Screening for ovarian cancer

We now need a definitive randomised trial

Computer simulation models predict that ovarian cancer screening could save at least three years of life per case1 and that screening is potentially cost effective.2 Such models have been developed and validated on data from published reports, but until now results from a randomised controlled trial have not been available. Jacobs et al recently described a randomised controlled trial of ovarian cancer screening in which 21 935 average risk postmenopausal women were randomised to three annual screens or to a control group.3 For their screening protocol Jacobs et al used the CA125 tumour marker as a first line screen and referred women to receive transvaginal sonography if CA125 concentrations were above 30 U/ml. If transvaginal sonography revealed an ovarian mass, women were referred for surgery. Designed as a pilot study, the trial was too small to show efficacy in terms of mortality reduction. However, its results are important for two reasons.

The Jacobs trial confirmed previous evidence4 that an acceptable positive predictive value is achieved by selecting women for evaluation by imaging on the basis of CA125 serum measurement. Neither transvaginal sonography nor CA125 is specific enough when used alone, but when the decision rule for surgical referral requires that both tests are positive the positive predictive value is 20%. Five operations …