Task 1: MRI-Only Diagnostic Decision¶

Goal: Use multiparametric MRI information and clinical data (e.g. PSA, PSA density, and prostate volume) to determine whether prostate biopsy is recommended before histopathological confirmation is available.

Input data

• structured-prompt.json: structured clinical variables, such as age, PSA, previous PSA, PSA velocity, PSA density, prostate volume, PI-RADS category, DRE result, previous biopsy status, comorbidities, and csPCa likelihood score
• prostate-biopsy-decision-clinical-data.json: radiology report, prior clinical notes, PSA trend, laboratory results, and family history
• prostate-modality-level-neural-representations.json: MRI neural representation

See Data Sources and Imaging for the complete input structure and variable descriptions.

Output required

• Binary biopsy recommendation ("yes" / "no")
• Structured reasoning including confidence in the biopsy decision, the importance assigned to the relevant clinical variables, the clinical-data sections/tools accessed by the agent during the decision, and a free-text explanation of the main factors supporting the decision

Ground truth: Histopathology-confirmed csPCa (ISUP Grade Group >= 2); negative cases confirmed by longitudinal PSA follow-up.

Expected output per case: Two JSON files (~5 KB) containing the biopsy recommendation and the corresponding structured reasoning.

Provided neural representation: MRI features from a validated MRI prostate zone segmentation model.

Primary metric: Case score, please refer to this page and the repo.