A new research alliance aims to use artificial intelligence to map the intricate geography of tumors, seeking to reveal novel drug targets for cancers that have been difficult to treat. The collaboration combines advanced spatial biology with deep protein analysis to understand how cells interact within a tumor's ecosystem.
The partnership integrates Vanderbilt Health’s Molecular AI Initiative, which creates high-resolution spatial maps of tumors, with Bertis’s proprietary AI-driven proteomics platform. Traditional bulk tissue analysis often misses critical details about how tumor, immune, and connective tissue cells are organized and communicate. The new approach uses AI to visualize these spatial relationships, isolating cell populations responsible for treatment response or resistance. Bertis then conducts deep proteomic profiling on these pinpointed areas, applying computational models to identify the most promising and druggable protein targets.
The initial focus will be on HER2-low tumors, a subtype of breast and other cancers that express low levels of the HER2 protein and have historically had fewer targeted treatment options. The teams aim to discover cell surface proteins suitable for next-generation therapies like antibody-drug conjugates. "Identifying therapeutic targets and understanding treatment response require a precise view of proteins, spatial context and tumor biology," said Dr. Tae Hyun Hwang of Vanderbilt Health, who is leading the effort. He noted the collaboration is designed to generate high-confidence targets to support future drug development.
Looking ahead, the partners plan to expand their research into additional cancer types based on early data. The initiative represents a hopeful step toward translating complex tumor maps into tangible clinical advances, potentially opening new therapeutic possibilities for patients with limited options.