DIGITAL TWINS OF ONCOLOGY PATIENTS: SIMULATION OF TOXICITY AND EFFECTIVENESS OF TARGETED THERAPY IN REAL TIME.

Abstract

 The traditional approach to developing cancer drugs is facing a crisis of effectiveness due to high biological heterogeneity. This requires a transition to quantitative, prognostic methods. An overview of the use of mathematical modeling to create digital twins of patients in oncology to predict the effectiveness and toxicity of targeted therapy in real time.The paper analyzes the principles of Model-Informed Drug Development (MIDD) and the triad of models: systems biology, quantitative systems pharmacology (QSP), and quantitative systems toxicology (QST). It describes methodologies for constructing virtual populations (using Latin Hypercube Sampling) and their individualization into digital twins. Organ-specific QST platforms (CiPA, DILIsym, Chaste, RENAsym) for predicting cardiotoxicity, hepatotoxicity, gastrointestinal toxicity, and nephrotoxicity are considered.It is shown that system modeling serves as a link between drug development stages (Transition Points), allowing data to be extrapolated, risks to be minimized, and clinical trial designs to be optimized. The mathematical superiority of simultaneous combination therapy over sequential therapy for overcoming resistance is justified.The integration of quantitative modeling into clinical practice opens a new era of personalized oncology. Digital twins, based on physiological principles and real-time data, are becoming a powerful tool for supporting medical decisions, improving patient safety, and increasing treatment effectiveness.