THE DIAGNOSIS OF PRIMARY SAR- 
COMA OF THE PLEURA FROM THE 
CELLS FOUND IN THE PLEURITIC 
EXUDATE. 
ALDRED SCOTT WARTHIN, M.D., 
OF ANN ARBOR, MICH..; 
INSTRUCTOR IN PATHOLOGY IN THE UNIVERSITY OF MICHIGAN. 
BY 
FROM 
THE MEDICAL. NEWS, 
October 16, 1897. [Reprinted from The Medical News, Oct. 16, 1897.] 
THE DIAGNOSIS OF PRIMARY SARCOMA OF 
THE PLEURA FROM THE CELLS FOUND 
IN THE PLEURITIC EXUDATE. 
By ALDRED SCOTT WARTHIN, M.D., 
OF ANN ARBOR, MICH.; 
INSTRUCTOR IN PATHOLOGY IN THE UNIVERSITY OF MICHIGAN. 
The microscopic examination of the sediment of 
pleural and peritoneal effusions for the purpose of 
establishing the diagnosis of new growths of these mem- 
branes before a clinical diagnosis can be determined 
is a diagnostic measure but rarely instituted, if the 
scantiness of the literature upon the subject is to be 
taken as an index. Much has been done with the 
chemistry of these fluids, but few careful investiga- 
tions have been reported of the cellular-constituents 
in those cases in which malignant tumors of the 
pleura or peritoneum were present. 
Such examinations have been much more frequently 
made of the urine, sputum, stomach-contents, feces, 
and vaginal discharges in the hope of establishing at 
the bedside the diagnosis of new growth, and not a few 
cases have been reported in which a correct diagnosis 
was made from the presence of tumor-cells or tumor- 
tissue in these excretions and secretions. Though 
in these days, when exploratory aspiration of almost 
every case of pleural or peritoneal effusion is made 
as a matter of course, and when the centrifugating- 
machine is so frequently utilized in the examination 
of the sediment, few cases of malignant growth diag- 
nosed in this way have been reported, This seems 2 
strange when one considers the large number of 
cases of ascites due to peritoneal metastases in ovarian 
and gastric carcinoma, and the frequent cases of 
pleuritic effusion due to secondary carcinoma of the 
pleura following primary tumors of the mammary 
gland, etc. 
It is evident that the possibility of diagnosticating 
new growths from the microscopic examination of 
effusions before a successful clinical diagnosis can be 
made would be a very important advance. The 
question then stands, To what extent is it possible, 
from the examination of the cellular elements of effu- 
sions, to establish the diagnosis of new growth of 
pleura or peritoneum ? 
The usual examination of the sediment of these 
fluids is confined to the direct microscopic examina- 
tion of the fresh specimen, or to the examination of 
stained preparations for bacteria. In the majority of 
cases the search is limited to leucocytes and red blood- 
cells; and more emphasis is laid upon these than upon 
the character of other cells which may happen to be 
present. In addition to the blood-corpuscles, note 
is also made of endothelial cells, fat drops, fatty acid 
crystals, Charcot-Leyden crystals, cholesterin, pig- 
ment, fibrin threads, micro-organisms, and foreign 
material. 
The presence of large numbers of large polymorph- 
ous cells, having numerous vacuoles or containing fat 
droplets, has led to the assumption in many cases 
that these cells were carcinoma cells, but from the 
fact that the endothelial cells in effusions exhibit such 
diversity of size and shape, and show also a similar 
degree of vacuolation and fatty degeneration, some 3 
investigators have declared it impossible to diagnos- 
ticate with certainty the existence of a new growth 
unless bits of tissue showing definite structure can be 
found. 
Quincke1 endeavored to establish the principle that 
carcinoma cells can be distinguished from endothelial 
cells by their size, their number, their grouping into 
masses, and perhaps by their glycogen reaction. 
Later Rieder,2 after a careful study of carcinoma cells 
in peritoneal and pleuritic fluids, concluded that the 
most important differential point between them and 
endothelial cells is in the large number of mitoses, 
especially asymmetrical division forms, which are 
found in the cells of tumors and not at all, or to a 
very slight degree in endothelial cells. The finding 
of many cells showing division figures he believed to 
speak conclusively for malignant growth. Neelsen, 
quoted by Rieder in the same article, diagnosticated 
carcinoma of the kidney from bits of tissue discharged 
in the urine, and which contained cells showing many 
asymmetrical mitoses. 
Since Rieder’s article, I have not been able to find 
any work bearing upon this point. The tumors which 
have been diagnosticated by a microscopic examina- 
tion of cells found in effusions have been in all cases 
carcinoma, or, as in Rieder’s case, atypical growths 
(sarcoma carcinomatosum), probably a mixed tumor 
arising from an ovarian teratoma. I have not been 
able to find any statement of a typical sarcoma of the 
peritoneal or pleural cavity, diagnosticated from the 
cells found in the effusion, but I have not been for- 
1 Dentsch. Archiv.f. Klin. Met/., vol. xxx. 
2 Ibid., vol. liv. 4 
tunate enough to secure all of the literature on pri- 
mary endothelioma of the pleura, so that my search 
must be looked upon as incomplete. The evident 
rarity of such cases makes the report of the one which 
came under my observation some time ago very im- 
portant because of its bearing upon the question of 
diagnosis. As far as I know, it is the only case in 
which sarcoma of the pleura has been correctly diag- 
nosticated by an examination of the pleuritic exudate. 
I am not at liberty to give the full clinical history 
of the case, my connection with it being limited to 
the examination of the effusion and the material ob- 
tained at the autopsy. The most important facts are 
as follows : 
X., a German, male, about forty-five years of age, 
gave an indefinite history of dyspnea, pain in the 
right side of the thorax, cough, and progressive weak- 
ness. He also had irregular fever. The exact dura- 
tion of these symptoms could not be ascertained be- 
cause of the clouded mental condition of the patient 
when first seen. 
The physical examination showed increased fulness 
of the right side of the thorax, with entire absence 
of movement. There was slight bulging in the fourth 
and fifth intercostal spaces between the nipple and 
sternal lines on the right side. There was a weak 
apex-beat in the left anterior axillary line in the fifth 
interspace. Vocal fremitus was absent over the right 
side. There was no friction fremitus. On percus- 
sion, dulness was found below the fourth rib on the 
right, rising to the level of the second rib when in a 
sitting position. Above this, there was dull tym- 
panitic resonance. Over the area of dulness there 
was no breath-sounds and no adventitious sounds. 
Above the second rib there were faint blowing breath 5 
sounds. The voice-sounds were only faintly trans- 
mitted above the second rib. 
A diagnosis of pleurisy with effusion was made, 
the patient was aspirated, and the fluid sent to the 
pathologic laboratory for examination. 
Pleuritic Exudate.—The fluid was brownish-red 
in color, with many yellowish or grayish flakes which 
settled quickly after agitation. The fluid was thick, 
but the amount obtained was not sufficient for the 
estimation of its specific gravity. On standing, fibrin 
was formed. The fluid coagulated solidly on boiling. 
The fresh sediment was examined, numerous stains 
and reagents being applied. Cover-glass prepara- 
tions were made and a portion of the sediment was 
hardened for future examination. 
The examination of the sediment, stained by allow- 
ing a drop of dilute methylene-blue solution to run 
under the cover-glass, showed the field to be full of 
spindle-cells of about the same shape and size, 
mostly single, but lying also in groups. There were 
numerous red blood-cells, leucocytes (mostly lymph- 
ocytes), fibrin threads, blood-pigment, and an abun- 
dance of fine granules, supposed to be cell debris. 
The leucocytes were not present in such numbers as 
to admit of the fluid being called purulent. En- 
dothelial cells were present in only a very small 
number. They showed various degrees of vacuo- 
lation. 
From the great number of spindle-cells present, 
the diagnosis of small spindle-cell sarcoma was made. 
The diagnostic points considered were the number 
of the cells and their character. The one alternative 
set against this diagnosis was a pleuritic effusion con- 6 
taining large numbers of fibroblasts, but the regular 
size and shape of the cells were opposed to this view. 
The patient gradually became worse. A rapidly 
progressing anemia and weakness, with constant de- 
lirium, were the most marked symptoms up to the 
time of death. The bulging on the right side be- 
tween the nipple and sternal lines became more 
prominent before death. 
The post-mortem examination was held four hours 
after death. The subject was very anemic and much 
emaciated. The pathologic conditions were con- 
fined to the thorax. The right half of the thoracic 
cavity was filled with a thick brownish-red effusion 
containing many large flakes and stringy masses. To 
these, blood-clots were attached. There were also 
large blood-clots floating in the fluid. The entire 
surface of the mediastinal pleura and of the anterior 
parietal pleura as far as the anterior axillary line was 
covered by a velvety growth; in some places the sur- 
face of the growth was covered by long strings of 
fibrin and blood-clot, firmly adherent to the tumor 
mass. In many areas the tumor itself consisted of 
long strings extending out into the pleuritic cavity. 
The greatest thickness of the growth (about three 
inches) was at the level of the fourth and fifth inter- 
costal spaces between the right nipple and sternal 
lines, corresponding to the area of prominence in 
the chest-wall. The right lung was completely ate- 
lectatic; the visceral pleural was covered with fibrin, 
but there were no evidences of a new growth. The 
heart was pushed to the left, the right edge being at 
the left sternal line. The left lung and pleura were 7 
normal, as were also the pericardium and heart. No 
metastases were found. 
Tumor.—The tumor was mottled red and grayish- 
white. It was so soft that it could easily be crumbled 
between the fingers, and so fragile that it could 
hardly be cut without falling to pieces. The surface 
was uniform, homogeneous, without stroma, but rich 
in large blood-vessels, having very thin walls. There 
were numerous hemorrhagic areas, also areas of ne- 
crosis, grayish-yellow in color. In the stringy areas, 
the blood-vessels were surrounded by the grayish- 
white tissue of the tumor, giving to the naked eye 
the appearance of a perivascular growth. 
Scrapings were made from the freshly cut surface ot 
the tumor, and it was found to consist of medium-sized 
spindle-cells like those found in the effusion. The 
fluid obtained from the thoracic cavity was very rich 
in the same cells, so rich that it was not necessary to 
collect the sediment for examination. The specific 
gravity of the fluid was not taken. 
Pieces of the tumor were hardened in alcohol and 
Muller’s fluid. Sections were prepared from these, 
and the growth was found to consist of spindle-cells 
grouped about the large but thin-walled blood-vessels. 
Only a single layer of endothelium separated the 
blood from the cells. That the tumor-cells arose 
from the endothelium there can be no doubt; they 
were crowded more closely just around the endo- 
thelium, and here the nuclei stained more deeply 
than those away from the vessels. In specimens 
hardened in alcohol, dividing endothelial cells were 
seen, and in some cases the cells grew into the lumen, 
forming buds on the inner wall of the vessels. In 8 
other cases, free masses of cells were found in the 
vessels. In the stringy areas covering the surface 
of the growth, all stages of development of the tumor 
tissue could be found, from the small capillary grow- 
ing out into the fibrin with only a single layer of en- 
dothelium forming its walls and with no tumor tissue 
about it, up to the large blood-vessel surrounded by 
a large cylindric mass of cells. Throughout the 
tumor the structure was essentially perivascular, so 
that the tumor consisted of cellular cylinders sur- 
rounding the blood-vessels. Mitoses'were abundant 
throughout; for the most part they were symmetrical. 
From its histologic structure the tumor should be 
styled a spindle-celled hemangiosarcoma; from its 
origin it should be classed with the endotheliomata. 
The hardened cover-glass preparations of the sedi- 
ment were not examined thoroughly until some time 
afterward. These were stained in hemalum and 
examined with an oil-immersion lens. The finer 
structure of the cells was then carefully studied. The 
cells were spindle-shaped with rather long and slender 
polar processes which did not branch as is commonly 
the case with the spindle-cells found in inflammatory 
exudates. The nucleus stained rather heavily; was 
oval, as a rule, but in many cells was slightly curved. 
The chromatin was very abundant, but had a very 
simple structure. Numerous mitoses were found, 
but in all cases they were regular and symmetrical. 
They showed all stages of development. 
As a rule, the cells were found singly, but groups 
were very common. They did not contain vacuoles, 
neither was any vacuolation of the nucleus observed. 
The rather abundant protoplasm was finely granular. 9 
The epithelioid appearance of the nucleus so com- 
mon to fibroblasts was not seen. Numerous necrosing 
cells were found, the protoplasm being more coarsely 
granular and the nucleus staining very poorly or not 
at all. 
The pleural endothelial cells were relatively scanty. 
They were very irregular in size and shape, and 
showed various degrees of vacuolation. They pre- 
Fig. i. 
Cells from effusion in primary sarcoma of the pleura, (a) En- 
dothelial cells, all the others are tumor-cells, some showing; cell divi- 
sion. (One-twelfth oil-immersion lens; ocular No. 2.) 
sented no characteristics differing from the same cells 
found in ordinary pleuritic effusions. 
In Fig. i the character and relative proportion of 
the cellular elements (blood-cells excepted) of the 
effusion in this case are shown. 
Since the examination of this case I have made a 
study of the cells found in the sediment of the exu- date in a large number of cases of diseased conditions 
of serous membranes—pleura, peritoneum, and peri- 
cardium. A description and comparison of the cells 
found under different conditions may be very valu- 
Fig. 2. 
Endothelial cells from sediment of effusion in early stage of 
acute serous pleurisy. (One-twelfth oil-immersion lens; ocular 
No. 2.) 
able in itself and of aid in establishing a basis for 
diagnosis, if such be possible. The specimens were 
taken from cases in which the diagnosis was confirmed 
post-mortem; in some instances the fluid was obtained 
directly from the body-cavities, in others by puncture 
previous to death and also at the post-mortem exam- 
ination. The cells were examined fresh, and fixed 
cover-glass preparations were made according to 
the method usually employed for the examination of 11 
blood. The cleaned cover-glasses were spread with 
a thin film of the sediment, allowed to dry in the 
air for a few minutes, and then fixed in a mixture of 
equal parts of absolute alcohol and ether (one to two 
hours in the fixing fluid is sufficient). After fixing, 
the covers were dried in the air and stained at con- 
venience. Hemalum or hematoxylon was the stains 
used most frequently. 
Acute Serous Pleurisy (Fig. 2).—The fluid was 
obtained in this case on the second day of the dis- 
ease. It was clear, pale yellow, and contained nu- 
merous small whitish particles. These, when ex- 
amined microscopically, were found to be masses of 
endothelial cells. From the large number of these 
desquamated groups of cells the pleurisy might be 
styled as desquamative. It is, however, the rule to 
find these in the very early stages of pleurisy. In 
these groups of cells the cell-membranes could not be 
made out. The cells were also found in great num- 
bers isolated or in smaller groups of two or three. 
As seen in Fig. 2, the cells were large, irregularly 
oval or round, and had an oval nucleus which stained 
less heavily than the nuclei of leucocytes or fibro- 
blasts. They were for the most part vacuolated, the 
vacuoles being due to hydropic degeneration and not 
to fatty change. Some of the nuclei were also vacu- 
olated. In many cases the vacuole occupied the en- 
tire cell, with the exception of the nucleus. In some 
cases the nuclei were partially extruded from the 
cell. Though the polymorphic character of the cells 
was very striking, no spindle forms were found in 
the early stage of simple pleurisy, though in the later 12 
stages they become abundant. No mitotic forms 
were observed. 
Acute Fibrinous Pleurisy (Fig. 3).—The effusion 
of both early and late stages of fibrinous pleurisy was 
examined. In all cases the fluid was cloudy, light 
Fig. 3. 
Cells from the sediment of acute fibrinous pleurisy, third week. 
Endothelial cells and various forms of fibroblasts. (One-twelfth 
oil-immersion lens; ocular No. 2.) 
yellow, and contained numerous flakes and strings 
of fibrin. The pleurae were covered by a thick 
layer of stringy fibrin. In both early and late 
stages the sediment contained vacuolated endothelial 
cells and numerous irregular spindle-cells, having in 
many cases finely branched polar processes. These cells had nuclei which stained more deeply than the 
endothelial cells, but in some cases they were epithe- 
lioid in character. Some of the nuclei of these 
irregular cells were vacuolated. In the early stage 
Fig. 4. 
Cells, probably fibroblasts wholly, from the sediment of peri- 
cardial fluid in the third week of acute fibrinous pericarditis. 
(One-twelfth oil-immersion lens; ocular No. 2.) 
the endothelial cells predominated, but in the older 
cases of fibrinous pleurisy the spindle forms were 
much more numerous. They are to be looked upon, 
I think without doubt, as fibroblasts. No mitoses 
were found in these cells. 
Tuberculous Pleurisy.—The exudate was obtained 
post-mortem. It was fibrinous, and similar in char- 
acter to that found in simple fibrinous pleurisy. The 14 
sediment of the fluid differed in no respect from that 
shown in Fig. 3, being made up of endothelial cells 
and fibroblasts. No mitoses were found. 
Acute Fibrinous Pericarditis (Fig. 4).—The case 
examined was seen post-mortem, death having oc- 
Fig. 5. ' 
Cells from scraping of normal pleura. Endothelial cells, with a 
few connective-tissue cells from basement structure. (One-twelfth 
oil-immersion lens; ocular No. 2.) 
curred during the third week. Both layers of the 
pericardium were covered by a thick layer of fibrin; 
in many places there were fresh adhesions which 
were easily separated by the fingers. There was a 
small quantity of thick grayish fluid between the layers. On examination, no endothelial cells were 
found in this fluid, but numerous fibroblastic forms. 
These were much more branching than the cells 
found in the pleuritic exudates. No mitoses were 
found. 
Scraping prom Normal Pleura (Fig. 5).—For 
the sake of comparison, the cells obtained by scra- 
pings from normal pleurae were also studied. The 
surface of the membrane was gently scraped with the 
edge of a sharp knife so as not to loosen the cells of 
the basement structure. As seen in Fig. 5, the cells 
obtained in this manner are almost wholly character- 
istic polymorphic endothelial cells; the few spindle- 
cells found were doubtless from the basement con- 
nective tissue. 
After a study of the sediment of many cases of 
effusion into peritoneal, pleural, and pericardial 
cavities, the following points seem to me to be worthy 
of emphasis : 
1. In all cases of this kind, when fluid can be ob- 
tained by exploratory puncture, cover-glass prepara- 
tions should be made by fixing in absolute alcohol and 
ether as described above. This should be done im- 
mediately after the withdrawal of the fluid from 
the body in order that the mitoses may be pre- 
served. 
2. In early stages of simple pleurisy the sediment 
contains endothelial cells and no fibroblasts. The 
endothelial cells may be desquamated in large pieces. 
3. In fibrinous pleurisy, tuberculous or septic, the 
sediment contains, in addition to endothelial cells, 
numerous fibroblastic forms. In later stages, when 
adhesions are forming, the endothelial cells may be 
entirely absent, and only fibroblasts be found. 16 
4. In the diagnosis ofsarcoma of serous cavities, the 
character and number of the cells and the presence 
of numerous mitoses are the points upon which the 
diagnosis must be based. It is, of course, evident 
that the quickly growing and soft forms of sarco- 
mata are the ones most apt to lead to the presence 
of cells in the fluid. The tendency of all sarcomata 
to break down after reaching a certain growth makes 
it extremely likely that in most cases free cells will 
be found. The diagnosis of spindle-celled sarcoma 
is easy, from the occurrence of large numbers of 
spindle-cells of similar size and shape, showing fewer 
branching processes than fibroblasts, and containing 
more mitoses. In the round-cell forms, the tumor 
cells are to be distinguished from the endothelial 
cells by their size, character of nucleus, and the re- 
lation of the nucleus to the protoplasm. The presence 
of mitotic forms will again be the most important 
point; in the small round-cell sarcoma, these are not 
so easily seen. The greatest difficulty will be in the 
differentiation of the cells of a polymorphic celled 
sarcoma from the fibroblastic forms found in fibrinous 
exudates. Aside from the general character of cell 
and nucleus, the presence of division forms must in 
all cases be looked upon as being the most important 
point. 
In conclusion, therefore, the presence of numerous 
cell-division forms in the cells of the sediment of 
serous exudates may be taken as strong (perhaps con- 
clusive) evidence that the effusion is due to the pres- 
ence of a new growth, inasmuch as mitoses are but 
rarely found in the cells of purely inflammatory 
exudates. The Medical News. 
Established in 1843. 
A WEEKLY MEDICAL NEWSPAPER. 
Subscription, $4.00 per Annum. 
The American Journal 
OF THE 
Medical Sciences. 
Established in 1820. 
A MONTHLY MEDICAL MAGAZINE. 
Subscription, $4.00 per Annum. 
COMMUTATION RATE, $7.50 PER ANNUM. 
LEA BROS & CO 
NEW YORK AND PHILADELPHIA.