Automated EMR-Based Tool Detects Immune-Related Colitis in Immunotherapy Patients

The condition is common and can be serious, disrupting treatment plans and driving hospital admissions. Rapid case identification across oncology services is challenging because manual chart review is slow and inconsistent. To help address this challenge, a newly introduced technology aims to use routine hospital data to quickly flag affected patients for timely evaluation.

At Peter MacCallum Cancer Centre (Peter Mac; Melbourne, Australia), investigators have developed an Australian-first digital tool to identify immune-related colitis in patients receiving immune checkpoint inhibitors. Immune-related colitis can affect up to 50% of patients on these therapies, underscoring the need for scalable surveillance. The new approach replaces labor‑intensive case‑note review with automated detection, supporting faster recognition and more consistent case ascertainment across clinical teams.

The technology, presented as a clinician‑verified digital phenotype, is a reproducible computer algorithm that uses existing electronic medical record (EMR) data to identify patients with high accuracy. By drawing only on routinely captured data, the method can be implemented without additional clinical workflows. This design supports standardization within and potentially across health systems, allowing care teams to detect at‑risk patients earlier in their course.

Development and validation were led by the Centre of Health Services Research in Cancer (CHSRC) and the National Center for Infections in Cancer (NCIC) in collaboration with clinicians at Peter Mac. It is the first digital phenotype of its kind to be validated for cancer patients in Australia. The work is published in JCO Clinical Cancer Informatics on June 16, 2026. 

According to the research team, the tool will help define the true incidence of immune‑related colitis, clarify care pathways in severe cases, and inform decisions about immune checkpoint inhibitor retreatment. The ability to assemble large, well‑characterized EMR‑derived cohorts also opens opportunities to partner with scientists on biomarker discovery that could guide risk stratification and earlier management.

Related Links
Peter MacCallum Cancer Centre