Endovaginal Ultrasound to Evaluate Endometrial Abnormalities—Reply

In Reply: Dr Bree and colleagues raise the question of whether hysterosonography is more sensitive for detecting benign endometrial abnormalities than EVUS. Bree et al found that EVUS detected only 32 of 54 cases of benign endometrial disease (sensitivity, 59%; 95% confidence interval [CI], 45%-72%). A summary of all available studies provides a more stable estimate of the sensitivity of EVUS and suggests a considerably higher sensitivity for detecting benign abnormalities than Bree et al report. Using a cutoff of 5 mm for our systematic review, we found that EVUS detected 92% (95% CI, 90%-93%) of endometrial abnormalities (defined as cancer, atypical and complex hyperplasia, and polyps) among 1306 women and 90% (95% CI, 88%-92%) of benign endometrial abnormalities among 849 women. We did not include women with uterine leiomyomata among those with benign endometrial abnormalities. Leiomyomata are an uncommon cause of postmenopausal endometrial bleeding (<1%), and excluding these women is unlikely to have substantially affected our results. The prospective trial referenced by Bree et al also reports a poor sensitivity of EVUS for detecting benign endometrial abnormalities.¹ In this study, EVUS was performed following endometrial biopsy, and thus it is not surprising that abnormalities may have been missed by EVUS if the lining of the uterus had already been substantially removed by the biopsy procedure or was obscured by bleeding.¹ Perhaps the sensitivity of EVUS presented by Bree et al, in part, reflects that endometrial biopsy preceded EVUS.