Cancer Digital Insights

About

The Cancer Digital Insights (CDI) Working Group is a collaborative initiative between the Cancer Center at Illinois (CCIL) and the National Center for Supercomputing Applications (NCSA), aimed at advancing cancer research at Illinois through data science, artificial intelligence, and machine learning.

To support this mission, the group launched CDI Planning Grants, which fund new, high-impact projects that leverage the combined strengths of CCIL and NCSA. Each award provides up to $40,000 in project support, GPU computing time, and expert guidance to help research teams explore innovative, data-driven approaches with strong potential for external funding and community benefit.

2025 CDI Planning Grants

Proposal Title: “An Interpretable AI Foundation Model for Predicting Immunotherapy Response across Cancer Types”
PI – Kun Wang (CCIL)
Co-PIs – Shirui Luo (NCSA) & Joseph Chan (Memorial Sloan Kettering Cancer Center)
This project will develop an interpretable AI model trained on cross-cancer immunotherapy datasets to improve prediction accuracy and generalizability across diverse patient cohorts—supporting more precise identification of who is most likely to benefit from immunotherapy.

Proposal Title: “A Spatially Aware AI-Augmented Retrieval Framework for Spatial Omics Data”
PI – Aiman Soliman (NCSA)
Co-PIs – Zeynep Madak-Erdogan (CCIL) & Volodymyr Kindratenko (NCSA)
This team will build a spatially aware retrieval-augmented AI framework paired with a high-throughput spatial omics pipeline to help researchers navigate and interpret large collections of spatial images and pathway patterns through efficient, conversational querying.

Proposal Title: “Virtual Spatial Proteomics for Colorectal Cancer Risk Stratification”
Yang Liu (CCIL) & Weihao Ge (NCSA), with Mohith Manjunath (NCSA) and team
This project will use label-free quantitative phase imaging (QPI) and AI to generate “virtual protein maps” as a more scalable alternative to high-cost spatial proteomics—advancing tools to identify recurrence-linked features in colon polyps and improve risk stratification.