Ultrasonographic Screening for Ovarian Abnormalities Does Not Decrease Mortality From Ovarian Cancer


In 1981, a study of 5479 women was performed in the United Kingdom with the purpose of determining whether ultrasonography of the ovaries might be useful in the early detection of ovarian cancer. At that time, the scans were performed with the transabdominal approach. The study participants were self-referred and were required to be in good health. Each woman had three screens, at one-year intervals. A detailed history including family history was obtained from all participants. Women who had "persistent" abnormal findings were recommended to have surgical intervention. The mean age of the cohort was 52.

Approximately one in every 17 women had at least one "persistent" abnormality. Of the 326 women who had such abnormalities, only 10% declined further investigation. Of those who agreed with intervention, approximately one-third had laparoscopy and two- thirds had laparotomy. Ninety percent of the women who had laparotomy had a bilateral salpingo-oophorectomy.

Nine cancers were identified, including four borderline and four metastatic ovarian cancers. Of the remaining 284 women, 88 had a benign epithelial tumor and 16 had a benign germ-cell tumor. The large majority of women had simple cysts, follicular cysts, or corpus lutea.

Years later (the article does not state when follow-up began), this cohort was traced. Ninety-five percent of the original participants were identified. The number of expected and observed deaths from all causes, and from cancer causes was calculated. Based on the expected number of deaths from ovarian cancer, the study had an 80% power to detect a decreased risk of ovarian cancer of 0.6 or an increase of 1.67.

Overall, the death rate of the women in this study was reduced from expected population rates. Crayford and colleagues point out that this is not a surprising finding since the participants in the project were all healthy volunteers.’ Overall, there was no decrease in deaths from ovarian cancer (Crayford JB, et al. Lancet 2000;355:1060-1063).

Comment by Kenneth L. Noller, MD

The early detection of ovarian cancer remains one of the major goals in the field of gynecology. Although only a few women in every 100,000 will develop the disease in any given year, the lifetime risk of developing the disease is approximately one in 70. In addition, because the disease is almost always detected at a late, nonsalvageable stage, most women with this diagnosis die from the disease. During the past two decades, there have been numerous attempts to develop a method that is both sensitive and specific that can be used for the early detection of ovarian cancer. Unfortunately, to date, none of the methods fulfill the criteria required for routine use as a screening test.

I particularly like this article because it involves the prospective follow-up of a screening trial for ovarian cancer. Although the 1981 screening was rather crude by modern ultrasonographic standards, enough information was collected on the cohort to make follow-up interesting. The results were not surprising. Despite performing some 300 surgical procedures for "persistent abnormalities" of the ovaries, the authors failed to reduce the risk of death from ovarian cancer.

Based on the article by Crayford et al, which of the following statements concerning screening for ovarian cancer is true?

a. Transabdominal ultrasonography can detect ovarian cancer but does not reduce mortality from the disease.

b. Removal of ovaries that show persistent abnormalities reduces the risk of ovarian cancer.

c. Surgical intervention is indicated in women with persistently abnormal ultrasonographic findings of the ovaries.

d. Transabdominal ultrasonography cannot reduce ovarian cancer deaths but does reduce overall cancer deaths.