CRCRpred

CRCRpred predicts individual colorectal cancer (CRC) risk scores from measured concentrations of mitochondrial apoptotic pathway proteins to stratify CRC patients by risk and survival outcome.

Key Features:

• Protein Concentration Analysis: Uses concentrations of Mitochondrial Outer Membrane Permeabilization (MOMP) proteins Bak, Bax, Bcl2, BclXL, and Mcl1 as input features.
• Risk Prediction Models: Implements multiple linear regression techniques including LassoLars and parameter optimization for model selection.
• Predictive Performance on Stage III Cohort: Trained and evaluated on a dataset of 134 stage III CRC patients, producing a Hazard Ratio (HR) up to 38.13 with p-value 0.0006 for risk differentiation.
• Survival Outcome Prediction: Predicts favorable versus unfavorable survival outcomes with reported accuracy up to 71.64%.

Scientific Applications:

• Prognostic Risk Stratification: Stratifies CRC patients by molecularly informed risk scores to support prognosis assessment.
• Biomarker Evaluation: Assesses the prognostic value of intrinsic apoptotic pathway proteins (Bak, Bax, Bcl2, BclXL, Mcl1) as biomarkers in CRC.
• Treatment Stratification Research: Supports research into personalized treatment planning by linking apoptotic protein profiles to survival outcomes.

Methodology:

Models were developed using multiple linear regression techniques including LassoLars with parameter optimization, trained and evaluated on a cohort of 134 stage III CRC patients, and performance was assessed using five-fold cross-validation.