Early results from a groundbreaking study conducted by researchers at Virginia Tech suggest that a simple saliva test, combined with advanced artificial intelligence (AI) algorithms, could help identify cancer patients at high risk for severe side effects from the chemotherapy drug 5-fluorouracil (5-FU). This drug, which has been in use since the 1950s, is a staple in cancer treatment, particularly for colorectal, breast, head and neck, pancreatic, and stomach cancers. However, about one in three patients has a genetic predisposition that hinders their body's ability to break down 5-FU, leading to toxic buildup and potentially life-threatening complications. The study, led by Carla Finkielstein, a professor at the Fralin Biomedical Research Institute at Virginia Tech, aims to add a layer of precision to 5-FU treatments by detecting mutations in the DPYD gene, which produces the enzyme crucial for breaking down the drug. Although the FDA has recommended genetic screening for 5-FU, it is not widely implemented, making the development of a simple and reliable saliva test particularly significant. The research team, including John Janiga, Dzenis Mahmutovic, Katherine Brown, Dulguun Myagmarsuren, Clinton Roby, and Douglas Grider from the Molecular Diagnostics Laboratory, and Mark Kochenderfer from Blue Ridge Cancer Care, initially focused on analyzing DNA from healthy volunteers and cancer patients to determine if saliva samples could accurately detect known DPYD mutations. Their preliminary results were positive, confirming that saliva samples are a viable alternative to blood samples for genetic testing. To further enhance the study's scope, the team employed advanced AI tools and 3D protein modeling to evaluate thousands of samples. This innovative approach allowed them to assess the structural and functional impact of previously unidentified mutations in DPYD, uncovering potentially harmful variants that conventional methods might have missed. Katherine Brown, the lead bioinformatician, highlighted the effectiveness of the AI-driven method in identifying new pathogenic mutations, particularly in colon cancer patients. The researchers have identified several newly recognized DPYD mutations in colon cancer patients that are predicted to impair 5-FU metabolism. Two of these mutations were labeled as "pathogenic" by multiple predictive tools and were confirmed by a clinical genetics database. The discovery of high-risk mutations in healthy individuals underscores the importance and feasibility of implementing broad population-level genetic screening using saliva samples. Finkielstein emphasized the potential benefits of early risk detection: "If we can flag high-risk patients early, we can tailor their treatment plans, reduce hospitalizations, and potentially avoid fatal complications. This is a real step forward in making cancer care safer, smarter, and more personalized." The team's findings could lead to expanded genetic testing and more tailored chemotherapy regimens, ultimately improving patient outcomes. The preliminary findings were presented at the 2025 ASCO Gastrointestinal Cancers Symposium and published as an abstract in the Journal of Clinical Oncology, indicating early interest in the study’s clinical potential. The collaboration with oncologists and pathologists has been crucial, providing essential clinical insights that have shaped the research direction. Industry experts are optimistic about the potential applications of this saliva test and AI approach. The combination of non-invasive sample collection and sophisticated AI analysis could revolutionize the way chemotherapy is administered, making it safer and more effective for a broader range of patients. Virginia Tech's Molecular Diagnostics Laboratory, known for its cutting-edge research in personalized medicine, is well-positioned to drive further advancements in this field. The study also highlights the growing importance of interdisciplinary research in addressing complex medical challenges.