Cancer Research DataMart

The UVACCC Cancer Research DataMart is a new, AI-driven resource allowing researchers to rapidly query and analyze vast collections of patient and other data from within UVA and beyond. Its AI analysis tools allows users to ask questions in plain language, explore cohorts, and run feasibility analyses without writing code or waiting on chart abstraction. The system is fast, taking only minutes versus the months or even years these processes took previously.

Cancer Center member faculty can access the DataMart here or via the button at top right.

For a demonstration, to request access, or if you have any questions, please email Elizabeth Mulcahy, PhD, at elizabeth.mulcahy@uvahealth.org or click the button at lower right.

Using the DataMart’s Cohort Explorer, researchers can build and refine patient cohorts interactively, filtering across clinical variables, diagnoses, treatments, and outcomes. The tool provides a fast, visual way to work with the data, without the need for structured query language (SQL).

Researchers can now request specimens by exporting a patient population within the DataMart and sharing it with it with BTRF Director Pat Pramoonjago, PhD.

Developed by Manifold AI, the DataMart’s AI analysis agent can query across datasets and instantly generate visualizations and surface patterns that would otherwise require a data analyst or custom scripts:

A Sankey diagram generated using the DataMart’s AI analysis agent visualizes a breakdown of UVA breast cancer patients by stage and their corresponding treatments.

A Kaplan-Meier plot generated using the DataMart’s AI analysis agent comparing the overall survival probability of pituitary tumor patients who received surgery to those who did not.

The DataMart securely integrates and analyzes data from multiple internal and publicly available resources:

• UVA Cancer Registry
• UVA Biorepository and Tissue Research Facility (BTRF) biospecimens
• ORIEN
• Open Targets Platform
• ClinicalTrials.gov
• Drug Gene Interaction Database
• Cancer Genome Atlas Program (TCGA) RNASeq
• Genotype-Tissue Expression (GTEx) Portal
• Cancer Cell Line Encyclopedia

Researchers can also bring their own data — clinical trial cohorts, lab datasets, external registries — and ingest it directly into the DataMart to leverage its AI analysis tools. The DataMart team will work directly with research teams to load and ingest the data.

The DataMart is regularly incorporating new and valuable publicly available datasets. If you have suggestions for additional public datasets, please email Dr. Mulcahy.