AI for Precision Cancer Treatment

Cancer is one of the deadliest diseases, responsible for nearly 10 million deaths annually, or about one in six deaths globally. In the U.S. in 2025, there were over 2 million new cases and approximately 618,000 deaths related to cancer. Two major factors contribute to this high death rate. First, early-stage detection remains insufficient, limiting the ability to identify tumors when they are smaller, less aggressive, and more responsive to treatment. Second, many current therapies, including radiation and chemotherapy, rely on inducing chemical stress to destroy cancer cells. These approaches often damage healthy tissues, leading to significant side effects, and tumors can adapt by developing protective mechanisms that reduce treatment efficacy over time.These challenges highlight the urgent need for new strategies for cancer treatment and early-stage cancer detection.

Single-atom catalysts (SACs), in which isolated metal atoms are atomically dispersed on supports, have emerged as promising candidates for enhancing therapeutic responses by modulating the tumor microenvironment. SACs in cancer therapy are rapidly surging but remain mostly at the preclinical stage (cell and animal studies). The team will coherently integrate quantum computing (QC) and machine learning (ML) into the rational design of SACs for cancer detection and therapy. QC enables the accurate simulation of strongly correlated electronic structures and catalytic reaction pathways beyond the reach of classical methods, thereby yielding more reliable databases. With reliable databases, ML models will efficiently navigate millions of candidate configurations to predict new SAC structures with high catalytic activity for cancer detection and therapy. Results will be disseminated through high-impact publications, interdisciplinary conferences, and open release of benchmark datasets and reproducible computational tools. This open-science approach will increase visibility and encourage other research groups to adopt the predictive framework.

This project addresses public health by enabling the predictive design of catalytic cancer therapies that amplify treatment effectiveness while reducing the time and cost required to discover new therapeutic materials. At the regional level, the project will strengthen collaboration between computational scientists, materials researchers, and cancer medicine experts. Nationally, the work will accelerate the development of more effective cancer therapies while reducing reliance on costly trial-and-error discovery, supporting U.S. leadership in advanced computing and AI for healthcare applications. Globally, catalytic platforms that enhance existing treatments, such as radiotherapy, could improve treatment accessibility and effectiveness, particularly in regions with limited advanced medical infrastructure. The predictive design approach developed in this project could transform how catalytic cancer therapies are discovered, enabling safer and more effective treatments that ultimately improve survival outcomes for millions of patients worldwide.