8
Benign Proliferative Reactions, Intraepithelial Neoplasia, and Invasive cancer of the Uterine cervix
Fig. 8.134 HSIL lesion in liquid-based cytology showing nuclear size variation, hyperchromasia, markedly increased N/C ratio, and irregular nuclear
membranes ("ceribriform”) with 3-dimensional structural abnormalities (Papanicolaou x HP).
is prepared. "Direct-to-vial" studies compare the performance
of LBC during a given time period with historic control data
using conventional cytology. Up till now more studies have
been undertaken with ThinPrep than with SurePath and most
of these studies address issues such as test performance in terms
of relative sensitivity, relative specificity, and quality of the cell
preparations.
Considering direct-to-vial studies,
overall
an
increased detection of LSIL as well as HSIL has been reported.
Arbyn and associates found a pooled ratio for detection rate of
HSIL in ThinPrep specimens versus conventional cytology of
1.72 (95% CI 1.42-2.08) and for SurePath specimens versus
conventional cytology of 1.42 (95% CI 1.14-1.89).261 LBC stud-
ies reported also reduced percentages of "unsatisfactory" slides,
forcing the National Health Service of the UK to introduce LBC
in their nationwide screening program.261 However, almost all
these studies had serious flaws in their design. None of the
published studies that used a cohort design allocated the stud-
ied cases randomly in both study arms. Many studies in favor
of LBC did not perform a blinded review of histology as gold
standard for follow-up. Moreover, no one study included thor-
ough follow-up of the negative smears.262 In a recent extensive
systematic review, Davey et al. showed that only very few LBC
studies are adequately designed to reliably compare conven-
tional cytology with LBC.263 Of 56 reviewed studies only 3 were
of high quality and the authors concluded that up till now no
claims can be made that LBC performs better than conventional
smears in detecting more high-grade lesions.
Except for the USA and the UK, no country has introduced
LBC in their screening programs nationwide. Before a definite
decision can be made that LBC is better than conventional Pap
smears or at least will perform equally it is necessary to obtain
large high-quality randomized, controlled, cost-effective studies
that also take into account the design of the screening programs
in individual countries.
Screening Programs for cervical cancer
A detection program for cervical cancer should not only be
directed toward detection of invasive cancer but should also
include detection of severe epithelial abnormalities that can be
considered potential precursors of invasive cancer. Sensitivity
figures should therefore not only encompass invasive cancer, as
in most reports on the accuracy of screening, but should also be
calculated for a spectrum of severe epithelial abnormalities con-
sidered as potential precursors. In most studies, the screening
history of only those women with invasive cancer is reviewed.
(See Chapter 3, Cytologic Screening Programs, for definitions of
sensitivity, specificity, and predictive value.)
To determine with maximum possible accuracy the sensi-
tivity of the cytologic diagnosis, however, the total number of
women from the same population and in the same time period
who were cytologically tested and found to have severe epithe-
lial abnormalities (severe intraepithelial neoplasia and invasive
cancer), or "true-positives," should be known as well. Because
it is clearly impossible to histologically examine each cervix
from each woman who has been screened, sensitivity must be
calculated in an indirect way. All women with negative smears
must be monitored for a sufficiently long time. The longer
the follow-up, the more the detection of cases that were in an
undetectable, asymptomatic, preclinical phase at the time of
the test and that otherwise cannot be distinguished from the
truly missed cases that become evident during the follow-up
period.
Day suggested avoiding the most common inaccuracies in
calculating sensitivity by using only incidence rates.264 Inaccura-
cies result from the inclusion of cases that become apparent dur-
ing follow-up but that were still in an undetectable preclinical
stage at the time of the negative cervical screening test; cases can
be missed at the time of screening and not become evident dur-
ing the time of follow-up, and in some of the women the lesion
may regress during follow-up. For cervical cancer, Day's method,
unfortunately, is of limited value, because incidence rates in the
absence of screening must also be known. The latter situation is,
however, fortunately becoming increasingly rare. Cervical cyto-
diagnosis has been the subject of many investigations evaluating
the quality of this method as a diagnostic test.184 In the literature,
false-negative rates in cervical cytology vary widely. A number of
studies have shown rates varying from 15 to 55% in the presence
of invasive cervical cancer265,266 and from 6 to 45% in the pres-
ence of squamous cell carcinoma in situ.126,206,266-268
Cecchini and co-workers calculated the protective interval
for CIN grade 3 to be 3 years.269 No differences in the risk for
203
previous page 204 ComprehensiveCytopathology 1104p 2008 read online next page 206 ComprehensiveCytopathology 1104p 2008 read online Home Toggle text on/off