"A solid triple but not a home run” is how Karen H. Lu, MD, characterized a study in The Lancet reporting a reduction in deaths from ovarian cancer with the use of multimodal ovarian cancer screening.¹ Dr. Lu’s remark was one of several, mostly but not universally, favorable and optimistic comments included in an article on the study in The New York Times.² In a follow-up interview with The ASCO Post, Dr. Lu, Chair of the Department of Gynecologic Oncology and Reproductive Medicine, Division of Surgery, at The University of Texas MD Anderson Cancer Center, Houston, explained what it would take to make it to home plate.

The authors of The Lancet study acknowledged that they need 3 additional years to allow the mortality data to mature, Dr. Lu noted. “So in 3 years, we should have a pretty firm answer about whether ovarian cancer screening reduces death from the disease—which would be a home run.”

One of the strengths of the study is its large scale, with more than 200,000 participants. Dr. Lu said that there are many other strengths as well. “The reason it is so powerful,” she pointed out, “is that the goal was to reach a gold standard for determining whether screening will be effective in terms of public health, which is seeing fewer deaths from the disease because of screening.”

Three Study Groups

The study reported in The Lancet randomly assigned 202,638 postmenopausal women aged 50 to 74 from 13 centers in the National Health Service Trusts in England, Wales, and Northern Ireland to multimodal screening, transvaginal ultrasound screening, or no screening in a 1:1:2 ratio. The multimodal screening used serum CA-125 testing interpreted over time by a patented Risk of Ovarian Cancer algorithm, with women triaged to one of three risk groups: normal (annual screening), intermediate (repeat CA-125 testing in 3 months), and elevated (repeat CA-125 testing and transvaginal ultrasound screening as a second-line test in 6 weeks). Dr. Lu is principal investigator of a single-arm study based on the same screening algorithm.

Results in the transvaginal ultrasound–screening groups were classified as normal (annual screening), unsatisfactory (scan repeated in 3 months), or abnormal (scan repeated within 6 weeks by a senior ultrasonographer). In both screening groups, women with persistent abnormalities had clinical assessment and additional tests by a trial clinician.

Women in the multimodal- and ultrasound-screening groups were compared separately with women in the no-screening group. At a median follow-up of 11.1 years, ovarian cancer was diagnosed in 1,282 (0.6%) women: 338 (0.7%) in the multimodal-screening group, 314 (0.6%) in the ultrasound-screening group, and 630 (0.6%) in the no-screening group. Among these women, deaths due to ovarian cancer occurred in 148 (0.29%) in the multimodal-screening group, 154 (0.30%) in the ultrasound-screening group, and 347 (0.34%) in the no-screening group.

Primary and Secondary Analyses

The relative mortality reductions of 15% in the multimodal- and 11% in the ultrasound-screening groups were not significant in the primary analysis. “In retrospect, it would have been preferable to specify a primary analysis that was weighted to reflect the predictable delay in mortality reduction in a screening trial of this type,” the authors stated. A prespecified secondary subgroup analysis excluding women who had an estimated change point in their CA-125 profile before randomization (prevalent cases) did find a significant mortality reduction in the multimodal-screening group, with an overall reduction in mortality of 20%, and 28% after 7 years of screening.

Commenting on the differing primary and subgroup analyses, Dr. Lu said that the study investigators “have done a tremendous job in carrying out and interpreting the study. I think that the results were very balanced, but we need to wait a little bit longer, and 3 years is a very short amount of time. Then we will hopefully have an answer.”

Roll Out in the United States

Results from a single-arm study at MD Anderson and six other sites using the same Risk of Ovarian Cancer algorithm has shown that the multimodal screening “really is quite practical,” Dr. Lu noted. “The vast majority of women—over 90%—just need to come back once a year for a blood test, and they could get that at their annual exam. It is a very small percentage, really less than 1% to 2%, that need to go on to have an ultrasound,” she said. A report on the first 11 years of that study, published in Cancer in 2013,³ concluded the screening strategy “demonstrated excellent specificity and positive predictive value” in a population of U.S. women at average risk for ovarian cancer but could not be considered practice-changing at that time.

“One of the things that we wanted to potentially be prepared for is if the UK study is positive, could this roll out in the United States, where the health-care system is less centralized? Would we get the same results, given that it is a much less coordinated health-care system? We’ve got to figure out how to deliver this type of personalized screening strategy to assure that our outcomes will be the same. That is one of the reasons we are conducting this study, to see if we can do that,” Dr. Lu said. “And we are doing that in multiple locations. It is going on simultaneously in New Jersey, Rhode Island, Miami, and Houston, and so far we have had similar findings.”

Balancing Benefit With Harm

In both the single-arm study led by Dr. Lu and the three-arm study reported in The Lancet, the multimodal-screening strategy is limited to postmenopausal women and should remain so, according Dr. Lu. “We know that ovarian cancer is a relatively rare cancer and that the incidence increases with age. In women under 50, the likelihood of ovarian cancer is quite low. We are always balancing benefit, which is detecting the cancer early, with harm, which is unnecessary surgery. That ratio will change when you start to screen women in whom ovarian cancer is less common,” Dr. Lu explained.

“If you look at the big picture, the vast majority of women who present to their physicians, present with very late-stage disease, and we have not been able to change that paradigm. So this would be a very important step forward. There have been incremental improvements in treatment over the past 20 years, but if we are going to make a home run, we really need an early-detection strategy,” Dr. Lu stated.

“Certainly there are women between the ages of 40 and 50 in whom early detection is important as well, but we need to be a little more cautious,” she said, about extending screening strategies to these women. They would include younger women considered at higher risk of ovarian cancer because of family history and BRCA mutations. “But I am also a very strong believer that in women with a family history of ovarian cancer, or of breast cancer at a young age, genetic testing is a powerful way to identify whether they are at high risk for ovarian cancer,” Dr. Lu stated.

“For women with a BRCA mutation, their risk for ovarian cancer can be as high as 40% to 50%. In the average population, lifetime risk is about 1% or 1.5%. There is a huge difference between lifetime risks of 1.5% and 40%, and that is why we need to identify those at high risk,” Dr. Lu said. “It is a small number of individuals, but if we can identify those individuals who have a BRCA mutation, then we can really make an impact in terms of risk-reducing surgery.”

Very Low Complication Rates

“Both venipuncture and transvaginal ultrasound used in the multimodal screening were associated with minor complications and very low complication rates,” the study authors reported. “Most postmenopausal women found transvaginal ultrasound screening acceptable, with 3.5% reporting moderate or severe pain during the scan.”

Dr. Lu agreed that the complications were “very minor. I would say that the main issues are: (1) We always want to be cautious about having too many unnecessary surgeries, and we haven’t really seen that. So that’s been important. And: (2) Some women naturally have a higher CA-125. They just run high, and it can cause anxiety, but it may not actually be ovarian cancer.” In general, however, neither the study cited in The Lancet nor the single-arm study led by Dr. Lu has shown an increase in anxiety among the women screened.

The authors of The Lancet article reported 14 false-positive surgeries per 10,000 screens in the multimodal-screening group and 50 false-positive surgeries in the ultrasound group. Dr. Lu said that is consistent with what is normally seen but pointed out: “When we say unnecessary surgery in our study, these were still women who had ovarian masses. It’s just that they turned out not to be cancer. We don’t know whether they ultimately would have developed symptoms or more worrisome disease.” ■

Disclosure: Dr. Lu Reported no potential conflicts of interest.

References

1. Jacobs IJ, Menon U, Ryan A, et al: Ovarian cancer screening and mortality in the UK Collaborative Trial of Ovarian Cancer Screening (UKCTOCS): A randomised controlled trial. Lancet. December 17, 2015 (early release online).

2. Grady D: Early detection of ovarian cancer may become possible. The New York Times, December 17, 2015.

3. Lu KH, Skates A, Hernandez MA, et al: A 2-stage ovarian cancer screening strategy using the Risk of Ovarian Cancer Algorithm (ROCA) identifies early-stage incident cancers and demonstrates high positive predictive value. Cancer 119:3454-3461, 2013.