5 Key Takeaways

• 1

  A machine learning radiomics model integrating PET/CT and MRI shows promise for prostate cancer risk stratification.

• 2

  The study evaluated 488 patients, with significant findings from both internal and external validation cohorts.

• 3

  Multimodal models outperformed single-modality approaches, achieving high AUC scores for clinically significant prostate cancer.

• 4

  Limitations include retrospective design, small validation cohorts, and lack of prospective validation affecting generalizability.

• 5

  The findings suggest further study of multimodal radiomics as a decision-support tool, not a replacement for biopsy.