Machine learning and deep learning for diagnosis of endometriosis: systematic review and meta-analysis.

What was studied

This review asked how well ML and DL models diagnose endometriosis. ML means machine learning. DL means deep learning. It followed PRISMA guidelines and a PROSPERO plan. The search ran through November 24, 2025. Risk of bias was rated with QUADAS-2. The pooled analysis used Stata 18.0. The review included studies that met its criteria.

What they found

Pooled sensitivity was 0.90. The range was 0.86 to 0.93. Specificity was 0.75. The range was 0.64 to 0.84. The SROC-AUC was 0.92. The range was 0.89 to 0.94. Disease probability went from 10 to 29 after a positive prediction. It went to 1 after a negative one. These are the abstract's own figures. Multicenter design and DL went with lower sensitivity. Results were robust.

What it means for patients

The authors say ML and DL models show high diagnostic performance. They see a use for non-invasive auxiliary diagnosis. They say more large, prospective, multicenter studies are needed. They also say multimodal models are needed. This is to reach clinical use.

Why it is not certain yet

Multicenter design and DL went with lower sensitivity. The authors say more prospective studies are needed before clinical use. The abstract gives no other limits. A review can only be as good as the studies in it.

Source

Machine learning and deep learning for diagnosis of endometriosis: systematic review and meta-analysis.

We restate facts from the source record in our own words. The original wording belongs to its publisher.