Theoretical Assessment of a Dynamic Risk Acceleration Index (ICAI) for Oncology Toxicity Prediction

Background Severe chemotherapy-related toxicity remains a major challenge in oncology care, often leading to dose reductions, treatment delays, or hospitalization. Conventional risk assessment tools such as the ECOG Performance Status and Charlson Comorbidity Index provide static evaluations but lack temporal sensitivity. A novel dynamic measure, the Individual Clinical Acceleration Index (ICAI), has been proposed to quantify the acceleration of individual risk probability. Objective To assess the incremental predictive value of ICAI compared to conventional oncology risk tools in forecasting Grade ≥ 3 chemotherapy-related toxicity, using previously published datasets as reference models. Methods Published data from prior studies of cancer patients receiving chemotherapy were used to simulate the integration of ICAI into existing predictive frameworks. Conventional models demonstrated moderate discriminative ability (AUC ~ 0.72–0.75). ICAI was incorporated as a dynamic variable, and predictive performance was evaluated using AUC, Net Reclassification Improvement (NRI), and bootstrap resampling for internal validation. Results Simulation analyses indicated that adding ICAI improved predictive performance, with AUC increasing to ~ 0.80, sensitivity rising from 68% to 83%, and NRI showing a positive gain of approximately + 12%. These findings suggest that dynamic risk acceleration provides meaningful incremental value over static models Conclusion ICAI enhances predictive accuracy when combined with established oncology risk tools. While these results are based on re-analysis of published data rather than new experimental evidence, they highlight the potential of ICAI to refine risk stratification. Future prospective studies are needed to validate its clinical utility.