Reproducibility in the assessment of postmenopausal ovaries with transvaginal ultrasound

Objective. The goal of this study was to evaluate the exterexaminer variation in the assessment of postmenopausal ovaries using transvaginal ultrasound (TW). Methods. One hundred eighty-eight cancer screening trial participants undergoing TW were reassessed by a second TVU examination. Results. Although first examiners tended to describe significantly larger left (P < 0.001) but smaller right (P = 0.036) ovaries, as well as fewer ovarian abnormalities, examiners agreed on the test interpretation 93% of the time (kappa = 0.846). In only two cases (1%) were the differences in interpretation such that the two examiners recommended different follow-up procedures. Conclusions. Because of the high fatality rate of ovarian cancer, early detection remains the best way to combat this devastating disease. TVU is one screening technique we are currently evaluating in a cancer screening trial. To help ensure screening test reproducibility, we have followed explicit protocols for training and certifying all TW examiners, as well as for conducting TVU examinations. This study demonstrates that by adhering to specific training, certification, and examination protocols, TVU reproducibility is excellent. Such protocols may well serve as a standard for TVU training and examination. (C) 2000 Academic Press.