Introduction

Radical prostatectomy (RP), external-beam radiotherapy (EBRT), and brachytherapy are standard treatments for localized prostate cancer but each may negatively impact bowel function. We sought to develop a nomogram predicting the probability of treatment-related bowel problems using prospective, patient-reported data.

Methods

Patient-reported data on treatment-related bowel problems was obtained from four prospective, longitudinal, health-related quality-of-life (HRQOL) protocols comprising 2,668 patients treated between 1999 and 2011 in the United States and Spain. A single HRQOL instrument was not uniformly used for each study, though a similar 5-point scale was used in each protocol to assess bother related to bowel problems. Bowel dysfunction was defined as bowel symptoms identified as a moderate-to-big problem by patients on survey responses before treatment and 2 years post-treatment. Multivariable logistic regression analysis was used to model the clinical information and follow-up data. Internal validation was performed using bootstrapping.

Results

Overall, 43 patients (2%) with complete data had bowel dysfunction prior to treatment. The rate of bowel dysfunction at 2 years in patients with no bowel dysfunction pretreatment was 2%, 10%, and 4% for patients treated by RP, EBRT, and brachytherapy, respectively (p<0.001). A nomogram based on pretreatment bowel bother, treatment modality, and race had a concordance index of 0.725 and predictions were well-calibrated with observed outcome on cross-validation. The predictive accuracy was not increased by the inclusion of additional demographic, tumor-related, or treatment-related variables, or by the creation of separate treatment-specific models.

Conclusions

A validated nomogram that predicts 2-year probability of bowel dysfunction after treatment for localized prostate cancer has been developed. The nomogram is anticipated to be useful for patient counseling regarding treatment options for localized prostate cancer.

Funding

NIH R01 CA 95662; NIH 1RC1CA146596