Secondary Logo

Journal Logo

Prostate Cancer: Localized: Surgical Therapy VI (MP64): Moderated Poster 64: Monday, September 13, 2021

MP64-17 PREDICTING LONG-TERM SEXUAL FUNCTION IN POST-RADICAL PROSTATECTOMY PATIENTS BASED ON MRI AND CLINICAL FACTORS: A MULTI-CENTER STUDY

Dupati, Ajith; Johnson, Anna; Mirza, Mahin; Mmonu, Nnenaya; Lane, Brian R.; Shankar, Prasad R.; Davenport, Matthew S.; Moriarity, Andrew; Ryan, Michael; Dabaja, Ali A.; Peabody, James O.; George, Arvin K.; Singh, Karandeep

doi: 10.1097/JU.0000000000002104.17
  • Free

INTRODUCTION AND OBJECTIVE:

Long-term sexual dysfunction is common following radical prostatectomy (RP) for localized prostate cancer (CaP). We sought to determine whether inclusion of data derived from preoperative multiparametric prostate MRI (mpMRI) improves the prediction of sexual function at 12 months and 24 months following RP.

METHODS:

The Michigan Urological Surgery Improvement Collaborative (MUSIC) is a consortium of 45 diverse urology practices that maintains a prospective registry of men with CaP. Using mpMRI data from 3 large hospitals linked to MUSIC registry data, we fit random forest models to predict patient reported sexual function at 12 and 24 months following RP; poor function was defined as EPIC-26 scores <73 and good function was defined as scores ≥73. We compared models using baseline clinical predictors (age, body mass index, grade group, PSA, EPIC-26 sexual domain score, urinary incontinence score), MRI predictors (membranous urethral length, maximum PI-RADS v2 score, median lobe size, prostate volume), and both clinical and MRI predictors. We assessed model discrimination using a 10-fold cross-validated C-statistic.

RESULTS:

We identified 854 patients at 12 months and 477 patients at 24 months, of whom 671 (79%) had poor sexual function at 12 months and 350 (73%) at 24 months. The combined models had strong discriminative performance with C-statistics 0.94 for 12 months and 0.96 for 24 months (Table). Adding MRI variables did not improve the model over clinical factors alone. The 5 most important variables for both 12 and 24 month combined models were baseline sexual domain score, age, body mass index, initial PSA, and prostate volume.

CONCLUSIONS:

Models using clinical factors alone achieved very high discriminative performance in the prediction of long-term sexual function following RP. Compared to clinical variables alone, adding preoperative MRI data does not significantly improve prediction of sexual function in patients at 12 or 24 months after RP.

Source of Funding:

Clinical data were collected through MUSIC, which is funded by Blue Cross Blue Shield of Michigan

© 2021 by American Urological Association Education and Research, Inc.