EarlySign’s Algorithm for Identifying Colorectal Cancer Risk Validated


Medial EarlySign, a developer of machine learning tools for data-driven medicine, has announced the results of new research with Oxford University. The study provides further validation for Medial EarlySign’s ColonFlag algorithm platform to identify individuals at risk of having colorectal cancer and support other approaches to early detection, including screening and active case finding.

The peer-reviewed study, Evaluation of a Prediction Model for Colorectal Cancer: Retrospective Analysis of 2.5 Million Patient Records, published in Cancer Medicine, evaluated Medial EarlySign’s risk stratification model for colorectal cancer. The UK research follows previous ColonFlag study results reported in the U.S. and Israel.

Patients older than 40 years of age with full blood count data were risk-stratified using ColonFlag and followed up for a diagnosis of colorectal cancer over a range of time intervals. Researchers focused on the 18 – 24 month interval prior to diagnosis, as it provides a greater opportunity to intervene and modify prognosis than shorter intervals. Secondary analysis was measured at 3-6, 6-12, 12-18, and 24-36 months before diagnosis, using the same method. The researchers also undertook a case-control analysis and a cohort study of patients undergoing full blood count testing during 2012, to estimate predictive values.

The study showed that the area under the curve (AUC) for the 18-24 month period was 0.776 and that performance improves as the time interval reduces. Additionally, for the population risk-scored in 2012, the positive predictive value at 99.5% specificity was 8.8% (being diagnosed within the next two years), with negative predictive value of 99.6%.

“Application of the Medial EarlySign algorithm to routinely collected primary care data provides an additional means of identifying those at risk of colorectal cancer,” said Dr. Tim A. Holt, Senior Clinical Research Fellow of Oxford University. “Our findings may provide a role for the risk score as a tool to assist case findings in a range of settings, and could also help target individuals who do not take up invitations for fecal occult blood test lab testing, or who refuse colonoscopy.”

Get more information at EarlySign.com.