Does Ovarian Cancer Screening Actually Improve Survival?

November 26, 2018
Dave Levitan
Dave Levitan

A screening study that began more than 30 years ago looked at ultrasound screening for at-risk women for ovarian cancer and its effects on disease-specific survival.

Results of a screening study that began more than 30 years ago found that ultrasound screening of at-risk women for ovarian cancer resulted in earlier stage at detection of the malignancy, and significant improvements in disease-specific survival out to 20 years. However, questions about the specific population in this study remain, and there is still uncertainty whether such a program would be beneficial.

“Although advances in surgical technique, postoperative care, and chemotherapy have resulted in a modest improvement in 5-year survival, most [ovarian cancer] patients continue to present with advanced-stage disease in which the cure rate is low and treatment cost is high,” wrote study authors led by John R. van Nagell Jr, MD, of the University of Kentucky-Markey Cancer Center in Lexington. “Because early-stage ovarian cancer is highly curable, screening at-risk women has been investigated by several different approaches as a method to increase earlier diagnosis and improve patient survival.” The effect of screening on survival, however, has remained unclear.

The Kentucky Ovarian Cancer Screening Trial began in January 1987; the current analysis includes a total of 46,101 women enrolled from then until 2017. It included asymptomatic women aged 50 years or older, or asymptomatic women aged 25 years or older with a documented family history of ovarian cancer in a primary or secondary relative. In total, 23.2% of the cohort had such a family history, and 43.8% had a family history of breast cancer.

The women underwent annual ultrasound screening, for a total of 298,418 scans (mean of 6.5 scans per participant). Results of the study were published in Obstetrics & Gynecology.

The screening program detected a total of 71 cases of invasive epithelial ovarian cancer, along with 17 epithelial ovarian tumors of low malignant potential; none of the women with the latter type of tumor experienced recurrent disease. As was found in earlier ovarian cancer screening trials, the program resulted in earlier detection of disease: 42% of cases were found at stage I, along with 21% at stage II, 37% at stage III, and none at stage IV. In a cohort of unscreened women with ovarian cancer, 30% had stage I or II disease when detected (P < .001), while 70% had stage III or IV disease.

All patients with stage I–IIIB disease achieved complete cytoreduction at surgery. Following surgery and systemic therapy, patients were followed for a mean of 7.8 years (range of 0.8 to 27.2 years); during that time, 17 patients died of other causes with no evidence of disease, and 41 patients remain alive. One patient diagnosed at stage I (3%) has died of ovarian cancer, along with 33% of those with stage II disease, and 38% of those with stage III disease.

The disease-specific survival (DSS) rate was 86% at 5 years, 68% at 10 years, and 65% at 20 years. In contrast, women with ovarian cancer who did not undergo screening had DSS rates at those time points of 45%, 31%, and 19%, respectively (P < .001). Those unscreened women were from the same geographic area, and were treated at the same institution with the same treatment protocols. Disease stage at detection was significantly associated with DSS.

“Without early detection, it is likely that many of these tumors would have advanced significantly before clinical diagnosis, making the chance of cure unlikely even with optimal treatment,” the authors wrote, adding that more research is still needed to better select which populations of women will benefit from this type of screening program.

In an accompanying editorial, Sharon E. Robertson, MD, MPH, and Jeffrey F. Peipert, MD, PhD, of the Indiana University School of Medicine in Indianapolis, pointed out that the incidence of epithelial ovarian cancer in this study was substantially higher than that of the general population (271 per 100,000 women vs 11.4 per 100,000 women), which improves the positive predictive value of screening. “When the test sensitivity and specificity as reported in this trial are applied to the general population, the positive predictive value falls to an unacceptable 0.7%,” they wrote, also pointing out that specific histologies are not reported and the tumors in the screened population may differ from those in the general population.

With this in mind, it is still difficult to say that ovarian cancer screening among a low-risk population of women would be beneficial. As Robertson and Peipert pointed out, the US Preventive Services Task Force this year recommended against screening for ovarian cancer. There is still a possibility that focusing on women with BRCA mutations or other risk alleles, especially as genetic testing improves and becomes more affordable, will signal a change in this realm.

“We agree that an ovarian cancer screening algorithm for women at high risk of developing ovarian cancer may detect this devastating disease at an earlier, potentially more treatable stage,” they concluded. “However, the evidence that screening for ovarian cancer improves survival remains elusive. More evidence is needed before we adopt a widespread screening program.”