Ovarian cancer accounts for
only 3% of female cancers but
is the fi fth leading cause of all
cancer-related deaths in American
women. A total of 15,520 women
died from ovarian cancer in the United
States in 2008.1 Nearly two-thirds are diagnosed
at an advanced stage (III and IV),
with a 5-year survival of approximately
33%. In contrast, the 5-year survival of
women with cancers confi ned to the ovary
is approximately 90%. This disparity
motivates research into screening for
early-stage ovarian cancer. An effective
screening tool that reduces the mortality
from ovarian cancer remains elusive.
back to top
Screening Tests
For Ovarian Cancer
The concepts of predictive value, sensitivity,
and specifi city are critical in determining
the usefulness of screening tests. "Sensitivity"
is the term used to describe the
proportion of patients with ovarian cancer
who will have a positive test, and "specificity" is a measurement of the proportion
of patients without ovarian cancer who
return a negative test result. The positive
predictive value (PPV) of a test refl ects the
probability that a patient who has a positive
test result truly has ovarian cancer.
The strength of a test's PPV is directly
related to the prevalence of the disease in
the screened population. Therefore, given
the relatively low prevalence of ovarian
cancer in the general population, it is diffi
cult for even a very sensitive or specifi c
tool to return an acceptable PPV. Most
epidemiologists regard 10% as the minimum
PPV to support a screening test.2
With these parameters, one woman out of
10 with a positive test will have ovarian
cancer, while the remaining 9 will be exposed
to surgical risks, costs, and recovery
in the absence of ovarian cancer.
back to top
Ultrasonography
Transvaginal ultrasound is used to evaluate
the size and structure of a woman's
ovaries. Healthy asymptomatic women
have a defined normal upper limit of
ovarian volume.3 Architectural features
that can be used to discriminate benign
from malignant ovarian neoplasm include
cyst wall septations greater than 3
mm and cyst papillations or solid components.
4 When ovarian morphology
indices using ovarian size and architecture
features have been retrospectively
applied, their reported sensitivity is
between 89% and 100%, with a specifi
city ranging between 70% and 83%.4,5
These values come from symptomatic
populations with increased baseline
prevalence and achieve higher PPVs
than would be expected in the general
population. When average-risk women
were screened with ultrasound, ovarian
cancer was confirmed in 1.3% to 8% of
women with a positive test.6-8
back to top
Cancer Antigen–125
Serum tumor markers are attractive tools
for ovarian cancer screening because
they are widely available, noninvasive,
and objective. Cancer antigen (CA)-125
is a serum protein elevated in approximately
80% of women with advancedstage
ovarian cancer but only 1% to 2%
of the normal population.9 Unfortunately,
this protein is within normal range
for more than half of women with stage
I ovarian cancer.10 The specifi city for this
test is poor, because a number of benign
conditions may result in elevated CA-
125 levels.
CA-125 has superior specificity in
postmenopausal women compared with
premenopausal women (98.5% and
94.5%, respectively).11 The examination
of CA-125 values over time improves the
estimation of ovarian cancer risk.12 When
33,621 CA-125 samples from 9,233 women
were retrospectively assessed, an increasing
trend in CA-125 was associated
with a sensitivity of 86% for the preclinical
detection of ovarian cancer (as compared
with 62% sensitivity when only a
single value of CA-125 was considered).13
This measure, called the risk of ovarian
cancer (ROC), was confi rmed to have a
high PPV (19%) in a subsequent prospective
study of postmenopausal women.14
back to top
Alternative Tumor Markers
Other tumor markers have been analyzed
in combination with CA-125 with the goal
of improving sensitivity, specifi city, and
PPV. The interpretation of tumor marker
panels relies on the pattern of their levels
in relationship to each other, rather
than the absolute levels of each marker.
Over 30 different serum tumor markers
have been evaluated in combination with
CA-125.15 In general, sensitivity has been
increased by 5% to 10%, but with an associated
decline in specificity.
OVA1™ is a serum test that evaluates
5 biomarkers (transthyretin, apolipoprotein
A-1, β2 microglobulin, transferrin,
and CA-125 II) in women with an
ultrasound-confi rmed adnexal mass who
are intended for surgery, with results indicating
the probability of malignancy.16
This test is FDA approved for use only in
combination with positive imaging (such
as ultrasound) and clinical evaluation to
determine the need for preoperative subspecialist
referral.
Human epididymis specific protein 4
(HE4) is a müllerian-derived glycoprotein
whose secretion is highly restricted
in normal tissues.17 It is FDA approved for
use in surveillance of recurrence or progression
in women with an established
diagnosis of epithelial ovarian cancer.
Neither of these tests are recommended
or approved for use in the screening of
asymptomatic women.
back to top
Symptoms Survey
In a national ovarian cancer survey, approximately
95% of women with ovarian cancer reported that they had symptoms
preceding their diagnosis (particularly
abdominal pain, bloating, early satiety,
and bladder or bowel dysfunction).18 A
symptom index has been created that
incorporates these commonly reported
experiences.19 This index has also been
described in combination with CA-125
measurement and transvaginal ultrasonography.
20,21
These studies failed to show an improvement
in specificity or accuracy for
the combined approach, possibly because
it was studied in high-risk women
(who may be more likely to overreport
symptoms) or, in the case of ultrasound,
applied retrospectively to those with
confirmed ovarian cancer, in which case
ultrasound alone proved to be most accurate.
There are ongoing studies to prospectively
evaluate the role of the symptom
index when applied for screening in
average-risk populations.
back to top
Combined Modality
Screening and RCTs
Ultrasonography, measurement of tumor
markers, and patient-reported symptom
indices in isolation have not demonstrated
the sensitivity, specifi city, and
PPV necessary to justify screening for
ovarian cancer in average-risk populations
of postmenopausal women. The
use of combined modality screening
(age-stratified CA-125 levels and subsequent
transvaginal ultrasound) has
been named the ROC algorithm. The
use of this algorithm in a prospective,
multicenter screening study of 3,238
postmenopausal, average-risk women
progressed only 8 women to surgery and
showed good positive predictive value
(37.5%), with 3 early-stage invasive cancers
found.22 Long-term survival data
from this group of women are not yet
available.
Large prospective randomized controlled
trials (RCTs) in both the United
States and United Kingdom are reviewing
combined modality assessment for
screening. The Prostate, Lung, Colorectal,
and Ovarian Cancer Screening Trial
is a US-based RCT that has enrolled
34,261 healthy, average-risk women ages
55 to 74 to receive either annual CA-125
testing plus transvaginal ultrasonography
or "usual care."7 The study is now
in follow-up phase. Ovarian cancer was
found in 5% of women undergoing surgery
following a positive test result, with
the majority of the cancers found in advanced
stages. The PPV of this combination
testing in the average-risk population
was 1.3%. Long-term follow-up is
not yet complete, and final results are
anticipated in 2014.
The UK Collaborative Trial of Ovarian
Cancer Screening is following 202,638
average-risk, postmenopausal women
who were randomized to no treatment,
multimodality screening (ultrasound and
CA-125), or yearly ultrasound screening.
23 At 3 years follow-up, multimodality
screening was associated with a 35.1%
PPV, compared to 2.8% for ultrasound
screening alone. Of the invasive cancers
identifi ed, 48.3% were stage I/II. The increased
detection of ovarian cancer in
this study has yet to be associated with
improved overall survival.
back to top
Guidelines
Routine screening of the general, average-
risk population for ovarian cancer is
not recommended by any professional
society.24,25
Conclusion
Ovarian cancer is a serious disease
made more challenging by our limited
ability to achieve diagnosis at an early,
curable stage. Routine screening of
average-risk, postmenopausal women
with current technologies is not recommended
at this time. However, women
who have a defi ned genetic predisposition
to ovarian cancer may benefit from
regular screening ultrasonography
and CA-125 evaluations. We await the
overall survival data from large RCTs
of screening average-risk women and
data from evolving research into new
serum markers. Clinicians and patients
should pay close attention to symptoms
and family history, reserving testing
for diagnostic rather than screening
purposes.
The authors report no actual or potential
confl icts of interest in relation to this
article.
back to top
Emma Rossi, MD, is a clinical fellow in Gynecologic
Oncology, Department of Obstetrics and Gynecology,
University of North Carolina at Chapel Hill.
Daniel L. Clarke-Pearson, MD, is Chair, Department
of Obstetrics and Gynecology, and Robert A. Ross
Professor, University of North Carolina at Chapel Hill
School of Medicine.