Postprostatectomy Urinary Continence Prediction: External Validation of a Model Incorporating MRI Derived Membranous...

Area of Science:

• Urology
• Radiology
• Medical Informatics

Background:

• Membranous urethral length (MUL) measured by MRI is linked to post-prostatectomy continence.
• Current predictive models for continence lack external validation, limiting their use in patient counseling.
• Objective: To evaluate a MUL-based continence prediction model in an external patient cohort.

Purpose of the Study:

• Assess the performance of a predictive model for post-prostatectomy continence.
• Validate the model using membranous urethral length (MUL) derived from pre-operative MRI.
• Evaluate the model's utility in an independent cohort of prostatectomy patients.

Main Methods:

• A prospective cohort of 500 patients undergoing prostatectomy (2015-2023) with pre-operative MRI was studied.
• Membranous urethral length (MUL) was measured retrospectively; continence was assessed post-operatively.
• The prediction model incorporated age, nerve sparing, prostate volume, surgical approach, and MUL. Performance was evaluated using c-statistics, calibration, and decision curve analysis.

Main Results:

• Median MUL was 11 mm. One-year continence (0 pads/24h) was achieved by 62% of patients (c-statistic: 0.64).
• The model showed poor calibration for predicting 0 pads/24h, with no net benefit.
• Continence (0-1 pad/24h) was achieved by 90% of patients (c-statistic: 0.69), with moderate calibration and net benefit shown by decision curve analysis.

Conclusions:

• A recalibrated prediction model using MRI-derived membranous urethral length (MUL) offers net benefit for predicting 0-1 pad/24h continence.
• The model's utility is limited for predicting complete dryness (0 pads/24h).
• Further refinement may be needed for precise prediction of complete urinary continence post-prostatectomy.