X-RAY STUDY OF ADVANCED TUBERCULOSIS OF 
THE LUNGS WITH AUTOPSIES 
THE DEGREES OF DENSITY OF GENERAL HOSPITAL No. 17* 
By JOSEPH WALSH, M.D. 
Philadelphia 
JAMES W. WOOD, M.D. 
Chester, Pa. 
AND 
CAPTAIN JAMES C. THOMPSON, M.C. 
Spartanburg, S. C. 
From our teachings we had the following general ideas in regard to 
the significance of different degrees of density on x-ray plates of tuber- 
culosis of the lungs: first, that the greater the density the greater the 
amount of fibrous tissue in the lesion; second, that a thickened pleura is 
capable of showing very marked density; third, that massive caseation 
shows less density than fibrous tissue and even less than many thickened 
pleurae, in fact often shows only mild semi-translucent snowstorm effects; 
fourth, that scattered tubercles are visible practically only in the greater 
evidence of the normal lung detail; fifth, that normal lung detail is pro- 
duced by bronchial tubes and blood vessels; sixth, that annular figures 
signify cavities and their absence the absence of cavities. 
Confidently we read our plates with those ideas in mind. Six autop- 
sies occurred and we compared the carefully sectioned lung with our read- 
ing. It failed to correspond in so many particulars that we set about a 
special study and proceeded to read our x-ray plates not from previous 
instruction but from the comparison of the autopsies. 
We sectioned the lungs most carefully, compared the plate with every 
part and tried to determine what each shadow meant. In addition, in 
order to discover how much of a shadow different normal and pathological 
tissues produced, we x-rayed a number of tissues separately and created 
a number of conditions in the way of cavities containing both air and 
fluid, localized caseous consolidations transferred from other lungs, watery, 
serous, blood and purulent effusions covering the lungs, and conditions 
representing thickened pleura over both human and pigs' lungs. We 
studied the lungs both within and without their thoracic cage, with and 
apart from the heart, in cut sections and under every circumstance which 
our imaginations could devise. Our plates were all stereoscopic. 
We found the study of pigs' lungs particularly instructive from the 
fact that the bronchial tubes are larger and extend nearer to the periphery, 
thus making them easier of recognition by the x-ray. 
* Published by permission of the Surgeon General, U. S. A. 2 
X-RAY STUDY OF ADVANCED TUBERCULOSIS OF THE LUNGS 
We made a number of x-ray pictures of the following structures: 
1. Normal pericardium 
2. Two thicknesses of pericardium 
3. Small piece of collapsed lung with tubercles 
4. Small piece of collapsed emphysematous lung with tubercles 
5. Small piece of collapsed lung with tubercles, the whole covered with 
pericardium 
6. A piece of lung with ordinary visceral pleura on side so that the 
pleura would be seen in optical section 
7. Piece of lung with small cavity; cavity was cut in half, cut side 
uppermost and penetrating about one-half the lung tissue 
8. Piece of blood vessel about the size of the adult carotid and one- 
tenth of an inch deep in optical section 
9. Piece of blood vessel empty and another filled with blood in longitu- 
dinal section 
10. Piece of adult trachea about one-eighth of an inch deep in optical 
section 
11. Piece of bronchus in longitudinal section 
12. Longitudinal section of dividing bronchus at hilum filled with 
blood. 
We learned that: 
First. Even a dense thick fibrous membrane, like pericardium, 
interferes very little with the ray and casts almost no shadow. This is 
true even when it is doubled. We were not surprised to find, therefore, 
that thickened pleura evidently apparent at autopsy could not be detected 
on the plate. Pathological conditions in the lungs are visible by their 
shadows only when the air content in this area is much less or much 
greater than that neighboring. We can imagine a pleura sufficiently 
thickened over one lobe to so compress it as to diminish the general air 
content and thus produce a shadow, but this would require the pleura 
to be probably a quarter of an inch or more in thickness which never 
occurs except over a massive lesion when the thickened pleura would be 
lost in the lesion itself. 
Second. Tubercles to the size of a pea in collapsed lung tissue are 
not differentiated, the whole mass appearing of equal density. Therefore, 
it is not the caseous debris per se which makes the shadow, but simply 
the comparison of the air content. 
Third. An artery the size and thickness of the carotid casts so little 
shadow even when filled with blood that it is likely the blood vessels 
per se have little or nothing to do with the hilum shadow or normal lung 
detail. The hilum shows up not on account of specific structures, but 
merely on account of the lack of air content. 
Fourth. The trachea and bronchial tubes cast a slightly more dense 
shadow than blood vessels, and from the study of bronchial tubes stand- 
ing alone and in the pigs' lungs, we believe that they are entirely respon- 
sible for normal lung detail. 
Fifth. A cavity in the lung becomes evident not on account of the 
density or thickness of its wall, but on account of the air in it, and the JOSEPH WALSH, M.D., JAMES W. WOOD, M.D., CAPT. JAMES C. THOMPSON 
3 
air usually so rarefies the tissue before and behind as almost to obliterate 
it. The cavity of the trachea, for instance, so modifies the dense tissue 
of the sternum and spinal column between which it lies that it is seen 
through them. As a consequence, it is usually impossible to diagnose 
whether the condition before or behind a cavity in the lung is normal 
or pathological. On account of this very rarefaction, however, a cavity 
sometimes intensifies special things, which are smaller in extent than the 
cavity itself, so that we feel confident, though we never had the opportu- 
nity to try the experiment, if a bladder filled with air were placed on the 
outside of the body in front of an imbedded metallic substance, or possibly 
a kidney or bladder stone, it would rarefy the tissue about the stone so 
as to make it more apparent. Since we wrote this sentence, we have 
learned that others have done this with success. In addition, no matter 
how thick the wall, a cavity never shows contour and always appears as a 
ring. 
To learn whether or not the inflammatory elements in the wall of a 
cavity are of importance in the production of the wall seen on the x-ray 
plate, an x-ray was taken of intestinal ulcers: 
1. An unopened piece of intestine with ulcer. The plate showed 
only the outline of the intestine with practically no indication of the ulcer. 
2. Piece of intestine opened with ulcer uppermost. This showed the 
ulcer in ill defined outline. 
3. Piece of intestine opened with ulcer covered with barium. This 
showed the barium covering the ulcer as well as the barium accidentally 
dropped over other places but did not outline the ulcer. 
The same three pieces of intestine were also covered by one-quarter 
inch of rectus muscle. All showed the intestine in faint outline and the 
last the barium in the intestine. In other words, the inflammatory ele- 
ments per se are not the cause of the cavity wall shadow. 
Artificial conditions were also produced in the following ways: 
1. A pig's bladder was blown up, one inch and a half in diameter with 
air and introduced into the center of a pig's lung. In spite of the thick 
walls of the bladder and the fact that it was not blown up to a great degree 
allowing the walls to fall in folds, the plate showed simply the typical 
appearance of cavity with ordinary walls. In other words, in spite of 
the muscular and connective tissue in the bladder walls, it was the air con- 
tent which was particularly visible, rather than the wall itself. Imme- 
diately below where the bladder was introduced there was a triangular 
space made in the lung tissue. In spite of the fact that this had no 
definite wall except above, where the bladder crossed it, this cavity showed 
up quite as well as the bladder cavity. 
2. A small three-quarter inch square of collapsed human lung contain- 
ing scattered tubercles was introduced into the middle of a pig's lung. 
This showed up as a single mass, the tubercles not being differentiated. 
The density was homogeneous and faint, comparing about with single rib 
or fibrosis density. It was sharply defined and showed the exact shape 
of the piece inserted. 
3. A mass of caseation one inch in diameter was introduced into the 4 
X-RAY STUDY OF ADVANCED TUBERCULOSIS OF THE LUNGS 
center of a pig's lung and showed slightly greater density than the preced- 
ing. It was about as dense as heart density, though slightly less homo- 
geneous. The edges were not so well defined and shaded off into normal 
lung. In other words, though caseation is more dense than collapsed 
lung and showed it on the plate, the margin of the caseated mass, due 
probably to slight differences in the pictures, was less clear cut. 
4. Pigs' lungs were infiltrated with barium sulphate both by insufflation 
after death and by attaching a bag filled with barium over the snout 
shortly before death, with the idea that it would be easier to trace the 
bronchial tubes. Subsequent studies proved these experiments unneces- 
sary since the bronchial tube in the pig is so evident throughout its course. 
Its shadow is manifested on the plate as a line of lessened density varying 
in size with the calibre of the tube, bounded by two white lines. We do 
not agree with the description that a bronchus is indicated by two flat 
bands of great density separated by an area of slight density, since slight 
density would convey the idea of greater density than normal when in 
reality it is less. On account of the air in the bronchial tube, it is often 
noticeable that when a bronchial tube passes in front of or behind an 
object the density of which corresponds with or is greater than the 
boundary line of the bronchus, the bronchus is still traceable as an area 
of lessened density, but the boundary lines are effaced. It is interesting 
to note that the barium in both cases confined itself almost entirely to the 
upper portions of the lungs, confirming the fact described by others that 
inhaled particles are carried naturally to the apices. 
5. The bladders of three pigs were filled, one with water, one with blood, 
and one with air, and x-rayed. The density of the water and blood was 
exactly the same. The bladder filled with air showed as a clear space 
bounded by a narrow white line. 
6. The bladder filled with water was flattened out to one inch in thick- 
ness, placed before a lung, and x-rayed. It showed the density of heart 
with lung detail visible behind it. Ordinarily a pleural effusion fails to 
show lung detail behind it, because, in addition to casting its own shadow, 
it compresses the lung, thus obliterating lung detail and increasing the 
shadow. The faint shadow immediately above a pleural effusion, com- 
monly stated to be due to thickened pleura, we believe due to upward 
compression of the lung. 
7. A bladder filled with pus was placed in front of the left lung, and a 
bladder filled with serum behind the right lung, and the whole was sur- 
rounded by a layer of skin and fat three-fourths of an inch thick and 
x-rayed. The pus showed slightly denser than the serum, and not quite so 
dense as the heart. 
From these studies, we thought we determined first that lung detail 
so called is due practically entirely to bronchial tubes, for if due to blood 
vessels it would be obliterated by the bladder filled with fluid. Moreover, 
their definition is due less to their walls than to the rarefaction of the 
tissue above and below by the air content, in as much as a sinus without 
walls through normal lung tissue shows up practically as well. We do 
not dispute the anatomical fact that the blood vessels accompany the JOSEPH WALSH, M.D., JAMES W. WOOD, M.D., CAPT. JAMES C. THOMPSON 
5 
bronchial tubes, but only that they do not participate in the production 
of normal lung detail shadow. 
In general it is evident that the hilum shadow is made up of bronchial 
tubes, bronchial glands, arteries, veins and the massive connective 
tissue at the root of the lung, but it is impossible to decide how much 
is due to each structure, since the shadow is the result of the absence of 
air in comparison with the lung rather than the presence of individual 
structures. The hilum shadow more than anything else demonstrates 
the necessity of having the plates absolutely uniform as to time expo- 
sure, distance, milliamperage, etc., for on some plates which appear 
practically perfect the bronchial tubes passing through the hilum are evi- 
dently shown, while on others apparently equally perfect they are not 
shown at all, and from this we judge that correspondingly large features 
might be missed in the lung due to similar unconscious differences in 
penetration. 
It appears that when the lesion extends from the anterior to the pos- 
terior aspect of the lung, the plate will show it, but if the lesion is reason- 
ably localized with considerable normal tissue before or behind it, or 
both, it may be entirely lost. 
The x-ray shadow is produced neither by caseation, nor fibrosis per se, 
but is merely a matter of more or less air than normal. Any sort of a 
consolidation of the lung tissue produces shadow, but the shadow is 
invisible if the consolidation is less than a certain size, probably I cm. 
in diameter, though multiple masses of even less size may produce shadow 
if these masses overlap, thus producing an optical mass of greater size. 
Cohn1 and Sewall and Childs2 assert that new tubercles cast no shadow, 
and whatever shadow is visible is due to the vascular engorgement about 
them. Our experiments prove that anything lessening the amount of air 
casts a shadow, but the shadow of new tubercles may be, and almost 
always is, too small to be seen. The vascular engorgement may be 
sufficient to make the shadow visible. From the numerous plates of lobar 
and broncho-pneumonias recently made, we know that congestion of the 
lung does produce a shadow, though the amount of congestion necessary 
has not been determined. From our studies we believe that a few widely 
scattered small tubercles are invisible, but that a large number become 
visible on account of their overlapping in optical section. The assertion 
of Sewall and Childs that lung detail is entirely and of Baetjer3 that it is 
mostly due to blood vessels does not seem justified. 
All of our autopsied cases showed both lungs so infiltrated that we had 
no opportunity for, the study of the early typical apical lesions on which 
so much of Dunham's excellent work has been done; in the study of our 
advanced lesions, however, we were forced to disagree with some of the 
prevailing ideas relative to the significance of various densities, and on 
this account have formulated our findings into the following series. On 
account of the unavoidable differences in technique of different plates, 
we thought it worth while comparing these densities when possible with 
something on the plate, so that in case the plate were over or under 
exposed a comparison would still be at hand. 6 
X-RAY STUDY OF ADVANCED TUBERCULOSIS OF THE LUNGS 
The following is the working scale we formulated: 
o° Pneumothorax. Perfect clearness. 
i° Compensatory emphysema or cavity, the latter being usually 
more homogeneous. 
2° Normal lung. 
3° Scattered tubercles. Slight mottling, usually seen neighboring 
an extensive lesion. 
4° Fibrosis. Density of moderate degree, simulating the density 
of a single rib near the base of the lung. 
5° Numerous scattered tubercles or caseous broncho-pneumonia. 
Density of greater degree than fibrosis in flaky or cottony 
areas resembling a thick snowstorm. About the density of 
crossing ribs near the base, but not so homogeneous. 
6° Massive caseation. Homogeneous density like frosted glass. 
Simulates heart density. Fluid cannot be distinguished from 
caseation except by location and demarkation. 
7° Calcification. Bright speck, greater than the density represent- 
ing sternum, heart and spine over one another. 
In spite of all this study and differentiation, however, our most recent 
reading of the x-ray plates of which we have autopsy protocols, after 
sufficient time had elapsed for us to forget the autopsy findings, was only 
50 per cent, perfect, 25 per cent, fair and 25 per cent, wrong. These 
failures were not, of course, failures in the diagnosis of massive lesions, 
but in the differentiation of fibrous tissue from caseation, the diagnosis 
of occasional cavities and the presence or absence of a few tubercles. 
The majority of cavities show annular figures and the majority of annular 
figures signify cavities, but the autopsy sometimes showed no cavity where 
there was a figure, and sometimes a cavity where there was none, and the 
number of them was sufficiently great to make the diagnosis very uncertain. 
In other words, we are still far from perfection; in fact, our x-ray readings 
are yet not as satisfactory as the original diagnosis from physical signs. 
We regret that we did not see the article by Sampson, Heise and 
Brown4 before our autopsies, since their explanation of the annular 
shadows as localized interlobar pneumothoraces seems as if it might 
cover a certain number of cases. As far as we know such local zed 
pneumothoraces have never been found at autopsy, and their presence 
or absence will require special sectioning of the lung with these in mind. 
It has been commonly stated (Barjon5, Honeii, Brown and others) 
that though the x-ray usually fails to show a small lesion at the apex as 
definitely as physical signs, any lesion which it is capable of showing it 
usually demonstrates as larger than the physical signs outline it, and the 
larger the lesion the more true this is. This statement has been repeated 
such a number of times that most clinicians believe it, yet we have never 
seen it based on the scientific grounds of comparison. If only the super- 
ficial examination of haste and inexperience is taken into consideration 
the statement is undoubtedly true; that it is far from true in comparison 
with the specialist's examination is evidenced by our autopsy findings com- 
pared with the physical diagnosis and x-ray findings. JOSEPH WALSH, M.D., JAMES W. WOOD, M.D., CAPT. JAMES C. THOMPSON 
7 
The following, table shows this comparison: 
Perfect Fair Wrong 
Physical Diagnosis62 per cent. 28 per cent. 10 per cent. 
X-ray Diagnosis43 per cent. 19 per cent. 38 per cent. 
We wish to reiterate that those wrong diagnoses were wrong only in 
details not in general involvement. 
The following is the comparison of the autopsy, physical diagnosis and 
x-ray findings on which the previous table is based. All of the physical 
examinations were made by one of us, and the plates were read by all 
three, both individually and in concert, as well as by three x-ray experts 
visiting the hospital as inspectors, and we have taken the best x-ray 
readings not the worst. 
Comparison of Autopsy, Physical Diagnosis and X-ray Findings 
Case No. 197 
Autopsy 
Physical Diagnosis 
X-ray Diagnosis 
Right Lung 
Upper lobe. One inch 
fibrous tissue then two large 
cavities and the rest of the 
upper lobe caseo-fibrosis. 
Pleura thickened. 
Large cavity surrounded by 
fibrosis. 
Plate taken after death. 
Large cavity surrounded 
by caseo-fibrosis. 
Middle lobe. Compensa- 
tory emphysema. Pleura 
thickened. 
Compensatory emphysema. 
Caseo-fibrosis. 
Lower lobe. Large cavity 
at apex surrounded by caseo- 
fibrosis. Two cavities just 
below. Rest compensatory 
emphysema. About 50 per 
cent, air bearing. Pleura 
thickened. 
Fibrosis at apex to fifth rib. 
Few scattered tubercles 
immediately below. From 
angle of scapula down 
compensatory emphyse- 
ma. 
Upper part caseo-fibrosis; 
rest normal. No ring. 
Left Lung 
Upper lobe. Entirely 
hollowed out by cavity. 
Pleura thickened. 
Entire lobe large cavity. 
Entire lobe large cavity. 
Lower lobe. Cavity one 
inch in diameter at apex. 
Caseation for an inch below. 
Rest of lobe numerous scat- 
tered tubercles of pea size. 
Pleura thickened. 
Cavity at apex with casea- 
tion to fifth rib with 
numerous scattered tu- 
bercles below. 
Dense shadow over entire 
lower half of lung diag- 
nosed massive caseation, 
though the autopsy 
showed numerous scat- 
tered tubercles. The cav- 
ity was not indicated. 
Case No. 395 
Right Chest 
Pneumothorax. 
Pneumothorax. 
Pneumothorax. 
Left Lung 
Upper lobe. Normal to 
third rib. Few scattered 
tubercles to fifth rib. Mass 
of coalesced tubercles to 
base. Pleura normal. 
Caseo-fibrosis to third rib. 
Scattered tubercles to fifth 
rib, massive caseation 
between fifth and seventh 
ribs. 
Normal to fourth rib, case- 
ous broncho-pneumonia 
to base. 
Lower lobe. Marked em- 
physema. Pleura normal. 
Emphysema. 
Large cavity in center of 
lobe. Fibrosis at apex. 
Another reading: Two cav- 
ities in center of lobe. 
Fibrosis at apex. 8 
X-RAY STUDY OF ADVANCED TUBERCULOSIS OF THE LUNGS 
Case No. 296 
Autopsy 
Physical Diagnosis 
X-ray Diagnosis 
Right Lung 
Upper lobe. Three large 
cavities one outside midclav- 
icular line with caseation. No 
fibrosis. No air bearing tissue. 
Pleura slightly thickened. 
Two large cavities sur- 
rounded by fibrosis. 
Caseation outside midcla- 
vicular line, fibrosis inside 
surrounding cavities. 
Another reading: Complete- 
ly fibrous. 
Middle lobe. Number o* 
large scattered tubercles- 
Pleura normal. Slight em- 
physema. About 70 per 
cent, air bearing. 
Compensatory emphysema. 
More recent and not so 
dense infiltration. An- 
other reading: Fibrosis 
and caseation. Another 
reading: Slight fibrosis. 
Lower lobe. Apex two 
large coalesced tubercles 
each one-half inch in diame- 
ter. A few scattered tuber- 
cles with emphysema below. 
About 85 per cent, air bear- 
ing. Pleura normal. 
Cavity at apex. Compen- 
satory emphysema below. 
Cavity between fifth and 
seventh rib, also at apex 
surrounded by fibrosis. 
Another reading: Normal. 
Left Lung 
Upper lobe. Four large 
cavities surrounded by case- 
ation. No air bearing tis 
sue. No fibrosis. Pleura- 
slightly thickened. 
Four large cavities sur- 
rounded by caseation. 
Four cavities surrounded by 
fibrosis. The fibrosis was 
diagnosed by several ex- 
perts. 
Lower lobe. Upper halt 
large and small cavities. 
Rest of lobe closely approxi- 
mated tubercles; 15 percent, 
air bearing. Pleura normal. 
Compensatory emphysema 
with scattered tubercles. 
Massive fibrosis with cavity 
towards apex. Another 
reading: Normal. 
Case No. 439 
Right Lung 
Upper lobe. Large cavity 
at apex extending to second 
interspace; three smaller 
cavities 1.5 to 2 cm. in diam- 
eter immediately below; 
rest of lobe caseo-fibro- 
sis (fibrosis predominating). 
No air bearing tissue. Pleura 
thickened. 
Cavity at apex and between 
third and fourth rib; rest 
of lobe caseous. 
Cavity at apex; five small 
cavities immediately be- 
low, only three of which 
could be found at autopsy, 
and these three were smal- 
ler than the x-ray ap- 
peared to show them; 
rest of lobe fibro-caseous, 
though on different plates 
the shadow varied, so 
that it was impossible to 
be sure whether fibrosis 
or caseation predominat- 
ed, yet the shadow was al- 
ways less than the heart 
shadow. The density va- 
ried from a snow-storm 
effect to a homogeneous 
density. A rather dense 
line at the base of the 
lobe, thought by some to 
be thickened interlobar 
pleura, proved to be dense 
fibro-caseation within the 
lobe. There was no spe- 
cial thickening of the 
pleura between the upper 
and middle lobes, though 
it was obliterated. JOSEPH WALSH, M.D., JAMES W. WOOD, M.D., CAPT. JAMES C. THOMPSON 
9 
Autopsy 
Physical Diagnosis 
X-ray Diagnosis 
Middle lobe. Few scat- 
tered tubercles; 97 per cent, 
air bearing. Pleura normal. 
Normal. 
Normal. 
Lower lobe. Several mas- 
ses of caseation at apex, 1 
to 3 cm. in size. At base 
posteriorly near spinal col- 
umn recent half gangrenous 
pneumonic process; rest of 
lobe contains a very few scat- 
tered tubercles; 85 per cent, 
air bearing. Pleura normal. 
Massive caseation at apex 
with scattered tubercles 
below. 
Normal. 
Left Lung 
Upper lobe. Practically 
entirely hollowed out by 
two large cavities; the upper 
extends from the apex to the 
base posteriorly and to the 
second interspace anteriorly 
outside midclavicular line; 
the second begins about 
the clavicle inside the mid- 
clavicular line and extends 
almost to base anteriorly. 
Pleura thickened. 
Large cavity at apex with 
small cavities below. 
Cavities with fibro-caseation 
below. 
Lower lobe. Cavity at 
apex 2 cm. in diameter filled 
with grumous-sanguinolent 
fluid. Another cavity 3 
cm. in diameter toward mid- 
dle of lobe, filled with simi- 
lar fluid. These cavities are 
surrounded by fibro-casea- 
tion. The lower half of the 
lobe shows numerous scat- 
tered tubercles. The hilum 
shows eight hard anthracot- 
ic glands. On the bronchial 
trunk extending from the 
hilum to the base are num- 
erous nodules of apparently 
anthracotic-lymphoid tissue 
varying in size from 2 to 8 
rrm. Pleura normal. 
Massive caseation at apex, 
scattered large tubercles 
below. 
Upper part fibro-caseous; 
lower part normal. The 
anthracotic glands were 
not visible. 
Case No. 495 
Right Lung 
Upper lobe. Apex for 
three-quarters of an inch 
shows considerable air bear- 
ing tissue with few scattered 
tubercles, then a cavity 3 
cm. in diameter extending 
from the clavicle to the sec- 
ond rib in the midclavicular 
line filled with thick yellow- 
ish, greenish, tenacious ma- 
terial. Small cavity at 
outer margin of lung under 
clavicle 1.5 cm. in diameter; 
rest of lobe shows numerous 
scattered tubercles, varying 
in size from 2 to 15 mm. 
About 40 per cent, is air 
bearing. Pleura normal. 
Caseation to second rib, 
numerous scattered tuber- 
cles below. 
Fibrosis upper half, casea- 
tion lower half. Cavity 
near junction of clavicle 
and sternum. (There 
was no cavity at this 
point. The cavity found 
in the middle of the lobe 
had probably never been 
evacuated). Another 
reading: Cavity at apex, 
fibrosis below, caseation 
at base. (As is evident 
this apparent cavity 
proved to be almost nor- 
mal air bearing tissue in 
an unexpected situation.) 10 
X-RAY STUDY OF ADVANCED TUBERCULOSIS OF THE LUNGS 
Autopsy 
Physical Diagnosis 
X-ray Diagnosis 
Middle lobe. Scattered 
tubercles. Slight emphy- 
sema; 80 per cent, air bear- 
ing. Pleura normal. 
Few scattered tubercles 
with emphysema. 
Emphysema and scattered 
tubercles. 
Lower lobe. Scattered 
tubercles. Slight emphy- 
sema; 70 per cent, air bear- 
ing. Pleura normal. 
Not recorded. 
Emphysema and scattered 
tubercles. 
Left Lung 
Upper lobe. Outside mid- 
clavicular line fibro-casea- 
tion and many cavities 
from 1 to 2 cm. in diam- 
eter. Inside midclavicular 
line caseous-pneumonia an- 
teriorly and air bearing tis- 
sue posteriorly. About 10 
per cent, air bearing. Pleura 
obliterated. 
Caseation with cavitation to 
fourth rib, pea to marble 
sized tubercles below. 
Cavity inside midclavicular 
line between first and sec- 
ond rib; rest of lobe 
caseation. Another read- 
ing: Cavity at apex and 
at sterno-clavicular junc- 
tion; rest of lobe casea- 
tion. 
Lower lobe. Upper half- 
(apex to angle of scapula) 
fibro-caseation with many 
cavities 1 to 3 cm. in diam- 
eter. Lower half caseous- 
broncho-pneumonic patches 
with small amount of air 
bearing tissue between them; 
20 per cent, of lobe air bear- 
ing. Pleura obliterated. 
Not recorded. 
Upper half caseation, lower 
half scattered tubercles. 
In this case the autopsy 
showed a tuberculous ab- 
scess of the ninth to twelfth 
vertebrae without clinical 
symptoms or physical signs 
which the x-ray diagnosed. 
Our investigations like those of other writers6 have taught us that 
conclusions should not be drawn from one plate, nor even from one 
autopsy comparison, because of the possibility of plate variation. Our 
current was not very uniform and hence we probably found more than the 
ordinary variation, but it has proven that the greatest caro must be 
exercised before judgment is passed. 
• We had only four cases out of over 200 within the four months of the 
study in which the x-ray * showed something undiscovered by physical 
diagnosis, the last two came to autopsy, the first two were determined 
clinically only. In the first two cases, Nos. 366 and 262, there is an area ap- 
parently of fibrosis under the right scapula in such a situation that it was 
easily passed over on physical examination unless particularly sought for. 
In one it represented almost surely the healed remains of a long passed 
abscess; in the other it probably represented the same, but there was 
question whether or not it was entirely cured. These two conditions were 
by measurement one and one-half inches in diameter, were located be- 
tween the fifth and eighth rib posteriorly, and centered about two inches 
from the spine. It is possible that they might have been interlobar instead JOSEPH WALSH, M.D., JAMES W. WOOD, M.D., CAPT. JAMES C. THOMPSON 
II 
of intrapulmonary abscesses. After a study of the literature7 and seeing 
some of Webb's, Bullock's, Marshak's and Dunham's plates we are 
inclined to wonder if a certain number of the abscesses diagnosed in the 
past as pulmonary may not have been interlobar. Yet in 500 autopsies 
at the Phipps Institute none was ever found. 
The third case in which the x-ray showed something undiscovered by 
physical signs was one of advanced tuberculosis of both lungs and tuber- 
culosis of the ninth to twelfth vertebrae. The latter gave neither symp- 
toms nor signs, but was very evident on the x-ray. 
The fourth case in which the x-ray showed something undiscovered 
by physical signs is rather important in that it gives us a possible explana- 
tion of cracked pot sound. All sorts of explanations have been offered to 
explain cracked pot over cavities. The x-ray finding in Case No. 296, 
subsequently proven by autopsy, showed a double cavity, one super- 
imposed on the other in two situations where cracked pot sometimes 
occurred. The x-ray shows a similar condition in Case No. 17, below a 
cracked pot sound. 
That the x-ray cannot always be relied on appeared in Case No. 17, 
in which the physical diagnosis shows a lesion from the fourth rib and 
sixth dorsal spine up with cavitation near apex, while the x-ray shows it 
to be only above a line extending from the junction of the third rib with 
the sternum to the outer junction of the clavicle with the first rib. In 
this case the physical findings are very much greater than the x-ray find- 
ings, and are explained by the fact that the x-ray has shown only the 
condition extending from front to back and not the superficial anterior 
lesion with normal or even emphysematous lung behind it. 
Of the x-ray work so far done, Dunham's8 is the most satisfactory and 
apparently the most accurate. It was a pleasure after making a number 
of comparisons to discover how many of our conclusions agreed. Whether 
or not his fan-shaped densities and prolonged or thickened linear markings 
are peculiar to tuberculosis, we are unable to state since all of our autop- 
sies showed advanced lesions on both sides. In addition all of our lesions 
were active so that we could not follow him into inactivity. 
In our readings we have followed the order and general method of 
Dunham as described in the American Review of Tuberculosis, and 
found it satisfactory: I plate (whether good, over or under exposed, etc.), 
II bones, III muscles, diaphragm, heart, IV trachea, V hilum, VI right 
lung, VII left lung. 
CONCLUSIONS 
1. Much of the teaching and speculation in regard to x-ray shadows in 
connection with tuberculosis of the lungs is far from correct as is proven 
by autopsy. 
2. The principal inaccuracy is that fibrous tissue shows more dense 
than caseation when the opposite is true; in fact, we found not only 
massive caseation, but even scattered tubercles, when they were numerous, 
to be denser than fibrosis. 
3. In the case of caseation or any other consolidation, like pneumonia, 12 
X-RAY STUDY OF ADVANCED TUBERCULOSIS OF THE LUNGS 
it is not the caseation or pneumonia per se which interrupts the rays and 
produces the shadow, but the lack of air in comparison with the air- 
bearing lung. 
4. Cavities in the lung are frequently indicated by annular figures, 
but cavities are also frequently found at autopsy with no rings on the 
plate; conversely, annular figures are frequently found on the plate 
with no corresponding cavities at autopsy. 
5. The pleura at autopsy was sometimes normal, sometimes thickened, 
but we were unable to distinguish the difference on the plate. 
6. In spite of careful study of six cases while comparing them with 
the sectioned lungs at autopsy, our subsequent readings of even these 
cases have never yet been accurate. 
7. The failures on the part of the x-ray appear to be due principally 
to the fact that no two plates are ever taken under exactly similar condi- 
tions of milliamperage, time exposure, and distance, producing a differ- 
ence between plates which may readily be attributed to the pathological. 
Again the pathological may be so over-exposed as to be lost. In other 
words each plate, is a law unto itself. 
8. Advanced tuberculosis with massive fibrosis or caseation, numerous 
tubercles, or large cavitation can usually be diagnosed by the x-ray, 
though still more accurately by physical signs. The best results are 
achieved by careful physical diagnosis compared with the x-ray plate 
and, if differences are found, submission of these to physical examination 
before conclusion is reached. 
9. Physical signs have frequently brought out lesions and extensions 
invisible on the x-ray; while the latter have shown only rare ones not 
found by the former. 
10. The x-ray showed a possible explanation for cracked pot sound 
over cavity, namely, two cavities, superimposed. 
11. Though there have been a number of careful autopsy comparisons 
with plates of advanced tuberculous cases, there is room for more; and 
there has been entirely too little comparison with early cases. The 
principal study still remaining, therefore, is a number of carefully taken 
plates of patients dying of other than lung disease with the expectation 
of finding at autopsy the occasional occurring early tuberculosis, and the 
comparative study of these lungs and plates.* 
REFERENCES 
1. Cohn: Die anatomische Bedeutung der Lungenrbntgenogramme und ihre Bezie- 
hunger zur Rontgendiagnostik der Lungentuberkulose, Zeitschrift fur Tuber- 
kulose, XVII, 1911, p. 217. 
2. Sewall and Childs: A Comparison of Physical Signs arid X-Ray Pictures of the Chest 
iu Early Stage of Tuberculosis, Archives of Internal Medicine, X, 1912, p. 45. 
3. Baetjer: Early Diagnosis of Pulmonary Tuberculosis by X-Rays, Twelfth Annual 
Meeting of the National Association for the Study and Prevention of Tubercu- 
losis, 1916, p. 150. 
4. Sampson, Heise and Brown: Study of Pulmonary and Pleural Annular Radiographic 
Shadows, National Tuberculosis Association, Fourteenth Annual Meeting, 
1918, p. 171. 
*Our sincerest thanks are due to our two pathologists, Captain Gayfree Ellison, M. C. 
and Captain Harold G. Palmer, IVL aid and interest made the work a pleasure 
instead of a drudgery. JOSEPH WALSH, M.D., JAMES W. WOOD, M.D., CAPT. JAMES C. THOMPSON 
5. Barjon: Radio-Diagnosis of Pleuro-Pulmonary Affections, translated by Honeii, 
Yale University Press, 1918. 
6. Hamman and Baetjer: Pulmonary Physical Signs and Roentgen Ray Failures in 
Healthy Adults, Archives of Internal Medicine, Nov. 1914, p. 757. 
7. Helm: Zur Rontgendiagnostik interlobarer prozesse, Fortschritte auf d. Geb. d. 
Rbntgenstrahlen, XXX, 1918, p. 169. 
8. Dunham, Boardman and Woolman: Stereoscopic X-Ray Examination of the Chest 
with especial reference to the Diagnosis of Pulmonary Tuberculosis, Bulletin 
Johns Hopkins Hospital, XXII, 1911, p. 248. 
Dunham: Stereoroentgenography, Pulmonary Tuberculosis, Southworth Co., Troy, 
New York, 1917. 
Dunham: A Manual of the Roentgenological Examination of the Chest, American 
Review of Tuberculosis, Vol. II, No. 9, Nov. 1918. 
13