Medical-AI trial: AI ultrasound screening for ovarian cancer in postmenopausal women — a field with no effective screening (ASTOM) (NCT07660718)

Ovarian cancer rarely causes early symptoms and is often advanced by the time it is found. Unlike breast or cervical cancer, no widely available, effective screening method is established, so early detection is hard, and it is the leading cause of death among gynecologic cancers in developed countries. Postmenopausal women in particular are at higher risk.

Gynecologic ultrasound is the first-line tool for evaluating the ovaries, but telling benign from malignant requires expert reading.

This study (ASTOM) examines whether AI can be brought into that ultrasound evaluation, as a single-center prospective observational pilot. Per the registry summary, at a gynecologic-oncology unit of an academic hospital (IRCCS) in Bologna, Italy, it evaluates in postmenopausal women whether applying AI to ultrasound can discriminate ovarian-lesion risk categories.

The primary outcome is the predictive performance of the integrated model in accurately telling apart the different risk categories. The size is 100 patients, with enrollment closed and analysis underway.

Against the deep-seated problem that ovarian cancer has no effective screening, the idea of boosting an already widely used tool, ultrasound, with AI is pragmatic. Raising the discriminative power of an existing test without adding new equipment could help pick up disease earlier in a higher-risk group.

That said, this is a single-center, small pilot at the stage of evaluating predictive performance; it does not establish the effectiveness of ovarian-cancer screening or clinical adoption.