Pangea Biomed has published data from a study showing that its AI-powered Enlight-DeepPT (Englight-DP) system can predict therapeutic responses in cancer patients using histopathology images.
The study was conducted in partnership with the Australian National University and the US National Cancer Institute, and the data was published in Nature Cancer journal.
The Enlight-DP platform uses AI and deep learning models to predict treatment responses across multiple cancer types via a two-step approach. First, the DeepPT platform infers gene expression and genome-wide tumour mRNA expression from haematoxylin and eosin (H&E)-stained histopathology slides. The Enlight platform then predicts treatment responses based on the inferred expression values.
"Enlight-DP bypasses the data availability limitations that hinder existing approaches by eliminating the need for dedicated training on new cohorts for each drug treatment," said Pangea’s CTO Ranit Aharonov, who co-led the study.
"This versatile solution can be applied across various cancer types and therapies, potentially transforming clinical practices and significantly improving patient outcomes."
The study evaluated Enlight-DP’s efficacy across five independent patient cohorts involving four different treatments spanning six cancer types. The platform had an overall odds ratio of 2.28 and a 39.5% increased response rate among predicted responders, compared to the baseline rate.
Pangea noted that Enlight-DP’s prediction accuracy “obtained without any training on the treatment data, is comparable to that achieved by directly predicting the response from the images, which requires specific training on the treatment evaluation cohorts”. Furthermore, the study demonstrated that the underlying DeepPT framework is superior to other available methods in predicting tumour mRNA expression.
Pangea is also planning further validation and prospective testing to strengthen Enlight-DP’s clinical utility and file for regulatory approval.
The use of AI in the diagnostic imaging market is set to rapidly grow and is expected to exceed $1.2bn by 2027, up from $336m in 2022 as per GlobalData analysis. GlobalData also projects the global revenue for AI platforms across healthcare to reach an estimated $18.8bn by 2027.
The use of AI in medical imaging and diagnostics received significant investment in the last year, with multiple companies raising capital to advance the use of AI in imaging. In October 2023, Philips and Quilbim partnered to develop AI-based imaging and reporting solutions for MR [magnetic resonance] prostate examinations. The collaboration leveraged Philips’ high-speed MR imaging and Quibim’s QP-Prostate software to streamline diagnosis.
