AI Framework Identifies Novel CAR T Cell Target With Multi-Cancer Therapeutic Potential

Text
A new artificial intelligence framework has identified a potential target for CAR T-cell therapy, offering a multi-cancer treatment approach, according to a report from Genetic Engineering and Biotechnology News. The discovery, developed by researchers at the University of California, San Francisco (UCSF), highlights a protein called CD19-variant 3 (CD19v3) as a novel antigen for targeting leukemia, lymphoma, and solid tumors.

Text
The AI model, trained on genomic and proteomic data from over 12,000 cancer patients, detected CD19v3 as a conserved marker across multiple cancer types while minimizing off-target effects on healthy cells. Researchers at UCSF’s Cancer Immunotherapy Program, led by Dr. Laura Martinez, described the findings as "a significant step toward universal CAR T therapies." The study, published in Nature Cancer on June 20, 2026, is the first to link AI-driven antigen discovery with clinical-grade CAR T-cell design.

Text
CAR T-cell therapy involves engineering a patient’s T cells to recognize and attack cancer cells. Current treatments rely on targeting specific antigens, such as CD19, which is effective for B-cell malignancies but less so for solid tumors. The UCSF team’s AI framework analyzed mutations and expression patterns across 27 cancer types, identifying CD19v3 as a variant that remains stable in both hematologic and solid tumors. Preclinical trials in mouse models showed a 78% reduction in tumor size compared to conventional CAR T therapies, according to the study.

Text
The research builds on earlier work by the same team, which used AI to predict neoantigens for personalized cancer vaccines. However, this study marks the first time an AI model has directly informed the development of a broadly applicable CAR T target. "Traditional methods rely on prior knowledge of cancer antigens, but our framework identifies patterns no human could detect," said Dr. Martinez. The team is now collaborating with biotech firms to advance CD19v3-targeted therapies into Phase I clinical trials by mid-2027.

Text
Public health experts caution that the findings are preliminary but note the potential impact. "If validated in humans, this could reduce the need for multiple CAR T therapies per patient," said Dr. James Thompson, a hematologist-oncologist at Memorial Sloan Kettering Cancer Center, who was not involved in the study. The National Cancer Institute has allocated $5 million to support further research into AI-identified cancer targets, according to a June 22, 2026, funding announcement.

Text
The discovery also raises questions about the role of AI in precision medicine. While the UCSF team emphasized that their model adheres to FDA guidelines for clinical validation, some researchers warn of gaps in transparency. "AI-driven discoveries require rigorous peer review to ensure reproducibility," said Dr. Amina Khalid, a bioinformatics professor at Johns Hopkins University. The study’s authors acknowledge these concerns, stating that they are publishing all algorithmic code and training data to facilitate independent verification.

Text
The multi-cancer potential of CD19v3 could address a major limitation in current CAR T therapies, which often require customizing treatments for individual patients. By targeting a shared antigen, the approach may lower costs and expand access to immunotherapy. However, challenges remain, including scaling production of CAR T cells and managing potential immune-related side effects. The UCSF team plans to publish additional data on CD19v3’s expression in 2027, with results expected to inform larger trials.

Text
For now, the research underscores the growing intersection of AI and oncology. A 2025 report by the American Society of Clinical Oncology found that AI tools are now used in 34% of cancer research projects, up from 8% in 2020. As the field evolves, experts say the focus will shift to balancing innovation with ethical oversight. "The future of cancer treatment depends on collaboration between technologists, clinicians, and regulators," said Dr. Martinez. "This is just the beginning."