SESAUA 2024: Addressing Racial Disparities in Prostate Cancer Prediction Models: External Validation and Comparison of Four Models of Pathological Outcome Prediction before Radical Prostatectomy

(UroToday.com) The 2024 Southeastern Section of the AUA (SESAUA) annual meeting featured a prostate cancer session and a presentation by Dr. Mahdi Mottaghi discussing racial disparities in prostate cancer prediction models, specifically the external validation and comparison of four models of pathological outcome prediction before radical prostatectomy. The development of widely used nomograms for predicting radical prostatectomy outcomes involved a relatively small proportion of Black men, potentially compromising the accuracy of predictions for this patient group. Additionally, most available models developed in tertiary care centers tend to overpredict outcomes in the general population. The Michigan Urological Surgery Improvement Collaborative (MUSIC) nomogram was developed based on the general urology patient population.1 This study by Dr. Mottaghi and colleagues presented at SESAUA 2024 aimed to implement the first external validation of the MUSIC model and compare the predictive performance of these models in Black men.

Dr. Mottaghi extracted data from 4,139 (n = 1,138 Black) men from the Shared Equal Access Regional Cancer Hospital database of the Veterans Affairs system. The objective was to externally validate the MUSIC model focusing on Black men and compare it to the Memorial Sloan Kettering, Briganti 2012, and 2017 Partin tables for the prediction of lymph node invasion, extra-prostatic extension, seminal vesicle invasion, and non-organ-confined disease.

The median PSA of Black men was 1 mg/dL higher, although these patients were younger by a median of three years and presented with a lower-stage disease. The mean number of examined lymph nodes was 9 (standard deviation 7) and the median was 7 (IQR 3-12). The preoperative grade group and the median number of positive (4, IQR 2-6) and negative cores (7, IQR 4-9) were similar between the two groups. All models showed 7-10% lower discriminatory accuracy for lymph node invasion in Black men, marking the most notable area of racial disparity:

image-0.jpg

Although the AUC for MUSIC and Memorial Sloan Kettering models were almost equal, Memorial Sloan Kettering overpredicted extra-prostatic extension and lymph node invasion among all participants, regardless of race. Calibration plots of seminal vesicle invasion showed better MUSIC performance for Black men, while Memorial Sloan Kettering was better for White men:

image-1.jpg

MUSIC showed the highest net benefit for Black men in predicting extra-prostatic extension, lymph node invasion, and seminal vesicle invasion (predictive probability: <25%). MUSIC also showed the most net benefit for extra-prostatic extension and lymph node invasion for White men:

image-2.jpg

Dr. Mottaghi concluded his presentation discussing racial disparities in prostate cancer prediction models with the following conclusions:

  • The MUSIC model exhibited higher net benefit over the Memorial Sloan Kettering model in predicting lymph node invasion outcomes among men in the SEARCH database, especially for Black men in this cohort

Presented by: Mahdi Mottaghi, Durham Veterans Affairs Medical Center, Durham, NC

Written by: Zachary Klaassen, MD, MSc – Urologic Oncologist, Associate Professor of Urology, Georgia Cancer Center, Wellstar MCG Health, @zklaassen_md on Twitter during the 2024 Southeastern Section of the American Urological Association (SESAUA) Annual Meeting, Austin, TX, Wed, Mar 20 – Sat, Mar 23, 2024. 

References:

  1. Otles E, Denton BT, Qu B, et al. Development and validation of models to predict pathological outcomes of radical prostatectomy in regional and national cohorts. J Urol. 2022;207(2):358-366.