Evaluation of an artificial intelligence model based on multiparametric transrectal ultrasound for localizing clinically significant prostate cancer by simulation of targeted biopsies - Summary - MDSpire

Evaluation of an artificial intelligence model based on multiparametric transrectal ultrasound for localizing clinically significant prostate cancer by simulation of targeted biopsies

  • By

  • Daniel L. van den Kroonenberg

  • Florian Delberghe

  • Auke Jager

  • Arnoud W. Postema

  • Katelijne C. C. de Bie

  • Johannes B. Reitsma

  • Marije Zwart

  • Hessel Wijkstra

  • Anna Garrido-Utrilla

  • Joost de Baaij

  • Jean-Paul A. van Basten

  • Henk G. van der Poel

  • Harrie P. Beerlage

  • Massimo Mischi

  • Jorg R. Oddens

  • November 6, 2025

  • 0 min

Share

Objective:

To validate the performance of an AI model based on multiparametric ultrasound (mpUS) for detecting clinically significant prostate cancer (csPCa) at the patient level through simulation of targeted biopsies, emphasizing the importance of this validation.

Key Findings:
  • Internal evaluation showed sensitivity of 0.82 and specificity of 0.43 for ISUP ≥ 2, indicating the model's performance in identifying significant cases.
  • External evaluation demonstrated sensitivity of 0.81 and specificity of 0.42 for ISUP ≥ 2, providing a comparative perspective.
  • For ISUP ≥ 3, sensitivity was 0.90 in internal and 0.96 in external evaluations, highlighting the model's effectiveness in more advanced cases.
Interpretation:

The AI model shows promising diagnostic performance for detecting csPCa, with high sensitivity but lower specificity, indicating potential for clinical application with further refinement and consideration of specificity implications.

Limitations:
  • The study is retrospective and may have biases related to patient selection, which could affect generalizability.
  • The simulation of biopsies may not fully replicate real-world clinical procedures, potentially impacting the applicability of results.
Conclusion:

The mpUS AI model demonstrates potential for improving the detection of clinically significant prostate cancer, warranting further prospective clinical evaluation to confirm these findings.

Original Source(s)

Evaluation of an artificial intelligence model based on multiparametric transrectal ultrasound for localizing clinically significant prostate cancer by simulation of targeted biopsies

  • by Daniel L. van den Kroonenberg, Florian Delberghe, Auke Jager, Arnoud W. Postema, Katelijne C. C. de Bie, Johannes B. Reitsma, Marije Zwart, Hessel Wijkstra, Anna Garrido-Utrilla, Joost de Baaij, Jean-Paul A. van Basten, Henk G. van der Poel, Harrie P. Beerlage, Massimo Mischi, Jorg R. Oddens

  • November 6, 2025

Related Content

CT and MRI radiomics in cardiovascular risk prediction: a systematic review and meta-analysis by the EuSoMII Radiomics Auditing Group

FDA Clears AI Guidance Tool for Mitral Repair

AI enhances real-time intraprocedural visualization and navigation.

The Loan Cap That Could Shrink the Doctor Pipeline

New federal limits on medical school borrowing may quietly reshape who becomes a physician—and where they practice.