Researchers at Peter MacCallum Cancer Centre in Melbourne have developed an electronic medical record-based algorithm to help identify cancer patients who have developed immune-related colitis, a significant side effect associated with immunotherapy.
The work has been led by the National Centre for Infections in Cancer, in collaboration with Peter Mac clinicians and the Centre of Health Services Research in Cancer. The team is developing a broader digital surveillance platform to monitor immune-related adverse events that can occur during cancer treatment.
The newly introduced tool uses a digital phenotype, described as a reproducible computer algorithm based on EMR data. It has been designed to identify affected patients with high accuracy by detecting multiple clinical indicators across routinely collected records. These include pathology results, colonoscopy or sigmoidoscopy procedures, imaging reports that mention colitis in CT or PET scans, and treatment data such as medication orders and administrations of steroids or immunosuppressants.
The research, published in JCO Clinical Cancer Informatics, has been described as the first validated digital phenotype of its kind involving cancer patients in Australia.
Immune checkpoint inhibitors are used across a broad range of cancers, but they can cause immune-related adverse events, including colitis. The condition can affect a substantial proportion of patients receiving immunotherapy. At present, clinicians often identify cases through manual review, a process that can be time-consuming and difficult to scale.
Dr Jasmine Teng, an infectious disease doctor and doctoral fellow with the National Centre for Infections in Cancer, said the tool was developed for retrospective surveillance rather than immediate clinical alerts. “It is not a predictive tool, nor a tool for near-real-time case finding,” she said.
The researchers believe the digital phenotype could support more accurate estimates of immune-related colitis incidence and provide a foundation for future clinical research. “In addition to surveillance, we envisage that this digital phenotype could assist with developing risk prediction tools, biomarker research, and support other work, such as mapping of the patient care trajectory,” Dr Teng said.
The team is now planning external validation with other centres and aims to expand its surveillance work. “We are keen to develop a surveillance platform to incorporate other immune-related adverse events,” Dr Teng added.
Click here for the original news story.
