Ultrasound tests of ovarian reserve; a systematic review of accuracy in predicting fertility outcomes.

We conducted a systematic review of studies evaluating the diagnostic accuracy of all the ultrasound based tests of ovarian reserve, including antral follicle count (AFC), ovarian volume and stromal blood flow in predicting fertility outcomes and, where appropriate, performed a meta-analysis to determine the predictive_value at each cut-off value described in the literature. Included in the analysis were 17 studies for AFC, six studies for ovarian volume and six studies for stromal blood flow. Meta-analyses showed that women with AFC less than four were 8.7 times more likely not to get pregnant after IVF (two studies; 95% CI, 2.4-31.7) than women with AFC four or more. The sensitivity and specificity of AFC to predict cycle cancellation was 66.7% and 94.7%, respectively. Women with an AFC of less than four were 37 times (two studies; 95% CI, 13.68-100.45) more likely to have their cycle cancelled than women with AFC of four o r more. Ovarian volume measurement, at a cut off value of 3 cm3, showed specificity for prediction of cycle cancellation and non-pregnancy of 92% (three studies, 95% CI, 89-94) and 93% (three studies, 95% CI, 87-97), respectively. The clinical value of Doppler studies for ovarian stromal blood flow was unclear.