13
Respiratory Tract
Table 13.31
Prediction of Histologic Type of Primary Lung Cancer from Specimens Obtained by Fine-Needle Aspiration
Histologic
diagnosis
Cytologic diagnosis
Squamous
cell carcinoma
Adenocarcinoma
Large-cell
carcinoma
Small-cell
carcinoma
Adenosquamous
carcinoma
Carcinoid
Unclassified
neoplasm
Squamous cell
carcinoma
47
1
5
1
0
0
2
Adenocarcinoma
8
27
19
0
6
0
4
Large-cell
carcinoma
2
0
10
0
1
0
2
Small-cell
carcinoma
1
0
0
18
0
0
1
Adenosquamous
carcinoma
1
0
0
0
2
0
0
Carcinoid
0
0
0
0
0
1
0
Correlation
47/59 (80%)
27/28 (96%)
10/24 (42%)
18/19 (95%)
2/9 (22%)
100 (0%)
Reproduced with perm
l
ission from Johnston WW: Cytologic correlations. In: Dail DH, Hammar SP (eds)
Pulmonary Pathology.
New York: Springer-Verlag, 1987.
combination of the two results improved the result to 96.6%.
In the benign group, 71.1, 70.1, and 74.2% of cases received
specific or nonspecific diagnoses by FNA, core biopsy, and their
combination respectively.570
In a similar type of study, Bocking and his associates com-
pared the diagnostic accuracy in 501 thoracic FNAs and core
biopsies in 482 patients. The overall sensitivities of the two
biopsy methods were equal (FNA 98.4% and core biopsy 98%),
but the typing accuracy was higher for core biopsy (87.2%) than
for FNA (83.5%).571
In a study of 52 patients, Cary and associates compared the
diagnostic accuracy of FNA versus core biopsy in each patient.
The diagnostic rate was 79% for FNA and 75% for core biopsy.
Use of both techniques increased the diagnostic rate to 88%.572
The experience of the investigators in correlations between
cytologic diagnoses from FNAs and histologic diagnoses is
recorded in Table 13.31. In this study, 159 patients with primary
lung cancer and with both needle aspirates and tumor tissues
available were reviewed. In 95% of cases diagnosed as small-
cell undifferentiated carcinoma by aspiration, the diagnosis
was confirmed by histologic study; 96% of adenocarcinomas
so diagnosed by aspiration were confirmed by histologic study.
In a similar fashion, other levels of correlation were as follows:
squamous cell carcinoma 80%, and large-cell undifferentiated
carcinomas 42%.
In those situations in which the cytologic-histologic cor-
relation is not high, it should not be concluded that the cyto-
logic interpretation is obviously an error and that the histologic
interpretation is correct. Although this was the original thesis
on which the discipline of cytology was founded, cytology has
matured and come of age, and it can now be appreciated that
in some situations the cytologic interpretation may be just as
correct as the tissue interpretation or, in some cases, more accu-
rately reflective of the nature of the lesion than the tissue exam-
ined. This last situation is being supported more and more by
investigations into the ultrastructure of lung tumors in which
both cytologic and histologic diagnostic interpretations had
been rendered.301,329,334-337,573 Indeed, some electron microscopic
studies are beginning to challenge most of the currently accepted
systems of diagnostic nomenclature for lung cancer. A growing
general awareness of the cellular heterogeneity present in most
of the non-small-cell lung carcinomas is evident.
Further studies in the literature suggest that the large-cell
carcinomas and poorly differentiated carcinomas, whether
of adenomatous or squamous differentiation, may show on
ultrastructural examination multiple patterns of differentia-
tion. A series of studies on the histogenesis of lung cancer have
shown that many poorly to well-differentiated squamous cell
carcinomas,
poorly to well-differentiated adenocarcinomas,
and giant cell carcinomas, diagnosed as such by conventional
light microscopic criteria, were in reality tumors exhibiting
dual differentiation toward both squamous cell carcinoma and
adenocarcinoma. Horie and Ohta examined by light and elec-
tron microscopy 26 human lung tumors classified as large-cell
undifferentiated carcinomas.337 On the basis of their observa-
tions, these investigators were able to subclassify all of these
tumors into squamous, adenosquamous, and giant cell carci-
nomas. Dingemans and Mooi investigated by electron micro-
scopy a series of 40 lung tumors that had been diagnosed by
conventional light microscopy as squamous cell carcinoma.301
At both the tissue level and the cellular level, the tumors showed
high variable ultrastructural details embracing both tonofibrils
and desmosomes on one hand and unmistakable adenomatous
differentiation on the other.
comparison of the Diagnostic Accuracy of Sputum
and Bronchial Material and Fine-Needle Aspirates
Considerable controversy exists about the relative roles that
sputum and bronchial cytology, on one hand, and FNA, on the
other, should have in the diagnostic workup of patients with
suspected lung cancer. One group supports a protocol that
359
previous page 356 ComprehensiveCytopathology 1104p 2008 read online next page 358 ComprehensiveCytopathology 1104p 2008 read online Home Toggle text on/off