13
Respiratory Tract
Table 13.27
Distribution by Diagnostic Categories of Fine Needle Aspirates
Diagnostic
Number
Percentage
category
of patients
of patients
Primary lung cancer
765
43.8
Neoplasm metastatic
to lung
261
15.0
Inconclusive
82
4.7
Benign
637
36.5
Total
1745
100.0
Reproduced with permission from Johnston WW: Cytologic correlations. In: Dail
DH, Hammar SP (eds)
Pulmonary Pathology.
New York: Springer-Verlag, 1987.
major concern occurs in those situations in which the cytologic
prediction of specific cellular differentiation was not reflected in
the tissues available for diagnosis. Thus, all cases of squamous
cell carcinoma and adenocarcinoma designated as such by orig-
inal cytologic diagnosis, but not sustained by the tissue diag-
nosis, were reviewed. For the squamous cell carcinoma group
diagnosed cytologically,
rediagnosis sustained the original
interpretation in 83% of the cases, and for the adenocarcinoma
group in 57% of cases. Undoubtedly, such observations support
the significance of the problems inherent in sampling in tissues
as well as cytologic specimens. Further, they support the view
that in some tumors, the morphologic features of the cells them-
selves may provide greater information about the true nature of
the neoplasm than the histologic material.
Diagnostic Accuracy of Specimens obtained
by Fine-Needle Aspiration
The diagnostic accuracy of FNA of lung as reported in the lit-
erature has shown sensitivity rates in the range of 75-95% of
patients with malignant neoplasms correctly diagnosed by this
method and specificity rates of 99% or more.
Tables 13.27 and13.28 show the experience of our laboratory
with FNAs from the lung. In nearly 60% of the cases, a malig-
nant neoplasm was detected. Approximately 44% of these were
primary lung neoplasms. In approximately 5% of patients, can-
cer was suspected but could not be conclusively diagnosed by
cytologic diagnosis alone (see Table 13.27). In 765 FNAs of the
lung, the specimens were interpreted as conclusively diagnostic
for a primary malignant neoplasm. These neoplasms are listed
in Table 13.28 and have been divided according to their his-
tologic type. Squamous cell carcinoma was the most frequent
diagnosis and was made in approximately 38% of patients. The
other major diagnoses in descending order of frequency were
large-cell undifferentiated carcinoma 27.2%, adenocarcinoma
13.5%, small-cell undifferentiated carcinoma 12.7%, and ade-
nosquamous carcinoma 3.4%. In 4.2% of patients, although it
was concluded that neoplastic cells were present, no opinion
could be reached with respect to their further classification.
Although the cytologic presentation of the major types of
lung cancer in FNAs has been discussed here in the appropri-
ate chapter sections, one important point is worthy of special
emphasis. The morphology of lung tumors in these specimens
is essentially the same as that in sputum and bronchial mate-
rial but with one additional characteristic: The FNA, because of
Table 13.28
Classification of Primary Lung Neoplasms Diagnosed by Fine
Needle Aspiration
Diagnostic
category
Number
of patients
Percentage
of patients
Squamous cell
carcinoma
290
37.9
Adenocarcinoma
103
13.5
Large-cell carcinoma
208
72.2
Small-cell carcinoma
92
12.7
Adenosquamous
carcinoma
26
3.4
Plasmacytoma
2
0.3
Carcinosarcoma
2
0.3
Carcinoid
3
0.3
Lymphomatoid
granulomatosis
1
0.1
Unclassified
33
4.2
Total
765
100.0
Reproduced with permission from Johnston WW: Cytologic correlations. In: Dail
DH, Hammar SP (eds)
Pulmonary Pathology.
New York: Springer-Verlag, 1987.
direct sampling within the tumor by the needle, should con-
tain large numbers of cancer cells and tissue fragments. The cell
block becomes a useful vehicle for their study. Microbiopsy tis-
sue specimens are not infrequently available for evaluation in
these preparations.
Indeed, the presence of only small numbers of
putative tumor cells in an FNA should be a significant warning to the
pathologist to exert extreme caution in rendering a conclusive cancer
diagnosis. This is the setting in which a false-positive diagnosis of
cancer is most likely to occur.
We have studied 288 consecutive patients from whom tissue
had been obtained within a reasonably short interval before
or after the aspiration. This group of patients was used to help
address the question of effectiveness of detection in our labo-
ratory. The comparison between the FNA diagnosis of cancer
and the histologic diagnosis of cancer in these 288 patients is
shown in Table 13.29. In 246 patients (85.4%), both the aspi-
rate and the histologic specimen reflected a diagnosis of cancer.
In 40 patients (13.9%), the tissue revealed a cancer that had
been missed by the needle aspiration. Review of these speci-
mens failed to reveal tumor cells that could have been over-
looked previously. In two patients, the histologic examination
did not show any cancer. These two cases, then, are considered
to be false-positive diagnoses for cancer, with a calculated rate of
0.7%. In both of these cases, the cardinal error of overinterpreta-
tion of only a few atypical cells prompted these errors. Sensitiv-
ity of FNA biopsy was found to be 86% and specificity 99.9%.
In 1988, Simpson and associates at the Mayo Clinic reported
on 233 consecutive cases and noted a sensitivity of 82% and a
specificity of 100%.548
From the standpoint of the conclusiveness of a negative diag-
nosis, Afify and Davila, in a review of 1181 fine-needle aspirates
from the lung, found a negative predictive value of 77%. Inad-
equate sampling was responsible for all false-negative cytologic
diagnoses in this series.549
357
previous page 354 ComprehensiveCytopathology 1104p 2008 read online next page 356 ComprehensiveCytopathology 1104p 2008 read online Home Toggle text on/off