Objective:

To review and analyze the role of artificial intelligence in predicting biochemical recurrence (BCR) of prostate cancer using MRI.

Approach:

• Introduction to Prostate Cancer: Prostate cancer (PCa) is a prevalent malignancy among men worldwide, with a rising incidence. In 2025, approximately 313,780 new cases of prostate cancer and 35,770 deaths are expected in the United States alone.
• Current Detection Methods: Current clinical detection of PCa relies on PSA testing, imaging techniques, and biopsies, which have limitations in specificity and reliability.
• Role of AI in Diagnosis: AI has emerged as a promising tool in PCa diagnosis, outperforming traditional methodologies in processing complex datasets and improving predictive accuracy.
• MRI-based AI Models: MRI-based AI models integrating radiomics and deep learning have shown potential in predicting BCR after prostatectomy or other treatments.

Key Findings:

• AI models demonstrate improved predictive accuracy for BCR compared to traditional methods, as shown in various studies.
• MRI-based AI models have shown robust AUC values in validation cohorts.
• Current studies face limitations including small sample sizes and lack of multicenter validation.

Interpretation:

While AI shows promise in enhancing the prediction of BCR in prostate cancer, further validation and exploration of its clinical applicability are necessary.

Limitations:

• Most studies are retrospective and exploratory.
• Limited sample sizes and multicenter external validation.
• Concerns regarding intermodel variability and overfitting.

Conclusion:

AI-based MRI models could significantly aid in BCR risk stratification and decision-making in prostate cancer management, pending further validation.

Sources:

• Cancers (Basel)
• Urology Times Journal
• European Urology
• Scientific Reports
• Abdominal Radiology
• Frontiers in Oncology
• Clinical Cancer Research
• Diagnostics (Basel)
• EngMedicine