82 Editorial | JNCI Vol. 104, Issue 2 | January 18, 2012
Advance Access publication on January 13, 2012.
© The Author 2012. Published by Oxford University Press. All rights reserved.
For Permissions, please e-mail: firstname.lastname@example.org.
For decades, investigators have sought a strategy for finding ovarian
cancer early enough to reduce the risk of dying of ovarian cancer.
In this issue of the Journal, Lim et al. (1) report on their study in
which women answered a dozen simple questions such as whether
they felt pelvic or abdominal pain in the recent past and, if so, when,
how frequently, and how severely. Symptom indices like this are
being promoted as easier or better ways to find ovarian cancer early,
under the assumption that early therapy can achieve a better outcome
than if the women wait until more or worse symptoms prompt
them to see a physician. The study revealed that symptom indices
as ovarian cancer screeners can be sensitive to the presence of
cancer in the period between 3 and 14 months before clinical diag-
nosis. On the other hand, they were not highly specific because a
substantial fraction of women without disease registered as “positive.”
Furthermore, these attractively simple screeners showed a dis-
appointing capacity to find cancer early, largely because the
symptom(s) arose close to the time of diagnosis. The study design
permits no calculation of years of life that might have been saved or
lost if such screeners actually were used—only a large and expensive
randomized trial would do that—but clinical gains likely would be
minor, and many women would undergo unnecessary diagnostic
procedures to assure that they are cancer free.
This sobering news follows hard on the heels of a large US
randomized trial finding no benefit, and indeed some harm, to
women who were screened annually with a transvaginal ultrasound
exam and a CA-125 blood test compared with a usual care control
group (2). It is plausible that other markers could perform better,
but here too, the results so far suggest caution. One serum marker,
HE4, performs better than does ultrasound as a second-line screen
following CA-125 (3), though it also suffers from bad timing: It
spikes relatively late in the year before diagnosis (4). Furthermore,
panels of other markers, some of which are in commercial use,
did quite well in initial reports but completely failed as a screening
method in confirmatory studies (5).
Why is the development of a sensitive and specific screening
method to detect ovarian cancer early so hard? The biology of
ovarian cancer, the arithmetic of screening, and the clinical charac-
teristics of the disease and its treatment collude to make it difficult
to find ovarian cancer early enough to matter. For instance, we now
suspect that many ovarian cancers arise in the fallopian tubes. The
cancer may stay there for months or years before either spreading or
migrating to the ovaries, with unknown effects on serum levels of
any biomarkers. Fortunately, ovarian cancer occurs only rarely, but
it follows that even a small fraction of false-positive test results will
produce many dangerous and costly follow-up procedures.
Most importantly, the clinical characteristics of ovarian cancer
pose some simple but profound challenges. Because of the location
of the ovaries deep in the pelvis, symptoms seldom announce the
presence of disease when it is still in its earliest stages. The symp-
toms used to derive the indices in this observational study by Lim
et al. (1) are most often associated with more advanced stage III
and IV disease, and overall, 68% of cancers are stage III or IV at
presentation (6). Bowel or urinary symptoms or bloating related to
malignant ascites or a palpable mass all suggest a substantial tumor
burden. At times, a large mass that is still confined to one ovary can
cause these symptoms and will still be amenable to primary sur-
gical removal. However, this scenario represents only a minority of
If a 1-cm mass contains approximately 109 cells, it will often
take several times that mass to actually cause clinical symptoms.
Although it is true that the ability to surgically debulk ovarian
cancer to a residual nodule of less than 1 cm—a tumor size that is
associated with better responses to chemotherapy and overall long-
term outcomes—the underlying biology of the disease may well
be more important. Tumor grade is an important variable in
determining ability to maximally debulk the tumor and response to
The potential for poor outcomes even with small tumors is seen in
other malignancies. In breast cancer, very small tumors with four or
more lymph nodes involved have a worse prognosis compared with
larger node-negative tumors that would be more likely to present with
symptoms (7). In aggressive triple-negative breast cancer, once the
cells metastasize to lymph nodes, the number of affected nodes is no
longer prognostic (8,9). It appears that the underlying tumor biology
is more prognostic than tumor size or even nodal involvement. In
prostate cancer, early results from the PIVOT trial suggest that
patients with a favorable prostate-specific antigen level and Gleason
score do as well with careful observation as they do with immediate
prostatectomy (10). Again, the underlying biology of the tumor,
rather than its presence or size, seems to predict outcome.
What does this mean for ovarian cancer? Will an earlier detec-
tion of a few months and intervention help most patients? Patients
already have a substantial tumor burden by the time they are symp-
tomatic. Is it all dependent on the size and location of tumor
spread or is it the kind of cancer cells that have determined that the
cells will spread, how they will respond to treatment, and how the
patient will eventually do?
For clinicians who employ a symptom index for screening, the
report by Lim et al. (1) suggests that further revisions of the ques-
tions and words would have only a minor impact on their ability to
distinguish the small number of women with silent ovarian cancer
from the vast number of cancer-free women. The study by Lim
et al. also shows that women respond to questions somewhat dif-
ferently in person compared with by telephone. The key clinical
Finding Ovarian Cancer
Patricia Hartge, James L. Speyer
Correspondence to: Patricia Hartge, ScD, Epidemiology and Biostatistics Program, Division of Cancer Epidemiology and Genetics, National Cancer
Institute, 6120 Executive Blvd, EPS 8090, Bethesda, MD 20892 (e-mail: email@example.com).
JNCI | Editorial 83
question concerns what steps to take next if a woman scores
high on the index. The relation between the symptom index and
CA-125 and HE4 warrants investigation. For example, the most
optimistic scenario for future designs of a screening program holds
that the right combination of questionnaires, blood tests, and
imaging will perform far better in finding the right women for
diagnostic evaluation than do any of the separate components.
If that combination can be found, the next question is whether the
resulting earlier treatments will prolong life.
No one would advocate delayed action after symptoms indi-
cate possible pathology. However, detecting ovarian cancer at an
earlier time but when the patient is already symptomatic is unlikely
to make a major difference in outcome. Should we be focusing
our efforts on the time between when a tumor is large enough to
cause mild symptoms and when it is large enough to cause symp-
toms that prompt women to call a physician? Or should we
work harder to understand the etiology of the disease, to find
methods of prevention, and to obtain better therapy for women
with biologically aggressive disease? These other research direc-
tions remain critical as the search continues for better ways to
find ovarian cancer early.
1. Lim AWW, Mesher D, Gentry-Maharaj A, et al. Predictive value of symp-
toms for ovarian cancer: comparison of symptoms reported by question-
naire, interview, and general practitioner Notes. J Natl Cancer Inst. 2012;
2. Buys SS, Partridge E, Black A, et al. Effect of screening on ovarian
cancer mortality: the Prostate, Lung, Colorectal and Ovarian (PLCO)
Cancer Screening Randomized Controlled Trial. JAMA. 2011;305(22):
3. Urban N, Thorpe JD, Bergan L, et al. Potential role of HE4 in multi-
modal screening for epithelial ovarian cancer. J Natl Cancer Inst. 2011;
4. Anderson GL, McIntosh M, Wu L, et al. Assessing lead time of selected
ovarian cancer biomarkers: a nested case-control study. J Natl Cancer Inst.
5. Cramer DW, Bast RC Jr, Berg CD, et al. Ovarian cancer biomarker per-
formance in prostate, lung, colorectal, and ovarian cancer screening trial
specimens. Cancer Prev Res. 2011;4(3):365–374.
6. Howlader N, Noone AM, Krapcho M, et al., eds. SEER Cancer Statistics
Review, 1975–2008. National Cancer Institute, Bethesda, MD. http://seer.
cancer.gov/csr/1975_2008/. Based on November 2010 SEER data submis-
sion, posted to the SEER web site, 2011.
7. Wo JY, Chen K, Neville B, et al. Effect of very small tumor size on cancer-
specific mortality in node-positive breast cancer. J Clin Oncol. 2011;29(19):
8. Hernandez-Aya L, Chavez-MacGregor M, Lei X, et al. Nodal status and
clinical outcomes in a large cohort of patients with triple negative breast
cancer. J Clin Oncol. 2011;29(19):2629–2634.
9. Comen E, Norton L, Massagué J. Breast cancer tumor size, nodal status,
and prognosis: biology trumps anatomy. J Clin Oncol. 2011;29(19):
10. Phillips C. Study questions benefit of surgery in some men with early-stage
prostate cancer. NCI Cancer Bulletin. 2011;8(11):2.
This work was supported by the Intramural Research Program of the Division
of Cancer Epidemiology and Genetics of the National Cancer Institute,
National Institutes of Health.
The authors have no conflicts of interest to disclose. The content is solely
the responsibility of the authors and does not necessarily represent the of-
ficial views of the National Cancer Institute or the National Institutes of
Affiliations of authors: Epidemiology and Biostatistics Program, Division of
Cancer Epidemiology and Genetics, National Cancer Institute, Bethesda,
MD (PH); Department of Medicine, New York University Langone Medical
Center, Clinical Cancer Center, New York University Cancer Institute, New
York, NY (JLS).