Nonvisualization of the ovaries on pelvic ultrasound: does MRI add anything?

The purpose of our study is to assess the utility of pelvic magnetic resonance imaging (MRI) in the event that either one or both ovaries are not visualized by pelvic ultrasound.

This HIPAA-compliant retrospective study was approved by our local institutional review board and informed consent waived. 1926 pelvic MRI examinations between March 2007 and December 2011 were reviewed and included if a combined transabdominal and endovaginal pelvic ultrasound had been performed in the preceding 6 months with at least one ovary nonvisualized. Ovaries not visualized on pelvic ultrasound were assumed to be normal and compared with the pelvic MRI findings. MRI findings were categorized as concordant or discordant. Discordant findings were divided into malignant, non-malignant physiologic or non-malignant non-physiologic. The modified Wald, the "rule of thirds", and the binomial distribution probability tests were performed.

255 pelvic ultrasounds met inclusion criteria with 364 ovaries not visualized. 0 malignancies were detected on MRI. 6.9% (25/364) of nonvisualized ovaries had non-malignant discordant findings on MRI: 5.2% (19/364) physiologic, 1.6% (6/364) non-physiologic. Physiologic findings included: 16 functional cysts and 3 hemorrhagic cysts. Non-physiologic findings included: 3 cysts in post-menopausal women, 1 hydrosalpinx, and 2 broad ligament fibroids. The theoretical risk of detecting an ovarian carcinoma on pelvic MRI when an ovary is not visualized on ultrasound ranges from 0 to 1.3%.

If an ovary is not visualized on pelvic ultrasound, it can be assumed to be without carcinoma and MRI rarely adds additional information.