Towards Precision Oncology: A New Predictive Machine Learning Model for Early Progression to Castration Resistant...

Area of Science:

• Oncology
• Bioinformatics
• Machine Learning

Background:

• Metastatic castration-resistant prostate cancer (mCRPC) is an aggressive malignancy with poor prognosis upon early progression.
• Predicting early progression is crucial for tailoring treatment strategies in mCRPC.
• Current prediction methods require enhancement for improved clinical utility.

Purpose of the Study:

• To develop and validate a novel machine learning (ML) model for predicting early progression (≤12 months) to mCRPC.
• To compare the predictive performance of the novel ML model against standard ML algorithms.
• To identify a robust tool for risk stratification in patients with mHSPC.

Main Methods:

• Retrospective analysis of 172 patients with metastatic hormone-sensitive prostate cancer (mHSPC) from the MSK-IMPACT cohort.
• Inclusion of 11 clinical, pathological, and genomic variables for model development.
• Evaluation using stratified fivefold cross-validation with Area Under the Curve (AUC) as the primary metric.

Main Results:

• The novel Rivality Index (RINH)-based model achieved a superior AUC of 0.86.
• Standard ML algorithms had AUCs not exceeding 0.67, indicating significantly lower predictive performance.
• The RINH model demonstrated 74% accuracy, 70% sensitivity, and 77% specificity.

Conclusions:

• The RINH model serves as a robust tool for risk stratification in mHSPC patients.
• This model has the potential to personalize therapeutic strategies for mCRPC.
• External validation in multi-center, prospective cohorts is necessary before clinical implementation.