Streamlined MRI sequences plus clinical data best classify rotator cuff tears

A streamlined MRI set (T2, PD coronal, PD transverse) with clinical metadata achieved AUC 0.9961 (95% CI 0.9823-0.9987) for rotator cuff tear classification in internal validation, outperforming full protocol.

• In a single-center study of 199 patients, a hierarchical multimodal model using T2, PD coronal, and PD transverse MRI with clinical metadata achieved an AUC of 0.9961 (95% CI 0.9823-0.9987) for rotator cuff tear classification.
• This streamlined three-sequence model outperformed the full four-sequence protocol (AUC 0.9909) and a single-sequence PD coronal model (AUC 0.9834).
• External validation is needed given the single-center design.

Automated summary

RadPigeon summaries are original and for information only. They are not clinical advice.