GE Healthcare’s AI models can predict patient responses to immunotherapy

GE Healthcare has revealed promising data showcasing the effectiveness of its artificial intelligence (AI) models in predicting patient responses to immunotherapies. The study, published in the Journal of Clinical Oncology Clinical Cancer Informatics, focused on leveraging electronic health record (EHR) data to forecast the efficacy and toxicity of cancer immunotherapy. The AI models demonstrated an accuracy rate ranging between 70% and 80% in a pan-cancer cohort, drawing primarily from routinely collected structured data such as diagnosis codes and medication information available in medical records. By utilising these existing inputs, GE Healthcare aims to develop models applicable in various clinical settings.

The collaboration between GE Healthcare and Vanderbilt University Medical Center involved retrospective analysis and correlation of immunotherapy responses for thousands of cancer patients. Deidentified demographic, genomic, tumour, cellular, proteomic, and imaging data were used to train the AI models. The predictive capabilities of the models extend to assessing the likelihood of an individual patient experiencing adverse reactions, offering potential benefits in selecting appropriate treatment pathways earlier, minimising unnecessary side effects, and reducing costs. 

GE Healthcare envisions widespread deployment and adoption of these models, with plans to seek regulatory authorisation for potential applications in clinical decision support and drug development. The company intends to collaborate with pharmaceutical companies, researchers, and clinicians to optimise and apply the AI models in therapy development and clinical practice.