PART TWO
Diagnostic Cytology
diagnosing significant endometrial pathology. Three different
reporting terminologies of normal and abnormal endometrial
cells in CV cytology are compared in Table 10.4.
Three cytologic findings in cervicovaginal cytology have been
suggested as indicators of endometrial carcinoma and disease.
The hyperestrogenic environment commonly seen in type I
carcinomas promotes increased maturation of the squamous
epithelium, resulting in a shift to the right in the vaginal matu-
ration index (MI) for a postmenopausal woman. This finding
alone, however, is a poor predictor of endometrial carcinoma
since it is frequently found in the absence of endometrial car-
cinoma.11 For example, both systemic drug and local treatment
can lead to this MI shift. A second cytologic finding, histio-
cytic exodus, has also been suggested as a useful predictor of
endometrial neoplasia since a stromal infiltrate of histiocytic-
type cells is frequently seen in both endometrial hyperplasia
and neoplasia. Although initial studies of histiocytic number
and exudate in postmenopausal women suggested that this
finding did indeed have utility in detecting endometrial carci-
noma,12 subsequent analyses have concluded that this finding
is too non-specific to be useful.11,13,14 Thus, with exclusion of
both MI and histiocytic exodus as useful predictors the detec-
tion of endometrial carcinoma in cervicovaginal cytology is
dependent upon the presence of cytologically abnormal or
normal endometrial cells.11,14 Reports on the prevalence of such
endometrial cells in CV cytology in cases of endometrial carci-
noma have been quite variable. Some studies have found that
only a minority of endometrial carcinomas exhibit endome-
trial cells, whereas others have reported that most endometrial
carcinomas are detectable by CV cytology.12,15,16 Endometrial
carcinomas which have detectable malignant cells in cervico-
vaginal cytology are more likely to be high-grade and higher
stage.17 A consensus that the variable, and low, sensitivity of
cervicovaginal cytology in detecting endometrial carcinoma
precludes its use as a screening tool for this disease has devel-
oped.18
There is evidence to suggest that liquid-based cytology is more
sensitive than conventional cytology in detecting endometrial
adenocarcinoma.19,20 The superior performance of liquid-based
cytology in this regard has been attributed to the more consist-
ent and optimal presentation of endometrial cells as compared
to that in conventional cytology.
Direct Endometrial Sampling
Direct sampling of the endometrial cavity can occur using
two methods. Aggressive endocervical sampling can reach
the lower uterine segment and retrieve cells and fragments
from this region.21 In addition, many proprietary endometrial
sampling devices for procuring endometrial cells have been
developed. Most of these devices use elongated brushes that
can be passed through the endocervical canal to reach the
endometrial cavity.22-24 The addition of a protective sheath
to minimize contamination from the exo- and endocervix
has improved the diagnostic accuracy that can be achieved.25
After the cells are procured the sample can be smeared on a
slide or placed into liquid fixation.22-24 Most commonly the
sample is fixed in a liquid media and prepared using centri-
fugation techniques. Tissue fragments can be filtered to pro-
duce cell blocks. The cell block technique has the advantage
that multiple slides can be prepared for both regular stain-
ing and ancillary tests such as immunohistochemistry.22 The
sampling devices are reported to be easy to use and are well
accepted by patients, causing less discomfort than histologic
biopsy.26
Direct cytologic sampling of the endometrial cavity27-32 has
been reported to be capable of detecting endometrial carcinoma
with high (> 90%) sensitivity.15,30,33-36 However, there has not
been widespread adoption of endometrial cytology in clinical
practice as either a diagnostic or screening tool, at least in North
America and Western Europe for three reasons principally.
Firstly, reporting of "suspicious" or "atypical" has led to a high
rate of non-specificity, often necessitating additional invasive
procedures for assessment (e.g. curettage and biopsy).30,31,33,37
Secondly, unsatisfactory endometrial sampling is not uncom-
mon.33 Thirdly, histologic sampling devices are often preferred
since they are capable of detecting not only endometrial carci-
noma with equivalent sensitivity as endometrial cytology, but
also additional clinically significant endometrial lesions, such
as polyps and hyperplasia. These important lesions are not reli-
ably detected by cytologic sampling alone.31,34,37 Some methods
combine use of cell block preparation and histologic prepara-
tion to compensate for this inadequacy of stand-alone cytologic
techniques.
Direct endometrial cytology may find a role in selective clini-
cal situations. It may be useful as an adjunctive technique to
endometrial biopsy and curettage in the assessment of women
with abnormal uterine bleeding, but only if laboratories are
properly trained to prepare, screen, and diagnose these speci-
mens.23,38 Some endometrial carcinomas of fundic and isthmic
origin may escape detection in curettage specimens, and sub-
sequent endometrial cytology may be useful for the detection
of these tumoral sites.39 Finally, endometrial biopsy in the out-
patient setting may be impossible or inadequate in the very
small atrophic uterus, whereas direct endometrial cytology may
be acceptable.40
Normal-Appearing Endometrial Cells
and Gestational Changes
Normal-Appearing Endometrial cells:
cytomorphology, Differential Diagnosis,
and Significance in cV cytology
Cells from the endometrial cavity that can be detected in exfo-
liative CV cytology include epithelial, stromal, and histiocytic
cells. These cells can be spontaneously shed or directly sampled
from the lower uterine segment with Pap sampling devices.
Normal-appearing endometrial cells (NEC) are of epithelial
and/or stromal type, and should be reported as a possible marker
for endometrial pathology only if they present in exfoliative CV
cytology as intact endometrial clusters.
There are many shared features of NEC cytomorphology
in CV and liquid-based cytology (SurePath (BD Diagnostics
Systems-TriPath, Burlington NC), ThinPrep (Hologic, Marl-
borough, MA), MonoPrep (MonoGen Inc., Lincolnshire,
IL)). The most typical appearance is that of the endome-
trial cell ball that contains a condensed core of stromal cells
surrounded by epithelial cells with more abundant paler
cytoplasm41-43 (Fig. 10.7). Fine nucleoli may be present but
nuclear degeneration may obscure this finding (Fig. 10.8).
Endometrial cell balls correspond to exodus or endometrial
breakdown noted in histologic samples,43 and are often most
252
previous page 252 ComprehensiveCytopathology 1104p 2008 read online next page 254 ComprehensiveCytopathology 1104p 2008 read online Home Toggle text on/off