A study published in July 2026 explored radiomics applied to pelvic MRI to identify patterns associated with different presentations of endometriosis.

The researchers extracted quantitative imaging features and used machine-learning models. Some classifiers showed promising signals, but false positives and limited specificity remained important issues.

Potential, but still experimental

The study highlights how differences in scanners, acquisition protocols and sample size can affect reproducibility. External validation is therefore needed before radiomics can be used broadly in routine clinical practice.

Source: arXiv, study on multiscale radiomics in pelvic MRI, July 2026.