Development and External Validation of a Machine Learning-Derived Risk Stratification Score for Metastatic Castration-Resistant Prostate Cancer
- 1Topazium Artificial Intelligence, Madrid, Spain
- 2Medicus, Buenos Aires, Argentina
- 3DLPHI Oncology Global Consultants, Vaud, Switzerland
Abstract
Purpose: Metastatic castration-resistant prostate cancer (mCRPC) shows marked clinical heterogeneity and poor long-term survival, highlighting the need for simple and reliable risk stratification tools beyond the limited usability of existing complex prognostic models. In this study, we developed and validated PROGRESS (Prostate Cancer Global Risk Evaluation and Stratification Score), a simplified prognostic score derived through machine learning‑guided feature selection distilled into a simple laboratory-based tool.
Methods: Baseline data from 2,035 patients with mCRPC (training set) were embedded into synthetic representations, reduced with UMAP, and clustered with DBSCAN to discover outcome-related subpopulations. Feature contribution analysis informed a multivariable model that yielded three independent predictors (prostate specific antigen [PSA], alkaline phosphatase [ALP], and AST) composing PROGRESS. External validation used two independent data sets (n = 1,239) across different therapeutic settings. Overall survival (OS) was analyzed with Kaplan-Meier and Cox regression; discrimination was assessed by ROC-AUC, AUC(t), and Harrel C-index.
Results: Unsupervised modeling identified two patient risk subpopulations with significantly different OS rates (median 27.4 v 17.7 months; hazard ratio [HR], 2.20 [95% CI, 1.91 to 2.54]; P < .00001). The three retained predictors (PSA, ALP, and AST) were combined into a composite score (PROGRESS) that showed strong discrimination in the training cohort (ROC-AUC 0.89; AUC(t) 0.60; Harrel C-index 0.69). Using a prespecified cutoff, patients classified by PROGRESS as increased risk had significantly shorter survival than low-risk patients (median 18.3 v 25.6 months; HR, 1.72 [95% CI, 1.50 to 1.97]; P < .0001). PROGRESS prognostic performance was consistent across all validation cohorts (all P < .0001).
Conclusion: By combining ML-based pattern discovery with classical modeling, PROGRESS provides a simple, objective, and readily implementable prognostic tool using three inexpensive routine laboratory measures. Its consistent performance across independent data sets supports clinical utility for individual risk stratification, trial enrichment, and early decision making. Prospective real-world validation is warranted.
Related
- Real-World PSA Response and Safety Outcomes of 177Lu-PSMA-617 in Heavily Pretreated Metastatic Castration-Resistant Prostate Cancer: A Single-Center Retrospective Validation Study
- Real‐World Overall Survival With Abiraterone Acetate and Enzalutamide in Patients With Chemotherapy‐Naïve Metastatic Castration‐Resistant Prostate Cancer From the Flatiron Electronic Health Records Database
- Immune biomarkers of response and resistance to ipilimumab and nivolumab in metastatic castration-resistant prostate cancer
- Niraparib in patients with metastatic castration-resistant prostate cancer and DNA repair gene defects (GALAHAD): a multicentre, open-label, phase 2 trial
- Global clinical trial activity in metastatic castration-resistant prostate cancer from January 2020 to March 2026: a landscape analysis
BibTeX
@article{castrolabrador2026development,
title = {{Development and External Validation of a Machine Learning-Derived Risk Stratification Score for Metastatic Castration-Resistant Prostate Cancer}},
author = {Laura Castro-Labrador and Rafael Zamora and Sergio Szyldergemajn and Pablo Gómez del Campo and José Del Castillo Izquierdo and José A. De All and Juan M. Domínguez and Carlos M. Galmarini},
journal = {JCO Clinical Cancer Informatics},
year = {2026},
month = oct,
doi = {10.1200/cci-26-00054},
url = {https://doi.org/10.1200/cci-26-00054},
note = {CC BY-NC-ND 4.0},
}
RIS
TY - JOUR
TI - Development and External Validation of a Machine Learning-Derived Risk Stratification Score for Metastatic Castration-Resistant Prostate Cancer
AU - Laura Castro-Labrador
AU - Rafael Zamora
AU - Sergio Szyldergemajn
AU - Pablo Gómez del Campo
AU - José Del Castillo Izquierdo
AU - José A. De All
AU - Juan M. Domínguez
AU - Carlos M. Galmarini
PY - 2026
DA - 2026/10/20
JO - JCO Clinical Cancer Informatics
DO - 10.1200/cci-26-00054
UR - https://doi.org/10.1200/cci-26-00054
AB - Purpose: Metastatic castration-resistant prostate cancer (mCRPC) shows marked clinical heterogeneity and poor long-term survival, highlighting the need for simple and reliable risk stratification tools beyond the limited usability of existing complex prognostic models. In this study, we developed and validated PROGRESS (Prostate Cancer Global Risk Evaluation and Stratification Score), a simplified prognostic score derived through machine learning‑guided feature selection distilled into a simple laboratory-based tool. Methods: Baseline data from 2,035 patients with mCRPC (training set) were embedded into synthetic representations, reduced with UMAP, and clustered with DBSCAN to discover outcome-related subpopulations. Feature contribution analysis informed a multivariable model that yielded three independent predictors (prostate specific antigen [PSA], alkaline phosphatase [ALP], and AST) composing PROGRESS. External validation used two independent data sets (n = 1,239) across different therapeutic settings. Overall survival (OS) was analyzed with Kaplan-Meier and Cox regression; discrimination was assessed by ROC-AUC, AUC(t), and Harrel C-index. Results: Unsupervised modeling identified two patient risk subpopulations with significantly different OS rates (median 27.4 v 17.7 months; hazard ratio [HR], 2.20 [95% CI, 1.91 to 2.54]; P < .00001). The three retained predictors (PSA, ALP, and AST) were combined into a composite score (PROGRESS) that showed strong discrimination in the training cohort (ROC-AUC 0.89; AUC(t) 0.60; Harrel C-index 0.69). Using a prespecified cutoff, patients classified by PROGRESS as increased risk had significantly shorter survival than low-risk patients (median 18.3 v 25.6 months; HR, 1.72 [95% CI, 1.50 to 1.97]; P < .0001). PROGRESS prognostic performance was consistent across all validation cohorts (all P < .0001). Conclusion: By combining ML-based pattern discovery with classical modeling, PROGRESS provides a simple, objective, and readily implementable prognostic tool using three inexpensive routine laboratory measures. Its consistent performance across independent data sets supports clinical utility for individual risk stratification, trial enrichment, and early decision making. Prospective real-world validation is warranted.
SN - 2473-4276
ER -