Recently published comprehensive reviews Current molecular pharmacology (2026, Volume 19, Pages 85-96) explore the rapidly evolving landscape of computational tools for predicting tumor drug resistance. The paper, led by Jia Wang, Hong-Rui Zhu, and corresponding authors Zhi-Chun Gu and Hou-Wen Lin of Shanghai Jiao Tong University School of Medicine, systematically maps how artificial intelligence, particularly machine learning and deep learning, is being leveraged to integrate multi-omics data from large-scale repositories such as TCGA and GDSC. These approaches can help decipher resistance mechanisms across chemotherapy, targeted therapy, and immunotherapy, while also suggesting new predictive aspects, such as cancer-associated thrombosis.
The authors highlight that standardized databases and advanced preprocessing pipelines are now essential to convert heterogeneous genomic, transcriptomic, and clinical data into reliable model inputs. However, they caution that data sparsity, batch effects, and the “black box” nature of many deep learning models remain major barriers to clinical adoption. ”The inherent trade-off between model accuracy and interpretability undermines clinician confidence and limits real-world adoption.To address this, this review advocates the integration of explainable AI frameworks, multimodal fusion strategies, and dynamic liquid biopsy monitoring to capture resistance evolution in real time.
Looking forward, the research team is calling for a paradigm shift toward specialized tools for high-risk subgroups, particularly patients with cancer-related thrombosis. By incorporating coagulation-related signatures and longitudinal thrombotic markers, these next-generation models have the potential to provide actionable predictions to guide anticancer and anticoagulant combination therapy. The authors also call for the establishment of uniform data standards, future clinical validation, and interdisciplinary collaboration to bridge the gap between computational innovation and bedside applications. ”Our goal is to go beyond general predictions and provide tailored insights for patients who need it most.This review concludes that with continued efforts in data integration, interpretability, and clinical translation, AI-driven resistance prediction has the potential to revolutionize precision oncology.
sauce:
Reference magazines:
Wang, J. Others. (2026). Oncology drug resistance prediction tools: database infrastructure, algorithmic innovations, and clinical translation. Current molecular pharmacology. DOI: 10.1016/j.cmp.2026.04.001. https://www.sciencedirect.com/science/article/pii/S1874467226000103?via%3Dihub
