Development and Validation of a Survival Nomogram and Calculator for Male Patients With Metastatic Castration-Resistant Prostate Cancer Treated With Abiraterone Acetate and/or Enzalutamide - Beyond the Abstract

Past clinical trials revealed the efficacy of docetaxel, abiraterone acetate (ABI), enzalutamide (ENZ), radium-223, cabazitaxel, prostate-specific membrane antigen-directed radioligand therapy with Lutetium-177, and Poly (ADP-ribose) polymerase inhibitors for metastatic castration-resistant prostate cancer (mCRPC) patients.


With several medication choices now available, clinicians should take care to select the most appropriate medicine to provide the optimal treatment. Prognostic prediction is important because of the lack of biomarkers for predicting the efficacy of individual mCRPC treatment. Using clinical information, the present study developed and validated a nomogram for predicting the prognosis of male patients with mCRPC who received ABI and/or ENZ treatment and developed an Excel-based prognostic calculator for clinical use.

A total of 568 mCRPC patients who received ABI and/or ENZ at Yokohama City University, Nagoya University, Kitasato University, and affiliated hospitals between 2012 and 2017 were evaluated. A nomogram for predicting overall survival (OS) was developed using a Cox proportional hazards regression model with age, initial PSA level, initial stage (M0/M1), Gleason score, time to CRPC, chemotherapy, PSA level at baseline, ALP level at baseline, and LDH level at baseline as predictors. Some of these factors were not independent prognostic variables in the multivariate analysis. However, they are established to be clinically important factors and were thus incorporated into the nomogram owing to their significant role in treatment decision making for CRPC.

Calibration was performed using the method described by Iasonos et al. The data were randomly separated into training and validation datasets to calibrate the nomogram prediction. The predictive ability was evaluated by comparing the predicted survival probability at 2 years with the observed survival probability, using the training and validation datasets. The discriminatory ability of the nomogram was assessed according to the concordance index (C-index). A 5-fold cross-validation was repeated 2000 times to estimate the C-index, and the mean values of the estimated C-index were determined using the training and validation sets.

The median (range) observational period was 13.3 (0.2-52.2) months. A total of 189 of the 568 patients (33.3%) died, and the median OS was 24.7 months. There was no difference in OS between patients treated first with ABI and those treated first with ENZ. A total of 202 patients (35.6%) were administered docetaxel systemic chemotherapy for CRPC and developed treatment resistance. None of the patients received either upfront ABI or ENZ during the initial treatment of metastatic hormone-naïve prostate cancer. The multivariate analysis showed that the time to CRPC (hazard ratio [HR] = 0.521, 95% confidence interval [CI] = 0.349-0.776, p = 0.001), chemotherapy naïve or not (HR = 1.681, 95% CI = 1.232-2.300, p = 0.001), baseline PSA level (HR = 1.439, 95% CI = 1.179-1.755, p < 0.001), baseline ALP level (HR = 1.827, 95% CI = 1.102-3.028, p = 0.019), and baseline LDH level (HR = 12.123, 95% CI = 5.343-27.51, p < 0.001) were independent risk factors for OS. Figure 1 shows the nomogram for predicting 1- and 2-year survival using five statically significant risk factors (i.e., time to CRPC, chemotherapy-naïve or not, baseline PSA, baseline ALP, and baseline LDH) and four clinically important risk factors (i.e., age, initial PSA, initial metastatic status, and Gleason score). The 2-year survival rate predicted by the nomogram correlated well with the actual 2-year survival rate in the validation set. The mean C-index was 0.72 for the training cohort and 0.71 for the validation cohort in the 5-fold cross-validation. We also developed an Excel-based calculator to conveniently apply these findings in daily clinical practice.

We developed and validated a nomogram for predicting the prognosis of patients with mCRPC who received ABI/ENZ treatment and an Excel-based prognostic calculator for use in clinical practice. The nomogram showed excellent predictive ability and may thus help clinicians make treatment decisions for patients with mCRPC.

figure 1.jpg 
Written by: Yasuhide Miyoshi, Departments of Urology and Renal Transplantation, Yokohama City University Medical Center, Yokohama, Kanagawa, Japan

Read the Abstract