Should all women at high risk for ovarian cancer undergo screening on a routine basis?

July 14, 2010

Citing different studies, two physicians reach different conclusions as to whether ovarian cancer screening is worthwhile.

ABSTRACT: Citing different studies, two physicians reach different conclusions as to whether ovarian cancer screening is worthwhile.

The National Comprehensive Cancer Network currently recommends ovarian screening for women at high risk. NCCN guidelines state that screening should be done with ultrasound, transvaginal ultrasonography, and CA125 serum marker testing every six months starting at age 35 or, depending on the family history, five to 10 years before the earliest diagnosis of ovarian cancer.

So if recommendations are in place, and high-risk women face a greater risk of dying from ovarian disease, why is there any argument about whether screening has value in these women? One issue is that, in the general population, ovarian cancer is relatively uncommon, and most data on ovarian cancer disease progression come from this larger pool of women. Also, the current tests for early detection are simply not sensitive enough to be used as screening tools in a general population.

Pro
Presented by DR. BETH KARLAN

Screening high-risk women could downstage them from being stage III and IV patients to being stage I and II patients, thereby improving the chance of long-term survival.
Data from the 22,000-woman pilot study demonstrated that screening could improve survival compared with a control group.
Data from UKCTOCS showed that multimodality screening meets the minimum standard of a 10% PPV.
Numerous biomarkers have been identified as potential candidates for early diagnosis.

Nonetheless, high-risk women should be judged by a different set of rules. Screening in this population could effectively lower these women's risk of having late-stage disease, said Beth Karlan, MD, arguing in favor of routine screening. Taking the opposite stance was Michael Birrer, MD, PhD, who emphasized that the current understanding of ovarian cancer and the tools for early detection do not meet the accepted criteria for sound screening.

Dr. Karlan is the Board of Governors Endowed Chair of Gynecologic Oncology and director of the Women's Cancer Research Institute at the Samuel Oschin Cancer Institute at Cedars-Sinai Medical Center in Los Angeles. She also is director of the division of gynecologic oncology at the Gilda Radner Cancer Detection Program at Cedars-Sinai. She is a professor of obstetrics and gynecology at the Geffen School of Medicine, University of California, Los Angeles.

Dr. Birrer is the director of gynecologic medical oncology at Boston's Massachusetts General Hospital Cancer Center. He is also co-chair of the National Cancer Institute's Gynecologic Cancer Steering Committee and chair of the committee on experimental medicine for the Gynecologic Oncology Group (GOG).

Current screening techniques are good enough

Women who have inherited a deleterious mutation (BRCA) are at a very high risk for ovarian cancer, more than an order of magnitude greater than the general population, Dr. Karlan said. When these high-risk women do present with disease, it is often metastatic and widespread, putting them at an equally high risk of dying from ovarian cancer.

"If we had a means of finding the ovarian cancer earlier, we could move a portion of those stage III and IV patients to stage I and II, and we could significantly improve the survival of women with ovarian cancer, even without any new therapies," Dr. Karlan said. "Today, survival of stage I ovarian cancer already exceeds 90% so ovarian cancer screening makes sense. It's a highly lethal disease that when caught early can be cured."

Dr. Karlan cited data from a pilot trial of 22,000 women who where randomized either to a control group or to an intervention group. The latter were offered three annual screening exams with CA125 testing; pelvic ultrasonography if CA125 was 30 U/mL or more; and referral for further examination if the ovarian volume was 8.8 mL or more on ultrasonography. All of the women were followed to see whether they developed invasive epithelial cancers of the ovary or fallopian tube (index cancers).

According to the results, 29 of the 468 women in the screened group with an elevated CA125 level were referred for a gynecologic opinion; screening detected an index cancer in six, and 23 had false-positive screening results. The positive predictive value was 20.7%. During a seven-year follow up, ten women with index cancers were identified in the screened group and 20 in the control group. The median survival of women with index cancers in the screened group was 72.9 months vs 41.8 months in the control group (P = .0112). However, the number of deaths from an index cancer did not differ significantly between the control and screened groups: 18 of 10,977 vs nine of 10,958 (Lancet 353:1207-1210, 1999).

Dr. Karlan acknowledged that "this was a pilot study and it was not powered to determine a difference in mortality. However, if you looked at the data, those in the screening group had survival that was almost twice that of the control group."

TABLE 1 UKCTOCS screening arms

Multimodal screening (MMS) arm consisting of annual CA125 screening interpreted using a risk of ovarian cancer algorithm plus transvaginal ultrasound as a second-line test.
Annual screening with transvaginal ultrasound (USS) alone in a 2:1:1 ratio using a computer-generated random number algorithm.

Data from that pilot study was used to fund and launch the randomized UK Collaborative Trial of Ovarian Cancer Screening (UKCTOCS), which will assess the effect of screening on mortality. In this trial that began in 2001, a total of 202,638 postmenopausal women (age 50-74 years) were randomly assigned to a control arm of no screening (n = 101,359) or one of two screening arms (see Table 1).

Women randomized to the multimodality screening (MMS) group had their blood tested for CA125 and those randomized to the ultrasound screening (USS) group were sent for transvaginal ultrasonography (TVUS). Women with abnormal screens had repeat tests, and women with persistent abnormality on repeat screens underwent clinical evaluation and, when appropriate, surgery.

Overall, 4,355 of 50,078 women (8.7%) in the MMS group and 5,779 of 48,230 women (12%) in the USS group required a repeat test, and 167 (0.3%) in the MMS group and 1,894 (3.9%) in the USS group required clinical evaluation. Ninety-seven of 50,078 women (0.2%) from the MMS group and 845 of 48,230 (1.8%) from the USS group underwent surgery. A total of 42 (MMS) and 45 (USS) primary ovarian and tubal cancers were detected, including 28 borderline tumors. Twenty-eight of the 58 invasive cancers (48.3%) were stage I/II. An additional 13 women (five on MMS, eight on USS) developed primary ovarian cancer during the year after the screen (Lancet Oncol 10:327-340, 2009). Unlike the pilot study, which used a single CA125 cutoff value for prompting ultrasound, this trial relied on the change in CA125 over time to identify women who should continue on with ultrasound testing, Dr. Karlan explained. Mortality data from the trial will be available in a few years and the data reported last year were not longitudinal, "but I think these are very hopeful data for the applicability of [screening] to even the general population. Over half of the cases of invasive cancer detected were stage I or stage II. They were not only able to find cases but also to find them early when we could actually have an impact on survival."

TABLE 2 UKCTOCS statistical measures

MultimodalTransvaginal ultrasound (USS)
Sensitivity89.4%84.9%
Specificity99.8%98.2%
PPV43.3%5.3%
Sensitivity89.5%75%
Specificity99.8%98.2%
PPV35.1%2.8%

In addition, experts have agreed that 10% would be the minimum acceptable positive predictive value (PPV) for a screening test, Dr. Karlan said. The UK trial has demonstrated that an MMS approach can reach, if not exceed, that minimum, she added (see Table 2).

Dr. Karlan than changed gears and highlighted results from a recent mathematical modeling study that looked at the preclinical natural history of serous ovarian cancer in order to define the target for early detection. Patrick O. Brown, MD, PhD, and Chana Palmer developed models for the growth, progression, and detection of occult serous cancers by analyzing published data on serous cancers in BRCA1 mutation carriers (PLoS Med 6:e1000114, 2009).

They came to several noteworthy conclusions:

Serous ovarian cancers spend, on average, more than four years as in situ stage I/II cancers and approximately one year as stage III/IV cancers before becoming clinically apparent.
During the occult period, serous ovarian cancers are less than 1 cm in diameter and not visible on gross examination of the ovaries and fallopian tubes.
The median diameter of a serous ovarian cancer when it progresses to an advanced stage is about 3 cm.

Based on these findings, in order to achieve 50% sensitivity in detecting tumors before advancement to stage III, an annual screen must detect tumors of 1.3 cm in diameter. An 80% sensitivity would require the detection of tumors that are less than 0.4 cm in diameter. And to achieve a 50% reduction in serous ovarian cancer mortality with annual screening, a test would need to detect tumors of 0.5 cm in diameter.

Dr. Karlan called the results "very exciting data that have really swung the screening window open. How much time do we have to find early ovarian cancer? These data really demonstrate that we may have more than four years, so even an annual screening exam may be able to find these early cases."

Another study has indicated that CA125 may find preinvasive cancers. Researchers in the Netherlands investigated whether CA125 concentrations were indicative of adnexal dysplasia and cancer in women at hereditary high risk of developing ovarian and/or tubal cancer (J Clin Oncol 25:1383-1389, 2007).

The investigators reported that the absolute value (P < .0001) and the serial change (P < .0001) of CA125 were predictive for ovarian cancer. For adnexal dysplasia, the absolute value of CA125 (P = .003) was predictive although the serial change in CA125 was not (P = .32). They found an odds ratio of six for adnexal dysplasia vs nondysplasia in women with the highest levels of CA125 (14 U/mL) compared with those with the lowest levels of CA125 (< 10 U/mL).

"The absolute value of CA125 was able to predict ovarian cancer, and was even able to predict adnexal dysplasia," Dr. Karlan said. "This is sounding pretty good. . .we may have a blood test that when used as a rate of rise based on longitudinal change may be able to find early cancers within a four-year screening period."

Finally, Dr. Karlan pointed out that many biomarkers have been identified as potential candidates for early diagnosis. "We have many methods now to prioritize the markers, to put them together as a composite marker, and to look at the marker's rate of rise. We also have biorepositories and/or tissue banks to validate the markers," she said. "These high-risk women can't wait any longer. . .we actually have tests that can find the tumors earlier."

Failing to meet the minimum of screening standards

TABLE 3 WHO criteria for screening tests

The disease should be an important health problem.
There should be a treatment for the condition.
There should be a latent stage of the disease.
There should be a test or examination for the condition.
The test should be widely accepted.
The natural history of the disease should be adequately understood.
The total cost of finding a case of the disease should be economically balanced in relation to medical expenditure as a whole.

For Dr. Birrer, it was not clear that ovarian cancer screening has met the basic criteria set out for any screening method. He pointed out that the World Health Organization has several principles a screening method must meet to be considered effective (see Table 3).

While ovarian cancer is no doubt a serious health problem and treatment is certainly available, Dr. Birrer questioned whether the natural history of the disease is readily understood. He also pointed out the lack of knowledge about latent disease and whether there really are widely accepted tests available. "I highlight those formal points because that's what is not clear to me," he said.

Parsing the current literature on the origins of the disease, Dr. Birrer suggested that the ovarian cancer community is still in the information-gathering stages. A 1996 paper compared the histologic features of the ovaries, removed for prophylaxis, of women at increased risk for cancer with those of women at no known increased risk. "Two unanticipated microscopic or near-microscopic malignant neoplasms, and other benign and borderline tumors, were discovered in the ovaries of the high-risk individuals," Dr. Birrer said.

In addition, the researchers pointed out that that among the ovaries of high-risk women, 85% presented with two or more and 75% presented with three or more of a variety of histologic features, including surface epithelial pseudostratification, surface papillomatosis, epithelial inclusion cysts, increased follicular activity, corpus luteum hyperplasia, and hilar cell hyperplasia (J Natl Cancer Inst 88:1810-1820, 1996).

But a 2000 study found no premalignant histologic, molecular, or cell biologic alterations in prophylactic oophorectomy specimens from BRCA1 heterozygotes. For this research, ovarian tissues from 18 BRCA1 heterozygotes and from 20 age-matched controls were examined in a blinded fashion for histologic evidence of surface epithelial pseudostratification, epithelial inclusion cysts, deep cortical invaginations of surface epithelium, increased stromal cell activity, and surface papillomatosis, the authors explained.

Although the researchers saw some histologic alterations, there was no difference in frequency between cases and controls. They also did not see any evidence of p53 overexpression in any ovarian tissue or any difference in epithelial cell proliferation or apoptosis between the two groups (Cancer 89:383-390, 2000).

"So there's some confusion," Dr. Birrer said. "That has led to the more recent observation in 2007 by David Kindelberger, MD, that enteropathy and carcinomas can be found not in the ovary but in the fingered ends of the tubes, raising the whole possibility that what we thought was ovarian cancer and we thought came from inclusion cysts on the surface of the ovary, actually comes from the tubes. This is still a work in progress. No one really knows whether these are the preneoplastic lesions for ovarian cancer. And we don't know how long it takes to go from there to the peritoneal cavity: Perhaps it takes weeks or can be measured in hours? We simply don't know the natural history of this disease" (Am J Surg Pathol 31:161-169, 2007).

Dr. Birrer quoted a paper by Dr. Karlan and colleagues stating that "for screening to be of value, it must, by definition, detect asymptomatic disease at an early stage" (Obstet Gynecol Clin N Am 34:641-665, 2007). He then outlined some of ways that current screening methods fall short.

He described CA125 as a "chronic underperformer" in ovarian cancer. The test is not very helpful in young women, he said. In addition, 50% of early-stage ovarian cancers have normal or nonelevated CA125. "And we have to remember too that there are a lot of false positives," he explained. "Anything that irritates the peritoneal cavity can elevate the CA125 and that's really a problem."

Con
Presented by DR. MICHAEL BIRRER

The natural history of ovarian cancer is not readily understood, including a lack of knowledge about latent disease.
CA125 testing is a "chronic underperformer." Half of early-stage ovarian cancers have normal or nonelevated CA125, and irritation of the peritoneal cavity can elevate the CA125, leading to false-positive results.
Numerous Dutch trials have demonstrated that screening with CA125 and TVUS still turns in an unimpressive early-stage detection rate.
Prophylactic BSO for BRCA1/2 carriers is the only effective strategy for reducing the risk of ovarian cancer.

CA125 testing is often supplemented with TVUS, which is accurate for detecting morphologic abnormalities, he said. Color flow Doppler and morphology indices make for sophisticated probability models, but the test is cumbersome, expensive, and still turns in low sensitivity and specificity, he said.

Dr. Birrer turned to published literature to illustrate why he thinks screening isn't ready for prime-time, starting with a study that he deemed one of the best examples of an ovarian screening study. In this study, 62 confirmed BRCA1/2 carriers were screened biannually with CA125 and TVUS.

At a median follow up of 17 months, there were 22 women with abnormal results. Of these 22, 12 had repeat tests that showed normal results, and 10 underwent surgery, revealing five cases of stage I/II ovarian/peritoneal cancer. The screening protocol turned in a sensitivity of 71% and a specificity of 91% (J Clin Oncol 20:1260-1208, 2002).

Dr. Birrer deemed the results "not bad, very encouraging," but the small patient population was a major limitation. Other studies have turned in more mixed results. A larger trial out of the UK enrolled 1,110 moderate-risk and high-risk women and screened them annually with CA125 and TVUS. The trial leaders reported that nine tumors were diagnosed with screening and two of those were stage I. The sensitivity of the screening protocol was 50%. The early-stage detection rate was 16.7%, which is "not very impressive when you think that 25% of all ovarian cancers are early stage," Dr. Birrer said (J Clin Oncol 23:5588-5596, 2005).

Then there are results from numerous studies out of the Netherlands. In one, the investigators concluded that monitoring 152 high-risk women with CA125 and TVUS did not prevent the diagnosis of advanced ovarian cancer. In this cohort, screening found four cancers with only one case of stage I disease, one case of stage III disease, and two cases of stage IV disease. CA125 and TVUS together turned in a sensitivity of 40%, a specificity of 99%, a positive predictive value (PPV) of 40%, and a negative predictive value (NPV) of 99% (Gynecol Oncol 100:20-26, 2006).

An 11-year study screened 113 BRCA mutation carriers and eight malignancies were found. But 83% of the invasive malignancies were detected at an advanced stage, Dr. Birrer said. "We know these patients are incurable. Twelve months of screening and the cancers that are being detected are at an advanced stage" (Int J Gynecol Cancer 16:545-549, supplement 1, 2006).

A study with a larger patient population enrolled 512 high-risk women of whom 265 women had a BRCA mutation. The median follow up was two years and, using annual screening with CA125 and TVUS, the authors found one stage IIIc cancer. "The conclusion to this paper: Surveillance programs that identify women with ovarian cancer are inefficient. . .in the prevention of ovarian cancer, prophylactic bilateral salpingo-oophorectomy (BSO) is a more efficient alternative," Dr. Birrer said (Br J Cancer 94:814-819, 2006).

More recently, a group at the University of Groningen screened 241 consecutive women with BRCA mutations who were enrolled in the surveillance program for hereditary ovarian cancer. Screening included annual pelvic examination, TVUS, and CA125 testing. The mean follow up was 1.8 years. The authors reported 46 cancers were found: 40 by screening and six interval cancers. Of those, 24 screen-detected and all six interval cancers were found at late stage. The screening protocol turned in a sensitivity of 40%, a specificity of 99%, a PPV of 40%, and an NPV of 99% (Int J Cancer 124:919-923, 2009).

"So [screening is] still missing 30 cancers out of this group, all of which are essentially incurable," Dr. Birrer said. "Conclusion: Current screening of women with BRCA1/2 mutations does not seem to be effective, and at this time prophylactic BSO for BRCA1/2 carriers is the only effective strategy for reducing the risk of ovarian cancer."

Dr. Birrer concluded by again quoting Dr. Karlan's study that screening for women at high risk should be considered for those patients who do not want to undergo prophylactic surgery (Obstet Gynecol Clin N Am 34:641-665, 2007).

"Patients must understand that these measures have not been conclusively proven to improve early detection or long-term survival," Dr. Birrer said. "We've already had close to 12 studies on this. Repeating these kinds of studies is not going to help."

Source: 2009 Oncology Congress.