation trial. Lancet. 2009;374(9701):1597- 1605. 34. Bennett J, Wellman J, Marshall KA, et al. Safety and durability of effect of contralateral-eye administration of AAV2 gene therapy in patients with childhood-onset blindness caused by RPE65 mutations: a follow-on phase 1 trial. Lancet. 2016;388(10045):661-672. ©Institute for Clinical and Economic Review, 2018 Page 75 Final Evidence Report – Voretigene Neparvovec 35. U.S. Food and Drug Administration (FDA). FDA Briefing Document Advisory Committee Meeting October 12, 2017. BLA 125610 Voretigene Neparvovec Spark Therapeutics, Inc. 2017; https://www.fda.gov/downloads/AdvisoryCommittees/CommitteesMeetingMaterials/BloodVac cinesandOtherBiologics/CellularTissueandGeneTherapiesAdvisoryCommittee/UCM579290.pdf. Accessed Oct 12, 2017. 36. Koenekoop RK, Lopez I, den Hollander AI, Allikmets R, Cremers FP. Genetic testing for retinal dystrophies and dysfunctions: benefits, dilemmas and solutions. Clin Exp Ophthalmol. 2007;35(5):473-485. 37. Wright AF. Long-term effects of retinal gene therapy in childhood blindness. N Engl J Med. 2015;372(20):1954-1955. 38. Cideciyan AV, Jacobson SG, Beltran WA, et al. Human retinal gene therapy for Leber congenital amaurosis shows advancing retinal degeneration despite enduring visual improvement. Proc Natl Acad Sci U S A. 2013;110(6):E517-525. 39. Bainbridge JW, Mehat MS, Sundaram V, et al. Long-term effect of gene therapy on Leber's congenital amaurosis. N Engl J Med. 2015;372(20):1887-1897. 40. Jacobson SG, Cideciyan AV, Roman AJ, et al. Improvement and decline in vision with gene therapy in childhood blindness. N Engl J Med. 2015;372(20):1920-1926. 41. Resnikoff S, Keys TU. Future trends in global blindness. Indian journal of ophthalmology. 2012;60(5):387-395. 42. International Council of Ophthalmology. VISUAL STANDARDS ASPECTS and RANGES of VISION LOSS with Emphasis on Population Surveys. Paper presented at: The 29th International Congress of Ophthalmology2002; Sydney, Australia. 43. Lloyd A., Nafees S., Gavriel S., Rousculp M.D., Boye K.S., Ahmad A. Health utility values associated with diabetic retinopathy. Diabet Med. 2008;25:618–624. 44. den Hollander AI. Omics in Ophthalmology: Advances in Genomics and Precision Medicine for Leber Congenital Amaurosis and Age-Related Macular Degeneration. Invest Ophthalmol Vis Sci. 2016;57(3):1378-1387. 45. Morimura H, Fishman GA, Grover SA, Fulton AB, Berson EL, Dryja TP. Mutations in the RPE65 gene in patients with autosomal recessive retinitis pigmentosa or leber congenital amaurosis. Proc Natl Acad Sci U S A. 1998;95(6):3088-3093. 46. Samardzija M, Wenzel A, Naash M, Reme CE, Grimm C. Rpe65 as a modifier gene for inherited retinal degeneration. Eur J Neurosci. 2006;23(4):1028-1034. 47. Jin M, Li S, Hu J, Jin HH, Jacobson SG, Bok D. Functional Rescue of Retinal Degeneration- Associated Mutant RPE65 Proteins. Adv Exp Med Biol. 2016;854:525-532. 48. YourGenome. What is gene therapy? 2017; https://www.yourgenome.org/facts/what-is-gene- therapy. Accessed Oct 21, 2017. 49. Verma IM, Somia N. Gene therapy -- promises, problems and prospects. Nature. 1997;389(6648):239-242. 50. U.S. National Library of Medicine. Trials with QLT091001. https://clinicaltrials.gov/ct2/results?cond=&term=QLT091001. Accessed Oct 23, 2017. 51. Koenekoop RK, Sui R, Sallum J, et al. Oral 9-cis retinoid for childhood blindness due to Leber congenital amaurosis caused by RPE65 or LRAT mutations: an open-label phase 1b trial. Lancet. 2014;384(9953):1513-1520. 52. Scholl HP, Moore AT, Koenekoop RK, et al. Safety and Proof-of-Concept Study of Oral QLT091001 in Retinitis Pigmentosa Due to Inherited Deficiencies of Retinal Pigment Epithelial 65 Protein (RPE65) or Lecithin:Retinol Acyltransferase (LRAT). PLoS One. 2015;10(12):e0143846. 53. Wen Y, Birch DG. Outer Segment Thickness Predicts Visual Field Response to QLT091001 in Patients with RPE65 or LRAT Mutations. Transl Vis Sci Technol. 2015;4(5):8. ©Institute for Clinical and Economic Review, 2018 Page 76 Final Evidence Report – Voretigene Neparvovec 54. Scholl HPN, Strauss RW, Singh MS, et al. Emerging therapies for inherited retinal degeneration. Science Translational Medicine. 2016;8(368):368rv366-368rv366. 55. Mills JO, Jalil A, Stanga PE. Electronic retinal implants and artificial vision: journey and present. Eye (London, England). 2017;31(10):1383-1398. 56. U.S. Food and Drug Administration (FDA). Humanitarian Device Exemption (HDE). https://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfhde/hde.cfm?id=h110002. Accessed Oct 22, 2017. 57. U.S. National Library of Medicine. Stem Cell Ophthalmology Treatment Study II (SCOTS2). https://clinicaltrials.gov/ct2/show/NCT03011541. Accessed Oct 22, 2017. 58. Woolf S. An organized analytic framework for practice guideline development: using the analytic logic as a guide for reviewing evidence, developing recommendations, and explaining the rationale.: Agency for Health Care Policy and Research;1994. 59. Data Presented Today at the American Academy of Ophthalmology 2016 Annual Meeting Reinforce Efficacy and Durability of Voretigene Neparvovec in RPE65-Mediated Inherited Retinal Disease [press release]. 2016. 60. U.S. National Cancer Institute. NCI Dictionary of Genetics Terms. https://www.cancer.gov/publications/dictionaries/genetics-dictionary? Accessed Oct 29, 2017. 61. Glossary PH. Health A-Z. 2017; https://www.ncbi.nlm.nih.gov/pubmedhealth/PMHT0024259/. Accessed October 30, 2017. 62. (CMS) CfMMS. Provider Education Article: Medicare Coverage of Rehabilitation Services for Beneficiaries with Vision Impairment. In: Services DoHH, ed2002. 63. Cook DJ, Mulrow CD, Haynes RB. Systematic Reviews: synthesis of best evidence for clinical decisions. Annals of internal medicine. 1997;126(5):376-380. 64. Moher D, Liberati A, Tetzlaff J, DG A. Preferred reporting items for systematic reviews and meta- analyses: the PRISMA statement. International Journal of Surgery. 2010;8(5):336-341. 65. Ollendorf DA, Chapman R, Pearson SD. Assessing the Effectiveness and Value of Drugs for Rare Conditions - A Technical Brief for the ICER Orphan Drug Assessment & Pricing Summit. In: Boston: Institute for Clinical and Economic Review; 2017: https://icer-review.org/wp- content/uploads/2017/02/ICER_Assessing-the-Value-of-Drugs-for-Rare-Conditions_051017.pdf. Accessed 2017-07-05. 66. Holladay JT. Visual acuity measurements. J Cataract Refract Surg. 2004;30:287-290. 67. Lange C, Feltgen N, Junker B, Schulze-Bonsel K, Bach M. Resolving the clinical acuity categories "hand motion" and "counting fingers" using the Freiburg Visual Acuity Test (FrACT). Graefes Arch Clin Exp Ophthalmol. 2009;247(1):137-142. 68. Robert D. Rondinelli M, PhD, . The Visual System. In: Guides to the Evaluation of Permanent Impairment, Sixth Edition. American Medical Association; 2009:281-319. 69. Jacobson SG, Cideciyan AV, Ratnakaram R, et al. Gene therapy for leber congenital amaurosis caused by RPE65 mutations: safety and efficacy in 15 children and adults followed up to 3 years. Arch Ophthalmol. 2012;130(1):9-24. 70. CDC/NCHS National Vital Statistics System. Life table for the total population: United States, 2014. 2014. 71. Brown MM, Brown GC, Sharma S, Kistler J, Brown H. Utility values associated with blindness in an adult population. Br J Ophthalmol. 2001;85(3):327-331. 72. Ternent L, Vale L, Boachie C, Burr JM, Lois N. Cost-effectiveness of internal limiting membrane peeling versus no peeling for patients with an idiopathic full-thickness macular hole: results from a randomised controlled trial. Br J Ophthalmol. 2012;96(3):438-443. ©Institute for Clinical and Economic Review, 2018 Page 77 Final Evidence Report – Voretigene Neparvovec 73. Brown MM, Brown GC, Lieske HB, Tran I, Turpcu A, Colman S. Societal Costs Associated with Neovascular Age-Related Macular Degeneration in the United States. Retina. 2016;36(2):285- 298. 74. Wittenborn J, Rein D. Cost of Vision Problems: The Economic Burden of Vision Loss and Eye Disorders in the United States. Presented to: Prevent Blindness America: NORC at the University of Chicago;2013. 75. Bennison C, Stephens S, Lescrauwaet B, Van Hout B, Jackson TL. Cost-effectiveness of ocriplasmin for the treatment of vitreomacular traction and macular hole. Journal of Market Access & Health Policy. 2016;4:10.3402/jmahp.v3404.31472. 76. Dunbar JA, Hsu V., Christensen, M., Black, B., Williams, P., Beauchamp, G. . Cost-utility analysis of screening and laser treatment of retinopathy of prematurity. Journal of AAPOS : the official publication of the American Association for Pediatric Ophthalmology and Strabismus. 2009;13(2):186-190. 77. Rothschild MI RR, Brennan KA, et al. The Economic Model of Retinopathy of Prematurity (EcROP) Screening and Treatment: Mexico and the United States. American journal of ophthalmology. 2016;168:110-121. 78. Mitchell P. A, L., Gallagher, M., et al. Cost-effectiveness of ranibizumab in treatment of diabetic macular oedema (DME) causing visual impairment: evidence from the RESTORE trial. The British journal of ophthalmology. 2012;96(5):688-693. 79. Haig J. BMFA. Cost-effectiveness of ranibizumab in the treatment of visual impairment due to diabetic macular edema. J Med Econ. 2016;19(7):663-671. 80. Yanagi Y FA, Barzey V, Adachi K. Cost-effectiveness of intravitreal aflibercept versus other treatments for wet age-related macular degeneration in Japan. J Med Econ. 2017;20(2):204-212. 81. Hopley C, Salkeld G, Wang JJ, Mitchell P. Cost utility of screening and treatment for early age related macular degeneration with zinc and antioxidants. The British Journal of Ophthalmology. 2004;88(4):450-454. 82. Vottonen P, Kankaanpaa E. Cost-effectiveness of treating wet age-related macular degeneration at the Kuopio University Hospital in Finland based on a two-eye Markov transition model. Acta ophthalmologica. 2016;94(7):652-656. 83. U.S. Securities and Exchange Commission (SEC). 10-K Spark Therapeutics, Inc. 2014. 84. Spark Therapeutics Inc. Patients and Families. 2017; http://sparktx.com/patients- families/inherited-retinal-diseases/. Accessed November 3, 2017. 85. Schölvinck A-FM, Pittens CACM, Broerse JEW. The Research Priorities of People with Visual Impairments in the Netherlands. Journal of Visual Impairment & Blindness. 2017;111(3):201- 217. 86. Simonelli F, Maguire AM, Testa F, et al. Gene therapy for Leber's congenital amaurosis is safe and effective through 1.5 years after vector administration. Mol Ther. 2010;18(3):643-650. 87. Testa F, Maguire AM, Rossi S, et al. Three-year follow-up after unilateral subretinal delivery of adeno-associated virus in patients with Leber congenital Amaurosis type 2. Ophthalmology. 2013;120(6):1283-1291. 88. Ashtari M, Cyckowski L, Marshall K, et al. Human visual cortex response to retinal gene therapy re-administration. Molecular Therapy. 2012;20:S114-S115. 89. Ashtari M, Cyckowski L, Zhang G, et al. Retinal gene therapy may alter connectivity of visual pathways. Molecular Therapy. 2013;21:S22. 90. Ashtari M, Nikonova ES, Marshall KA, et al. The Role of the Human Visual Cortex in Assessment of the Long-Term Durability of Retinal Gene Therapy in Follow-on RPE65 Clinical Trial Patients. Ophthalmology. 2017;124(6):873-883. ©Institute for Clinical and Economic Review, 2018 Page 78 Final Evidence Report – Voretigene Neparvovec 91. Sanders GD, Neumann PJ, Basu A, et al. Recommendations for Conduct, Methodological Practices, and Reporting of Cost-effectiveness Analyses: Second Panel on Cost-Effectiveness in Health and Medicine. Jama. 2016;316(10):1093-1103. 92. Hui DJ, Chen Y, Antrilli T, et al. Safety study by validated immunoassays in a phase III study of subjects with inherited retinal dystrophy due to mutations in the gene encoding human retinal pigment epithelium-specific protein 65 (RPE65) injected with adeno-associated viral vectors. Molecular Therapy. 2016;24:S72-S73. 93. Ashtari M, Cook P, Zhang H, et al. Brain pathways enabling vision in LCA patients before and after gene therapy. Molecular Therapy. 2016;24:S105. 94. Maguire AM, Simonelli F, Pierce EA, et al. Safety and efficacy of gene transfer for Leber's congenital amaurosis. N Engl J Med. 2008;358(21):2240-2248. ©Institute for Clinical and Economic Review, 2018 Page 79 Final Evidence Report – Voretigene Neparvovec Appendix A. Search Strategies and Results Table A1. PRISMA 2009 Checklist # Checklist Item TITLE Title 1 Identify the report as a systematic review, meta-analysis, or both. ABSTRACT Structured 2 Provide a structured summary including, as applicable: background; objectives; data Summary sources; study eligibility criteria, participants, and interventions; study appraisal and synthesis methods; results; limitations; conclusions and implications of key findings; systematic review registration number. INTRODUCTION Rationale 3 Describe the rationale for the review in the context of what is already known. Objectives 4 Provide an explicit statement of questions being addressed with reference to participants, interventions, comparisons, outcomes, and study design (PICOS). METHODS Protocol and 5 Indicate if a review protocol exists, if and where it can be accessed (e.g., Web Registration address), and, if available, provide registration information including registration number. Eligibility Criteria 6 Specify study characteristics (e.g., PICOS, length of follow-up) and report characteristics (e.g., years considered, language, publication status) used as criteria for eligibility, giving rationale. Information 7 Describe all information sources (e.g., databases with dates of coverage, contact with Sources study authors to identify additional studies) in the search and date last searched. Search 8 Present full electronic search strategy for at least one database, including any limits used, such that it could be repeated. Study Selection 9 State the process for selecting studies (i.e., screening, eligibility, included in systematic review, and, if applicable, included in the meta-analysis). Data Collection 10 Describe method of data extraction from reports (e.g., piloted forms, independently, Process in duplicate) and any processes for obtaining and confirming data from investigators. Data Items 11 List and define all variables for which data were sought (e.g., PICOS, funding sources) and any assumptions and simplifications made. Risk of Bias in 12 Describe methods used for assessing risk of bias of individual studies (including Individual studies specification of whether this was done at the study or outcome level), and how this information is to be used in any data synthesis. Summary 13 State the principal summary measures (e.g., risk ratio, difference in means). Measures Synthesis of 14 Describe the methods of handling data and combining results of studies, if done, Results including measures of consistency (e.g., I2) for each meta-analysis. Risk of Bias 15 Specify any assessment of risk of bias that may affect the cumulative evidence (e.g., Across studies publication bias, selective reporting within studies). Additional 16 Describe methods of additional analyses (e.g., sensitivity or subgroup analyses, meta- Analyses regression), if done, indicating which were pre-specified. RESULTS ©Institute for Clinical and Economic Review, 2018 Page 80 Final Evidence Report – Voretigene Neparvovec Study Selection 17 Give numbers of studies screened, assessed for eligibility, and included in the review, with reasons for exclusions at each stage, ideally with a flow diagram. Study 18 For each study, present characteristics for which data were extracted (e.g., study size, Characteristics PICOS, follow-up period) and provide the citations. Risk of Bias 19 Present data on risk of bias of each study and, if available, any outcome level within Studies assessment (see item 12). Results of 20 For all outcomes considered (benefits or harms), present, for each study: (a) simple Individual summary data for each intervention group (b) effect estimates and confidence Studies intervals, ideally with a forest plot. Synthesis of 21 Present results of each meta-analysis done, including confidence intervals and Results measures of consistency. Risk of Bias 22 Present results of any assessment of risk of bias across studies (see Item 15). Across Studies Additional 23 Give results of additional analyses, if done (e.g., sensitivity or subgroup analyses, Analysis meta-regression [see Item 16]). DISCUSSION Summary of 24 Summarize the main findings including the strength of evidence for each main Evidence outcome; consider their relevance to key groups (e.g., healthcare providers, users, and policy makers). Limitations 25 Discuss limitations at study and outcome level (e.g., risk of bias), and at review-level (e.g., incomplete retrieval of identified research, reporting bias). Conclusions 26 Provide a general interpretation of the results in the context of other evidence, and implications for future research. FUNDING Funding 27 Describe sources of funding for the systematic review and other support (e.g., supply of data); role of funders for the systematic review. From: Moher D, Liberati A, Tetzlaff J, Altman DG. The PRISMA Group (2009). Preferred Reporting Items for Systematic Reviews and Meta-Analyses: The PRISMA Statement. PLoS Med 6(6): e1000097. doi:10.1371/journal.pmed1000097 ©Institute for Clinical and Economic Review, 2018 Page 81 Final Evidence Report – Voretigene Neparvovec Figure A1. PRISMA Flow Chart Showing Results of Literature Search for Voretigene Neparvovec 360 potentially relevant references screened 306 citations excluded Population: 240 Intervention: 52 Comparator: 0 Outcomes: 3 Study Design: 11 54 references for full text review 40 citations excluded (conference abstract duplicated peer-reviewed publication, wrong intervention, no outcome of interest 14 TOTAL • 1 RCT o 1 publication o 1 poster o 2 conference abstracts • 3 single-arm studies o 8 publications o 2 conference abstracts ©Institute for Clinical and Economic Review, 2018 Page 82 Final Evidence Report – Voretigene Neparvovec Appendix B. Previous Systematic Reviews and Technology Assessments We did not identify any completed technology assessments or peer-reviewed systematic reviews of voretigene neparvovec for Biallelic RPE65-Mediated Retinal Disease, however the National Institute for Health and Care Excellence (NICE) in the UK has published a draft scope for an appraisal of the clinical and cost effectiveness of voretigene. The appraisal’s completion date is yet to be confirmed. NICE: Voretigene neparvovec for treating inherited retinal dystrophies caused by RPE65 gene mutations https://www.nice.org.uk/guidance/indevelopment/gid-ta10200/documents NICE has proposed to appraise the clinical and cost effectiveness of voretigene neparvovec within its marketing authorization for treating inherited retinal dystrophies caused by RPE65 gene mutations. ©Institute for Clinical and Economic Review, 2018 Page 83 Final Evidence Report – Voretigene Neparvovec Appendix C. Ongoing Studies Title, Trial Sponsor, Study Design Treatment Arms Patient Population Key Outcomes Estimated ClinicalTrials.gov Completion Identifier Date Clinical Trial of Gene Phase I/II 1. Low dose Inclusion Criteria Primary Outcome Measures October 2018 Therapy for the AAV-RPE65 • Age ≥3 • Adverse events related to Treatment of Leber Non- subretinal • Early-onset severe retinal dystrophy treatment Long-term Congenital Amaurosis randomized administration consistent with RPE65 deficiency follow-up until (LCA) (OPTIRPE65) Secondary Outcome April 2023 Single group 2. Intermediate Exclusion Criteria Measures MeiraGTx UK II Ltd assignment dose AAV-RPE65 • Females who are pregnant or • Visual function subretinal breastfeeding • Retinal function NCT02781480 Estimated administration • Participation in another research study • Quality of life Enrollment: involving investigational therapy for 27 3. High dose ocular disease within last 6 months AAV-RPE65 subretinal administration ©Institute for Clinical and Economic Review, 2018 Page 84 Final Evidence Report – Voretigene Neparvovec Appendix D. Clinical Effectiveness Supplemental Information Additional Endpoints from Clinical Trials Pupillary light reflex (PLR) Pupillary light reflex (PLR) was measured in early voretigene neparvovec studies as a secondary endpoint. However, in the phase III study, a decision was made by the study sponsor ( ) to make PLR an exploratory instead of secondary endpoint due to issues with control (no untreated eye), measurement (nystagmus) and maintenance of the pupillometer. Exploratory data on PLR are not reported in the phase III manuscript, supplement, or two-year poster. Prior trials stated improved pupillary response after treatment and cite individual study participant data (no aggregate results).33,34 Ocular motility testing The original pilot study of three participants who underwent low dose voretigene neparvovec treatment in the worst-seeing eye used digital eye-movement video to assess ocular motility including nystagmus. Each of the three enrolled individuals had frequent ocular movements of varying degrees at baseline. Following treatment, all three participants had reduced monocular and binocular nystagmus frequency and amplitude which lasted out to 1.5 years.86 It has been hypothesized that improvements in visual acuity may stem from reduced nystagmus.87 The phase III study did not provide baseline or follow-up nystagmus data. Functional magnetic resonance imaging (fMRI) To assess whether treatment with voretigene neparvovec altered the visual cortex responsiveness to light, a longitudinal functional MRI study was performed on a subset of participants enrolled in the Phase I and II studies.88,89 Participants scheduled to receive contralateral treatment in the Phase I follow-on study underwent baseline magnetic resonance imaging (MRI) to identify baseline cortical response. Images were looked at both in the areas of the brain associated with the untreated and originally treated eye and provided a baseline with which to assess changes after contralateral eye treatment (Phase I follow- on study). Follow-up data out to three years shows that in the seven participants imaged, all but one had increased cortical activation following treatment.90 Levels of activation varied widely depending on ©Institute for Clinical and Economic Review, 2018 Page 85 Final Evidence Report – Voretigene Neparvovec the subject’s age, disease progression and location of voretigene neparvovec injection (see Appendix Figure D1).90 A longitudinal regression using mixed effects showed associations between visual cortex activation and clinical measures of visual function. Full-field light sensitivity and pupillary light response were positively correlated with improvement while visual acuity and visual field were not.90 Figure D1. Longitudinal Functional Magnetic Resonance Imaging (MRI) in subset of Phase I follow- on study subjects90 MLMT Scores from Individual Participants in Study 301 There is an interest in whether VN improves sight for younger participants (age less than 10) compared to participants age 10 and older. There were no subgroup analyses performed; however, two figures were provided to the FDA Advisory Panel that provide baseline and one-year score on the MLMT for both eyes and first treated eye (see Figures D2 and D3 below). ©Institute for Clinical and Economic Review, 2018 Page 86 Final Evidence Report – Voretigene Neparvovec Figure D2. MLMT Score Using Both Eyes at Baseline and One Year in Study 301 35 ©Institute for Clinical and Economic Review, 2018 Page 87 Final Evidence Report – Voretigene Neparvovec Figure D3. MLMT Score Using First Eye at Baseline and One Year in Study 301 35 ©Institute for Clinical and Economic Review, 2018 Page 88 Final Evidence Report – Voretigene Neparvovec Appendix E. Comparative Value Supplemental Information Figure E1. Impact Inventory91 ©Institute for Clinical and Economic Review, 2018 Page 89 Final Evidence Report – Voretigene Neparvovec Appendix F. Data Extraction Summary Table Author & Year Study Design and Interventions (n) Major Inclusion & Patient Key Outcomes Harms of Publication Duration of & Dosing Exclusion Criteria Characteristics (Trial Name) Follow-up Schedule Quality rating Russell, Lancet Phase III, open- 1) Voretigene Inclusion Age, yrs 1 year Mean MLMT (SD) 1 year TEAEs (2017)21 label, randomized neparvovec (VN) ≥ 3 years old; biallelic Mean (SD) Both eyes (mITT) in control trial [AAV2-hRPE65v2 RPE65 gene mutation; 1) 14.7(11.8) 1) 1.8(1.1) intervention Manuscript or LUXTURNA™] both eyes 20/60 or 2) 15.9(9.5) 2) 0.2(1.0) group (n=20) Follow-up: 1 year (n=21) worse or visual field < P=0.0013 # events, # Study 301 20° in any meridian; Sex, N(%) First eye/ Second eye patients (% pts) 11 1.5x10 vector sufficient viable retinal Female 1) 1.9(1.2)/ 2.1(1.2) Fair Performed by 5 genomes (vg) per cells; able to perform 1) 12(57) 2) 0.2(0.6)/ 0.1(0.7) Increased surgeons at 2 eye in 0.3ml mobility test (MLMT) 2) 6(60) P=0.0005/ P=0.0001 intraocular hospitals: subretinal but unable to pass at 1 pressure: 5, 4(20) CHOP/U of Iowa injection LUX Age group, N(%) Goldmann visual field, sum Exclusion <10 yrs/ ≥10 yrs total degrees Cataract 4, 3(15) 2) Control group, Participation in gene 1) 9(43)/12(57) Mean (SD) eligible to receive therapy or 2) 4(40)/6(60) 1) 673.9(423.7) Retinal tear 2, VN after 1 year investigational drug 2) 397.8(367.3) 2(10) (n=10) study; used high dose MLMT passing P=0.0059 retinal compounds in level, N(%) Eye inflammation past 18 months; <125 lux/≥125 Humphrey visual field, 6, 2(10) intraocular surgery in lux foveal sensitivity/macula past 6 months; 1) 12(57)/9(43) threshold (dB) Macular hole 2, 1 contraindications to 2) 4(40)/6(60) Mean (SD) (5%) - same eye, operative meds; 1) 25.8(9.1)/ 24.0(8.0) full-thickness conditions that 2) 21.5(8.9)/ 15.8(7.4) macular hole preclude outcome P=0.18/ P=0.0005 spontaneously interpretation resolved ©Institute for Clinical and Economic Review, 2018 Page 90 Final Evidence Report – Voretigene Neparvovec Author & Year Study Design and Interventions (n) Major Inclusion & Patient Key Outcomes Harms of Publication Duration of & Dosing Exclusion Criteria Characteristics (Trial Name) Follow-up Schedule Quality rating Russell, ARVO, See Russell 2017 See Russell 2017 See Russell 2017 See Russell 2017 MLMT- Mean bilateral Total adverse (2017)29 change score (SD) events: 32 in 19 Follow-up: 2 years 1) Original 1) 1.9 (1.1) subjects (66%) Poster for original intervention at 2 2) 2.1 (1.6) intervention group years Cataract: 7 Study 302 and Averaged over both eyes events in 4 1 year from 2) Delayed White light FST Mean subjects (14%) crossover controls intervention at 1 change (cd.s/m2) (SD) “delayed year 1) 2.27 −log10 (1.65) Retinal tear: 3 intervention” 2) 2.86 (1.49) events in 3 subjects (10%) Visual Acuity LogMAR mean (SD) change Retinal deposits: 1) −0.16 (0.36) (+8 letters) 3 events in 3 2) −0.09 (0.22) (+4.5 subjects (10%) letters) Macular hole: 2 Visual Field mean (SD) events in 2 change subjects (7%) Goldmann sum total degrees on GVF III4e Eye inflammation: 1) 311.6 (295.3) 4 events in 2 2) 194.3 (244.7) subjects (7%) Humphrey macula threshold 1) 6.45 (7.35) dB 2) 5.23 (9.92) dB ©Institute for Clinical and Economic Review, 2018 Page 91 Final Evidence Report – Voretigene Neparvovec Author & Year Study Design and Interventions (n) Major Inclusion & Patient Key Outcomes Harms of Publication Duration of & Dosing Exclusion Criteria Characteristics (Trial Name) Follow-up Schedule Quality rating Russell, AAO See Russell 2017 See Russell 2017 See Russell 2017 See Russell 2017 MLMT- Mean bilateral Cataract: 10 (2017)30 change score (SD) events in 5 Follow-up: 3 years 1) Original 1) 1.8 (1.0) subjects (17%) Presentation for original intervention at 3 2) 2.1 (1.6) intervention group years Increased and Averaged over both eyes intraocular 2 year from 2) Delayed White light FST Mean pressure: 7 crossover controls intervention at 2 change (cd.s/m2) (SD) events in 5 “delayed year 1) -2.04 (1.43) subjects (17%) intervention” 2) -2.69 (1.41) Retinal tear: 3 Visual Acuity LogMAR mean events in 3 (SD) change subjects (10%) “unchanged” Retinal deposits: Visual Field mean (SD) 3 events in 3 change subjects (10%) Goldmann sum total degrees on GVF III4e 1) 282.2 (256.6) 2) 182.6 (309.9) Hui, Molecular See Russell 2017 See Russell 2017 See Russell 2017 See Russell 2017 18/21 intervention subjects None Therapy tested negative for T cell (2016)92 Safety study Immunological responses against AAV2 assays designed and RPE65 across all Abstract Follow-up: N/A to monitor timepoints. One subject cellular immune was positive against AAV2 responses. capsid at baseline (55.0 ©Institute for Clinical and Economic Review, 2018 Page 92 Final Evidence Report – Voretigene Neparvovec Author & Year Study Design and Interventions (n) Major Inclusion & Patient Key Outcomes Harms of Publication Duration of & Dosing Exclusion Criteria Characteristics (Trial Name) Follow-up Schedule Quality rating SFU) and positive against RPE65 at the 1 year timepoint (171.7 SFU). Another subject was positive at 1 year for RPE65 only (170.0 SFU). Positive responses were considered very weak with respect to threshold cutoff values. One subject displayed a moderate response (518.3 SFU) against RPE65 at baseline only. Positive T cell responses prior to vector administration are unlikely to be related to gene transfer. Maguire, Lancet Single arm Voretigene Inclusion Age Range: 8-44 Nystagmus frequency (Hz) No serious (2009)33 comparative study neparvovec (VN) LCA diagnosis; 90 days Range 0-3.0 adverse events to contralateral or AAV2- molecular diagnosis of Female, N(%) reported Manuscript eye with some hRPE65v2 RPE65 mutations; age Total: 5/12 Visual acuity normal age- 8-44 yrs; visual acuity < (41.7) improved in 9 patients, Macular hole (1 Study 101 matched 1) Low dose: 20/200 or visual field although not significantly; 2 subject) @ day 14 individuals for 1.5x1010 vector less than 20° Nystagmus unchanged, 1 worsened. post treatment pupillary light genomes (vg) per frequency at Not correlated with dose or reflexes eye in 0.3ml Exclusion baseline age. Vector detected subretinal Participation in trial of Range 0.3-4.2: in tears of 6 Dose escalation injection investigational drug in Visual field improvements subjects ranging (n=3) past 6 months; in 12 patients but with ©Institute for Clinical and Economic Review, 2018 Page 93 Final Evidence Report – Voretigene Neparvovec Author & Year Study Design and Interventions (n) Major Inclusion & Patient Key Outcomes Harms of Publication Duration of & Dosing Exclusion Criteria Characteristics (Trial Name) Follow-up Schedule Quality rating Performed by 2 condition that substantial variability. from 1-4 days physicians at 2 2) Medium dose: precludes accurate Younger subjects saw more after procedure. centers: 4.8x1010 vector measure of endpoints; improvement; correlated CHOP genomes (vg) per lack of sufficient retinal with amount of viable Vector in blood or Seconda eye in 0.3ml cells; ocular surgery in retina serum detected Universitá degli subretinal past 6 months; in 2/3 high dose Studi de Napoli injection (n=6) sensitivity to surgical Pupillary reflex improved in subjects meds; neutralizing treated eye of all subjects Follow-up: 2 yrs 3) High dose: antibodies AAV2 of compared to untreated eye 1.5x1011 vector 1:1000 genomes (vg) per Mobility test eye in 0.3ml 4 children increased subretinal accuracy and speed after tx injection (n=3) Testa, Five Italian See Maguire 2009 See Maguire 2009 Age: 11-26 Best Corrected Visual Acuity None in third yr Ophthalmology patients from (2013) 87 Maguire 2009 Gender: F 2/5; NP01: M: 3/5 Baseline RE: 2.0 Manuscript NP01, NP02, NP03, Year 3 RE: 1.52 NP04 and Eye injected: R NP15 4/5; L 1/5 NP02: Baseline RE: 2.0 Follow-up: 3 yrs Vector volume: Year 3 RE: 1.49 150-300 µL NP03: Concentration: Baseline RE: 1.48 1.0-5.0 (10 per Year 3 RE: 0.96 µL) NP04: ©Institute for Clinical and Economic Review, 2018 Page 94 Final Evidence Report – Voretigene Neparvovec Author & Year Study Design and Interventions (n) Major Inclusion & Patient Key Outcomes Harms of Publication Duration of & Dosing Exclusion Criteria Characteristics (Trial Name) Follow-up Schedule Quality rating Baseline LE: 1.02 Year 3 LE: 0.57 NP15 Baseline RE 0.85 Year 3 RE: 0.56 Bennett, Lancet Phase I follow-on Voretigene See McGuire 2009 See McGuire Goldman visual field, N(%) No AEs related to (2016)34 (Study 101) neparvovec (VN) 2009 Increase ≥20 sum total AAV vector or AAV2- Descriptive stats, N(%) Age at re- degrees: 6(60) Manuscript Treatment of hRPE65v2 calculated by ICER; 3 administration Decrease ≥20 sum total Dellen formation contralateral eye years Range 11- degrees: 3(30) (uneven surface Study 102 Dose: 1.5x1011 46 No change: 1(10) of cornea) - 3 See McGuire 2009 vector genomes patients (vg) to 1 patient from Change in full-field light Follow-up: 3 years contralateral eye first injection sensitivity threshold (dB) Cataracts – 2 pts in 0.3ml not eligible for (>10dB): 8(80) subretinal follow-on SAEs: injection because of Visual acuity Post-op bacterial glaucoma in No change: 8(80) endophthalmitis N= 11 although contralateral Improved: 1(10) with intraocular final data eye Worsened: 1(10) pressure and provided on 10 optic atrophy – 1 patients (one Time between Mobility-at lower light pt patient had post- first and second levels op eye infection) injection (years) No change: 2(20) Myopia requiring Range: 1.71- 1 level: 1(10) correction of >10 4.58 2 levels: 3(30) diopters and 3 levels: 3(30) retinal thinning – 5 levels: 1 (10) 1 pt ©Institute for Clinical and Economic Review, 2018 Page 95 Final Evidence Report – Voretigene Neparvovec Author & Year Study Design and Interventions (n) Major Inclusion & Patient Key Outcomes Harms of Publication Duration of & Dosing Exclusion Criteria Characteristics (Trial Name) Follow-up Schedule Quality rating Pupillary light reflex improvement in all patients Humoral and cell- mediated response to AAV2 at 4 weeks – 1 pt Ashtari, Single arm See Bennett 2016 See Bennett 2016 See Bennett fMRI baseline (after 1st gt, None described Opthalmology N=7 (from 11 that 2016 before contralateral eye) (2017)90 had contralateral and after tx of contralateral eye tx in 102) eye. Increases in right, left Manuscript and total hemisphere Follow-up: 3 yrs cortical activation after gene therapy was associated with improved clinical outcomes of white red and blue light full field stimulus threshold and pupillary reflex Ashtari, See Bennett 2016 See Bennett 2016 See Bennett 2016 See Bennett Tractography results NA Molecular 2016 showed higher RT tract Therapy Pre-tx for 10 GT density for LCA2 patients in (2016)93 contralateral eye 11 matched the hemisphere ipsilateral compared to controls to their untreated eye and Abstract treated eye in For fMRI a higher GS tract density Maguire ipsilateral to their treated (post-tx reported eyes. Control subjects above in Ashtari showed symmetrical tracts Opthamology for both RT and GS 2017) pathways. ©Institute for Clinical and Economic Review, 2018 Page 96 Final Evidence Report – Voretigene Neparvovec Author & Year Study Design and Interventions (n) Major Inclusion & Patient Key Outcomes Harms of Publication Duration of & Dosing Exclusion Criteria Characteristics (Trial Name) Follow-up Schedule Quality rating Follow-up: N/A Ashtari, Science See Bennett 2016 See Bennett 2016 See Bennett 2016 See Bennett Results from DTI, diffusion NA Translational 2016 tractography, and fMRI Medicine 10 LCA2 patients Diffusion tensor along with correlation of (2016)93 and imaging (DTI) to these data with nystagmus 11 examine the measures, age, and length Manuscript demographically effect of of time after treatment in matched sighted deprivation and LCA2 patients suggested controls matched subsequent that retinal gene therapy for age, gender, unilateral retinal may promote ethnicity, and gene therapy on remyelination of handedness. the organization geniculostriate fiber axons and/or as well as local changes Follow-up: N/A reorganization of within the V1 favoring the white matter treated eye. These microstructure in observations suggest that V1. the functional plasticity for patients receiving retinal gene therapy may be related to structural changes in the brain. Ashtari, 5 pts from Bennet See Bennett 2016 See Bennett 2016 See Bennett All subjects showed NA Molecular 2016 after 2016 significant increased Therapy contralateral cortical activations after re- (2012)88 injection administration. While younger subjects showed Abstract Follow-up: N/A considerable activations, there were no significant cortical responses for the ©Institute for Clinical and Economic Review, 2018 Page 97 Final Evidence Report – Voretigene Neparvovec Author & Year Study Design and Interventions (n) Major Inclusion & Patient Key Outcomes Harms of Publication Duration of & Dosing Exclusion Criteria Characteristics (Trial Name) Follow-up Schedule Quality rating baseline of the older subjects. fMRI results revealed significant improvement in visual function of 5 LCA patients who received re- administration of AAV2-hRPE65v2 to their contralateral eye with no adverse effect to the functionality of their previously treated eye. Younger subjects’ cortical activations at baseline may be due to less advanced retinal degeneration as compared to older patients. Bennett, Single arm safety Voretigene See Bennett 2016 Age: Visual acuity, logMAR No surgical Science report on first 3 neparvovec (VN) CH12: 46 (higher=worse) complications Translational patients to receive or AAV2- CH11: 27 post-contralateral injection Medicine contralateral hRPE65v2 NP01: 29 CH12: 2.0 No serious AEs (2012)32 treatment as part CH11: 0.58 of Phase 1 Follow- See Bennet 2016 Sex: all Female NP01: 1.6 AE: Manuscript on (see Bennett Years after first Surface irritation 102 for full results) treatment: Full-field light sensitivity Sprained ankle Subset 102 for CH12: 2.1 CH12: no improv Headache safety 2/3 pts F/U until CH11: 2.3 CH11: increased day 180 NP01: 3.7 NP01: increased ©Institute for Clinical and Economic Review, 2018 Page 98 Final Evidence Report – Voretigene Neparvovec Author & Year Study Design and Interventions (n) Major Inclusion & Patient Key Outcomes Harms of Publication Duration of & Dosing Exclusion Criteria Characteristics (Trial Name) Follow-up Schedule Quality rating 1 pt final data is Vector present in average 180 & 365 Visual acuity in fMRI total visual cortex tears and blood logMAR changes with high contract after injection up (higher=worse) at 30/60 days post injection to day 3 pre-contral (mm2) inject: CH12: 1729/8110 Some transient CH12: 2.6 CH11: 9658/13366 immune CH11: 0.64 NP01: 241/784 response, one pt NP01: 1.83 with high Pupillary light reflex background improved after treatment forming units of contralateral eye Mobility test: CH12 no data reported CH11 10 lux (p=0.015) NP01 5 lux (p=0.005) Simonelli, See Maguire 2008 See Maguire 2008 See Maguire 2008 See Maguire Objective measures No serious Molecular 2008 Pupillary light reflex at 1.5 adverse event Therapy Follow-up: 1.5 yrs years showed sustained through 1.5 years (2010)86 improvement in velocity and amplitude in treated Mild increase in Manuscript vs. untreated eye short run serum neutralizing Follow-up to Reductions in ocular antibodies to Maguire 2008 motility maintained (and AAV2 in 2 lowered) at 1.5 yrs patients –normal compared to baseline in at 1 year both eyes ©Institute for Clinical and Economic Review, 2018 Page 99 Final Evidence Report – Voretigene Neparvovec Author & Year Study Design and Interventions (n) Major Inclusion & Patient Key Outcomes Harms of Publication Duration of & Dosing Exclusion Criteria Characteristics (Trial Name) Follow-up Schedule Quality rating Prior macular Electroretinography: no hole (subject 2) change from baseline to 1.5 not expanded years (response flash) Subjective measures Visual acuity same at 1.5 years (see Maguire, 2008) Pts 1&3 flipped best eye after treatment Mobility testing showed slight continuous improvement between day 30 and 1.5 years. Maguire, JAMA Safety Voretigene Inclusion Ages: 19, 26, 26 Reductions in Nystagmus Patient 2: (2008)94 neparvovec (VN) LCA diagnosis; RPE65 All pts: right eye Freq after injection (Hz), Outer lamellar N=3 or AAV2- mut; age 8-27; visual treated right/left eye cyst in fovea Manuscript hRPE65v2 acuity ≤ 20/200 or Female, N (%): 2 1) 1.2/1.2 noted on day 5 LCA2 visual field less than 20 (66.7) (difference -0.8/-0.8) after injection; Pilot/Safety 1) Subject 1 degrees 2) 0.9/0.9 Macular hole 2 Follow-up: 1 2) Subject 2 Nystagmus Freq (difference -0.1/-0.1) weeks post- month 3) Subject 3 Exclusion before injection 3) 1.4/1.1 surgery Participation in study (Hz), right/left (difference -0.1/-0.27) Single injection- of investigational drug eye Patient 1: subretinal or ocular surgery in 1) 2.0/2.0 Pupillary Light Reflex Tear showed past 6 months; 2) 1.0/1.0 All pts had greater AAV2 on Day 1 Dose: 1.5x1010 vg conditions that 3) 1.5/1.37 response to light stimulus; post surgery, no in 150 µl of preclude endpoint evidence of ©Institute for Clinical and Economic Review, 2018 Page 100 Final Evidence Report – Voretigene Neparvovec Author & Year Study Design and Interventions (n) Major Inclusion & Patient Key Outcomes Harms of Publication Duration of & Dosing Exclusion Criteria Characteristics (Trial Name) Follow-up Schedule Quality rating phosphate- interpretation; lack of Visual Acuity each eye ~3x as sensitive to dissemination, no buffered saline sufficient viable retinal (logMAR), light humoral immune supplemented cells by OCT; right/left eye response with Pluronic F-68 complicating systemic 1) 2.0/1.72 Visual Acuity, logMAR NF Prill Poloxamer diseases or abnormal 2) 2.0/1.04 right/left eye (difference in Patient 2: 188 and chicken β baseline labs; 3) 1.5/1.05 treated eye) neutralizing actin (CBA) sensitivity to meds for 1) 1.72/1.74 (0.28) antibody titers promoter surg; presence of Visual Field 2) 1.55/1.04 (0.45) increased post- neutralizing antibodies (degrees), 3) 1.16/1.03 (0.34) surgery but to AAV2 above 1:1000 right/left eye diminished over 1) 41/36 Visual Field, degrees time 2) 62/55 (difference in treated eye) 3) 147/203 1) 177/26 (136/-10) 2) 213/75 (151/20) 3) 210/160 (63/-43) ©Institute for Clinical and Economic Review, 2018 Page 101 Final Evidence Report – Voretigene Neparvovec Appendix G. Public Comments This section includes a summary of the public comments prepared for the Midwest CEPAC Public Meeting on January 25, 2018 in Kansas City, MO. This summary was prepared by those who delivered the public comment at the meeting. A video recording of all comments can be found here, beginning at minute 1:43:30. Thomas Ciulla, MD, Spark Therapeutics Ophthalmology Medical Lead Prior to the approval of LUXTURNA™ (voretigene neparvovec-rzyl), nearly all patients with RPE65- mutation associated retinal dystrophy progressed to complete blindness, as there were no pharmacologic treatment options. To help ensure that all eligible patients have access to this novel treatment, Spark has priced LUXTURNA responsibly, has created outcomes-based payer rebate arrangements that recognize its longer-term efficacy, and has explored alternative payment options that would allow for installment payments. Spark also has an interest in ensuring ICER’s report on LUXTURNA does not limit patient access to the therapy due to a base-case economic model that makes headlines. In fact, ICER’s report demonstrates that LUXTURNA is cost effective for many patients when ICER uses health utilities that reflect the rapid decline in quality-of-life experienced by patients as they progress to blindness, uses indirect costs that address the high societal impact of blindness, considers a lifetime treatment effect, and/or uses a higher cost per QALY threshold for an innovative therapy. Unfortunately, payers who do not review the report in detail could easily miss these points, and inappropriately limit access to LUXTURNA. Consequently, the results from ICER’s own sensitivity analyses that demonstrate the cost- effectiveness of LUXTURNA must be highlighted more prominently in the report, so that readers more fully understand the value of the therapy, and so that no patient in need is denied access. ©Institute for Clinical and Economic Review, 2018 Page 102 Final Evidence Report – Voretigene Neparvovec Appendix H. Conflict of Interest Disclosures Table H1 through H3 contain conflict of interest (COI) disclosures for all participants at the January 25, 2018 public meeting of the Midwest CEPAC. Table H1. ICER Staff and Consultant COI Disclosures Name Organization Disclosure • Reiner Banken, MD, Msc ICER None • Josh Carlson, PhD, MPH University of Washington None • Geri Cramer, BSN, MBA ICER None • Sonya Khan, MPH ICER None • Molly Morgan ICER None Steve Pearson, MD, MSc ICER None David Rind, MD, MSc ICER None • Marita Zimmerman, MPH, PhD University of Washington None ©Institute for Clinical and Economic Review, 2018 Page 103 Final Evidence Report – Voretigene Neparvovec Table H2. CEPAC/CTAF Panel Member COI Disclosures Name Organization Disclosure Eric Armbrecht, PhD St Louis University Center for Health Outcomes * Research Aaron Carroll, MD, MS Indiana University School of Medicine * Reem Mustafa, MD, MPH, PhD University of Kansas Medical Center * Harold Pollack, PhD University of Chicago * Rachel Sachs, JD, MPH Washington University in St. Louis * Shumei Yun, MD, PhD MO Department of Health and Senior Services * Paul Heidenreich, MD, MS Stanford University * Rob Aseltine, PhD University of Connecticut * Austin Frakt, PhD Department of Veterans Affairs * Claudio Gualtieri, JD AARP * Brian O’Sullivan, MD Giesel School of Medicine at Dartmouth College * Rev. Edward Westrick, MA American Diabetes Association * *No conflicts of interest to disclose, defined as more than $10,000 in health care company stock or more than $5,000 in honoraria or consultancies during the previous year from relevant health care manufacturers or insurers. ©Institute for Clinical and Economic Review, 2018 Page 104 Final Evidence Report – Voretigene Neparvovec Table H3. Policy Roundtable Participant Disclosures Name Affiliation Disclosure Katelyn Corey VN Trial Participant Receipt or potential receipt of anything of monetary value, including but not limited to, salary or other payments for services such as consulting fees or honoraria in excess of $5,000. Patrick Gleason Prime Therapeutics Prime Therapeutics employee Christine Kay Vitreo Retinal Associates None Janet LaBreck Former Commissioner, Spark Consultant Rehabilitation Services Administration Bill Martin Express Scripts, Inc ESI employee ©Institute for Clinical and Economic Review, 2018 Page 105 Final Evidence Report – Voretigene Neparvovec