A review published in Current Molecular Pharmacology (2026, Volume 19, Pages 85-96), led by Jia Wang, Hong-Rui Zhu, Zhi-Chun Gu, Hou-Wen Lin and colleagues at Shanghai Jiao Tong University School of Medicine, investigates computational approaches to predict tumor drug resistance, according to News-Medical. This article shows you how machine learning and deep learning The model integrates multi-omics data from repositories such as: T.C.G.A. and GDSC To study resistance across chemotherapy, targeted therapies, and immunotherapies. The authors report that while standardized databases and robust preprocessing pipelines are essential for model input, there are challenges such as data sparsity, batch effects, and the black-box nature of deep models. News-Medical quotes Dr. Zhi-Chun Gu. “The inherent trade-off between model accuracy and interpretability undermines clinician confidence and limits real-world adoption.” News-Medical reports that the review advocates for explainable AI, multimodal fusion, longitudinal fluid monitoring, and specialized tools for high-risk subgroups such as patients with cancer-related thrombosis, and calls for uniform data standards and prospective clinical validation. Professor Hou-Wen Lin said: “Our goal is to go beyond general predictions and provide tailored insights to patients who need it most.”
