Statistics from Altmetric.com
Ovarian cancer is the most fatal gynaecological malignancy. Most cancers are detected at advanced stages when therapeutic options are limited. Epithelial ovarian cancers are heterogeneous with respect to cell of origin, molecular pathways, morphological appearance, risk factor associations and survival.1 ,2 High-grade serous carcinomas are most common and have the worst prognosis. Currently, proposed ovarian cancer screening strategies include transvaginal ultrasound and CA-125 blood testing. In a previous randomised controlled trial evaluating these strategies, no mortality benefit was observed.3
In the UK Collaborative Trial of Ovarian Cancer Screening (UKCTOCS), around 200 000 women aged 50–74 were enrolled and randomly assigned to three groups: 50% to receive no screening, 25% to annual transvaginal ultrasound screening (USS) and 25% to annual multimodal screening (MMS) with CA-125 testing using the ‘risk of ovarian cancer algorithm’ followed by transvaginal ultrasound. Women with persistent abnormalities were evaluated by a trial clinician who made management and treatment decisions. Only women who underwent a biopsy or surgery triggered by screening were considered screen-positive. The primary outcome was death due to ovarian cancer during follow-up. End points included primary non-epithelial ovarian cancer, borderline and invasive epithelial ovarian cancer. Six annual screens were initially planned, but due to lower than expected mortality in the control arm, additional screens were offered to women and follow-up was extended.
The sensitivity for detection of ovarian cancers diagnosed within a year of screening was 84% for MMS and 73% for USS. A significant shift to lower stages compared with no screening was observed for MMS. In the prespecified analysis, a non-significant reduction in ovarian cancer mortality was observed (15% for MMS and 11% for USS). Excluding prevalent cases, a significant mortality reduction of 20% was observed for MMS. A post-hoc analysis using a weighted log-rank test showed significant mortality reductions of 22% and 20% for MMS and USS, respectively, with larger reductions in later study years.
Implementation of population-based cancer screening programmes is complex and requires evaluation of benefits, harms and cost. Currently, the evidence about the efficacy of ovarian cancer screening remains inconclusive. In the Prostate, Lung, Colorectal and Ovarian Cancer Screening Trial (PLCO), a non-significant increase in ovarian cancer mortality was observed among screened women.3 While the stage shift and the mortality reductions observed in UKCTOCS are promising, the mortality benefit was not significant in the primary analysis. PLCO and UKCTOCS were powered to detect mortality differences of 30–35%, but the UKCTOCS data demonstrate that a 15–20% reduction may be more realistic. Thus, confirming the UKCTOCS results would require even larger trials.
The focus of most screening efficacy trials is on cancer-specific mortality reduction. However, for a complete evaluation of benefits and harms of screening, assessment of short-term and long-term adverse events related to the screening intervention is crucial. In PLCO, where management of screen-positive women was not standardised and reflected real-life practice, 15% of the screen-positive women had serious complications.3 In UKCTOCS, which had a more organised management, serious complications were rare. Implementation studies are important to evaluate the population effectiveness of the screening programmes outside of the highly controlled and standardised clinical trials.
Current data do not support moving to implementation and new, larger efficacy trials are unlikely to happen. However, cost-effectiveness analyses using a range of estimates on benefits, harms and cost of ovarian cancer screening from PLCO and UKCTOCS can predict scenarios in which ovarian cancer screening either could, or would not be, cost-effective. Following these analyses, efforts can be targeted to improve screening tests, to identify high-risk populations, or to minimise harm from false positive results.4
Implications for practice
There is insufficient evidence to justify population-based ovarian cancer screening. The association of mortality reduction with a stage shift suggests that it can serve as a surrogate end point in biomarker studies. Targeting screening to high risk populations is promising, but apart from inherited BRCA mutations, there are currently no markers that can identify a disease-enriched population that could benefit from ovarian cancer screening.
Contributors NW developed the concept and wrote the commentary.
Funding Intramural Research Program of the National Cancer Institute.
Competing interests None declared.
Provenance and peer review Commissioned; internally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.