Combining 3D Segmentation and LLM on mpMRI Boosts Clinically Significant Prostate Cancer Detection

A combined model integrating 3D volumetric segmentation and an LLM-based classifier achieved AUROC 0.900 (95% CI 0.882-0.918) for csPCa on mpMRI, significantly outperforming single-modality metrics in multi-institutional external validation.

• Retrospective development on 3,896 patients; external validation on 1,154 patients with scanner/protocol heterogeneity.
• Both combined models showed significant improvements in net reclassification and integrated discrimination (P<0.05 for most).
• Limitations: retrospective design, though externally validated across multiple institutions.

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