Prediction of the Risk of Biochemical Relapse After Radical Prostatectomy for Prostate Cancer Using Radomics on Pre-therapeutic MRI

Prediction of the Risk of Biochemical Relapse After Radical Prostatectomy for Prostate Cancer Using Radomics on Pre-therapeutic MRI

With 50% of post-operative biochemical failure, efficient predictive models are needed to guide post-operative management.

Radiomic features are quantitative features extracted from medical imaging, supposed to be correlated with tumor heterogeneity.

We aim to build and test three predictive models (clinical, radiomic and combined models).

Inclusion Criteria:

• Age > 18yo
• Patients underwent radical prostatectomy
• High-risk prostate cancers: at least 1 criteria (pt3a/pT3b/pT4, R1, Gleason score > 7)

Exclusion Criteria:

• No available pre-operative MRI
• No analyzable pre-operative MRI
• proof of lymph-node involvement (cN1/2 or pN1/2)
• post-operative PSA > 0.04ng/mL