Federal Project Grant Award Summary

The University of Texas MD Anderson Cancer Center received a $421,685 Project Grant from the National Cancer Institute (NCI) under the Cancer Cause and Prevention Research program (CFDA 93.393) on June 1, 2026, with completion targeted for May 31, 2028. This award supports the development of novel statistical methodologies and software tools to enhance risk assessment for rare cancer patients at individual medical institutions by adaptively leveraging data from external population-based cohorts. The research addresses a critical gap in current transfer-learning approaches by accommodating scenarios where target cohorts possess more detailed covariates—such as newly developed biomarkers and comprehensive genetic information—than source cohorts derived from cancer registries or research consortiums.

The project deliverables include: (1) development of a transfer-learning-based Cox proportional hazards model that accommodates missing covariates in source cohorts to enable effective information borrowing; (2) creation of a group-specific transfer-learning Cox model that permits flexible, heterogeneity-aware information borrowing at the subgroup level; and (3) dissemination of publicly available, user-friendly software packages to ensure reproducibility and broad application across multiple datasets. Although the methodology is disease-agnostic, the research will demonstrate application to rare cancer populations, directly supporting the NCI's Cancer Cause and Prevention Research program objectives by advancing screening, early detection, and risk assessment capabilities.