A two-stage ovarian cancer screening strategy substantially improved early-stage detection among average-risk postmenopausal women, according to long-term follow-up from a large prospective study.
Combining serial CA125 values and conditional referral for transvaginal sonography (TVS) led to a sensitivity of 74% for detection of ovarian and borderline cancers. Cancers detected by the (ROCA) for CA125 analysis were stage I-II in 70% of cases. The strategy reduced detection of late-stage disease by 30-34% compared with historical controls.
After 21 years of follow-up, the screening strategy had a positive predictive value (PPV) of 50% for ovarian cancer and 74% for any cancer, far exceeding the 10% minimum accepted PPV study endpoint, reported Robert Bast, MD, of the University of Texas MD Anderson Cancer Center in Houston, and coauthors in the .
"We can't judge mortality because we don't have an appropriate control arm," Bast told . "What we can do is look at stage shift, and the stage shift we observed was pretty dramatic [>40%]."
Bast and colleagues had three additional biomarkers that can detect cancers missed by CA125 analysis. Researchers have developed a second algorithm that combines all four biomarkers, and initial evaluation of the new algorithm could begin in February.
Later this year, a 4-year National Cancer Institute-funded study will investigate the specificity of the four-biomarker algorithm.
"I think the hope long term is that if that study is promising, if we detect at least as many or more early-stage cases and if we don't have an inordinate number of ultrasounds or negative surgeries, that we could think about collaborating in larger trials that almost certainly are going to be looking at DNA biomarkers, and possibly look at whether we can actually show a mortality advantage," said Bast.
Basics of Two-Stage Screening
Initiated in 2001, the (NROSS) evaluated the two-stage screening strategy in almost 8,000 postmenopausal women. The participants had annual CA125 analyses by ROCA. If the value did not change and the ROCA risk remained <1:2,000, no further action was needed, and the women returned annually for CA125 assessment until 2011.
A ROCA risk >1:500 triggered immediate TVS. Women with an intermediate risk returned in 3 months for repeat CA125 analysis. A further risk increase prompted referral for TVS.
At the same time as NROSS, the U.K. Collaborative Trial of Ovarian Cancer Screening began a randomized evaluation of the two-stage strategy. From 2001 to 2005, investigators allocated 202,638 postmenopausal women 1:1:2 to the two-stage strategy, annual TVS, or no screening.
The initial report from UKCTOCS showed a 14% stage shift with two-step screening but no improvement in ovarian cancer mortality, the primary endpoint. A after 16.5 years of follow-up showed a statistically significant 20% annual reduction in the hazard for overall mortality with two-stage screening versus no screening. The overall reduction was driven by larger annual reductions during years 7-14.
An from NROSS after 11 years of follow-up showed a 40% PPV for detections of invasive ovarian cancer and a 99.9% specificity with the two-stage screening process.
Updated Results
Though encouraging, the two-stage screening strategy is not yet ready for clinical practice.
"We may need to wait another decade or two for a disruptive technology to come along before we even think about ovarian cancer screening," said Bast. "I think this study suggests that we should not give up on this strategy because it's adequately specific, and we have biomarkers that can improve on this [and] that have already been identified."
The present update showed that 2% of participants were referred annually for TVS as a result of elevated risk by ROCA analysis, and 95.5% of the women actually had TVS, which raised suspicion for cancer in 4.8% of cases. Ultimately, 34 patients were referred for surgery as a result of TVS findings.
The annual false-positive rate for follow-up CA125 analysis (intermediate risk) was 5.8%, for TVS (high risk) 2%, and 0.2% for surgery, resulting in a specificity of 92.3% for intermediate- and high-risk ROCA, 98% for elevated risk triggering TVS, and 99.8% for ROCA and TVS leading to surgery.
The 34 operations detected 15 invasive ovarian cancers and two borderline tumors, resulting in a 50% PPV. Additionally, 12 of 17 ROCA-detected were stage I or II, and five of the 12 were associated with CA125 <35 U/mL. The five late-stage cancers consisted of three high-grade serous, one high-grade Mullerian, and one clear cell. No more than two operations were required to detect any ovarian cancer.
Seven cases of uterine endometrial cancer were detected by abnormal ROCA and follow-up TVS, six of which were stage I. Overall, 25 cancers and borderline tumors were detected by 34 operations, resulting in a PPV of 74% for any malignancy. All ROCA-associated operations were associated with pathological findings predicted by TVS, the authors noted.
"The magnitude of late-stage reduction ... is likely to be an important factor for reducing mortality," Bast and coauthors stated. "An association has been observed between stage shift and mortality in screening trials for non-small cell lung and breast cancers. A 20% or greater magnitude reduction in late stage has been associated with a significant decrease in mortality in randomized trials of mammography, and stage shift is an integral component of models that predict the mortality advantage of mammographic screening."
"While the UKCTOCS had a <20% stage shift, it remains to be determined whether a higher stage shift, as occurred in NROSS, would significantly reduce mortality," they wrote.
Disclosures
NROSS was sponsored by the University of Texas MD Anderson Cancer Center. The study was supported by the National Cancer Institute and multiple foundations and philanthropic organizations.
Bast disclosed relationships with InterVenn biosciences, Greenfire Bio, and Fujirebio Diagnostics (related to discovery of a CA125 assay).
Primary Source
Journal of Clinical Oncology
Han CY, et al "Normal-risk ovarian cancer screening study: 21-year update" J Clin Oncol 2024; DOI: 10.1200/JCO.00141.