PART TWO
Diagnostic Cytology
Fig. 19.145 This cell block of pleural effusion of an 8-year-old
depicts a fragment of spindle cell neoplasm, diagnosed as malignant
neurolemmoma (H&E x LP).
True histiocytic neoplasms exhibit a cytologic spectrum, rang-
ing from cells that have the appearance of normal histiocytes to
malignant forms that have a similar general appearance, but in
which the cytoplasm is much reduced. These conditions have
rarely been reported in serous effusions, and we have seen only
two examples in our laboratory, both from young children. In
each case, the neoplastic cells had the appearance of small his-
tiocytes; however, in both patients the condition was fatal, and
at necropsy numerous organs were infiltrated by these benign-
appearing cells (Fig. 19.146).
Reporting of Results and Statistics
Reporting
To be of diagnostic value to clinicians, reports on the presence or
absence of neoplastic cells in serous fluids should be expressed
as unequivocally as possible. When neoplastic cells are present
in a fluid, they are usually present in large numbers and are read-
ily recognizable as such. Seldom should it be necessary to resort
to special technical methods to determine whether neoplastic
cells are present. The incremental yield of "positive" results from
the use of such methods is extremely small, scarcely affecting the
statistical analysis of results.
We have never used a numeric classification, such as the
Papanicolaou system, in our reporting of cytologic specimens,
preferring to use telegraphic English prose rather than Roman
numerals. Typical examples of our reporting are as follows:
• Examination for neoplastic cells: Positive, for
adenocarcinoma.
• Examination for neoplastic cells: Negative. Numerous
neutrophilic leukocytes.
• Examination for neoplastic cells: Negative. This effu-
sion shows the pathognomonic cytologic picture of
rheumatoid pleuritis.
It should be understood by clinicians that reporting a speci-
men as being "negative" for neoplastic cells does not preclude
neoplasm as the cause of the effusion. As mentioned previously,
effusions may be caused by neoplastic obstruction of lymphatic
Fig. 19.146 This cell block of pleural fluid from an infant depicts numerous
cells of malignant histiocytosis (unclassified). The cells look like miniature,
non-phagocytic histiocytes. At necropsy many areas were infiltrated by these
cells (H&E x MP).
or blood vessels or inflammation secondary to the presence of
neoplasm or both. The only time we comment with any fre-
quency with respect to non-neoplastic cells is when numerous
neutrophilic leukocytes are present. We do not comment on the
presence of mesothelial cells, macrophages, or lymphoid cells,
believing that to do so would be providing the clinician with
useless information.
We do not report cells as being "consistent with neoplasm,"
which is merely a way of stating that cells that may be neoplastic
but that, on the other hand, may not be neoplastic are present.
As much as possible, we avoid reporting cells as "suspicious" of
being neoplastic, believing that such a report is not much more
useful to the clinician than no report at all. When we have been
driven to issuing a "suspicious" report, it is because of the pres-
ence of only one or two cells that are strongly suggestive of neo-
plasm. Such a report implies that we are almost certain that the
fluid contains neoplastic cells, and in such cases examination
of additional material, perhaps from the same specimen, has
generally enabled us to make a definite diagnosis of neoplasm.
Reliability of Positive and Negative Reports
In their monograph, Spriggs and Boddington analyzed the sta-
tistical results of six published series, which were the only large
series published in the previous 25 years from which relevant
figures could be extracted.9 Some of these series dealt exclusively
with pleural or peritoneal effusions, whereas the others dealt
with both types of effusion. The number of specimens in the
six series ranged from 159 to 2198. The total number of cases
covered by these series was 6001, with an average of 1000.
In their analysis of these series, Spriggs and Boddington cal-
culated sensitivity, specificity, predictive value of positive reports,
and predictive value of negative reports. To give examples,
sen-
sitivity
of 0.70 means that in cases of cancer, 70% were given
unequivocal positive reports and 30% were reported as either
negative or suspicious.
Specificity
of 0.95 means that in cases
without cancer, 95% were reported as negative and 5% were
given false-positive or suspicious reports. A
predictive value of
positive
of 0.97 means that 3% of positive reports were false, and
a
predictive value of negative
of 0.80 means that 20% of negative
reports were from cases with cancer.
572
previous page 566 ComprehensiveCytopathology 1104p 2008 read online next page 568 ComprehensiveCytopathology 1104p 2008 read online Home Toggle text on/off