Multimodal AI integrating mpMRI with clinical and molecular data in prostate cancer

Combining multiparametric MRI with clinical variables incrementally improves prostate cancer AI model discrimination over MRI alone. Integration with PSMA PET and pathology further aids staging and prognosis. Domain shift and missing-modality robustness remain key barriers to cl…

• Combining mpMRI with clinical variables yields incremental improvement in model discrimination, calibration, or net benefit over single-modality models.
• Fusion with ultrasound and PSMA PET may aid biopsy targeting and staging; pathology-clinical fusion may support prognosis.
• Successful translation requires modality harmonization, missing-modality robustness, uncertainty estimation, and prospective impact studies.

Automated summary

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