GE HealthCare’s AI can accurately predict patient responses to immunotherapy, as demonstrated in collaborative studies with VUMC and UME. The models, showcasing 70-80% accuracy, may guide more personalized treatment decisions, potentially minimizing side effects and costs. This innovation is set to enhance clinical trial selection and is part of GE’s larger immuno-oncology efforts, including novel PET tracer development, indicating a significant stride towards precision oncology care.
GE HealthCare has revealed supportive data indicating that Artificial Intelligence (AI) can play a crucial role in forecasting how patients will respond to immunotherapy treatments before they begin.
AI algorithms developed by GE HealthCare are proficient in anticipating patient reactions to immunotherapy treatments with an accuracy range of 70-80%, as demonstrated by a comprehensive study across various cancers. This breakthrough will be shared at the upcoming Society for Immunotherapy of Cancer (SITC) conference in San Diego, resulting from collaborative efforts between GE HealthCare, Vanderbilt University Medical Center (VUMC), and the University Medicine Essen (UME) in Germany. The AI technology, first honed using a database of over 3,000 immunotherapy patient records from VUMC, was further confirmed with data from 4,000 UME patients. These models are adept at forecasting treatment effectiveness and predicting possible adverse effects for individuals, potentially guiding more precise and personalized treatment options early on, while also avoiding unnecessary side effects and costs.
Given that immunotherapy treatments have the potential to outperform traditional cancer treatments and that the current landscape boasts approximately 5,000 immunotherapies under development, these AI models could be pivotal. They offer the promise of enhancing the selection process for candidates in clinical trials, potentially expediting and improving the odds of successful outcomes. Post-regulatory approvals in various regions, GE HealthCare aspires to market these models for advancing drug development and aiding in clinical decision-making.
The creation of these AI models involved retrospective analyses by GE HealthCare and VUMC, linking the treatment outcomes of VUMC cancer patients to a vast array of anonymized data, including demographic, genomic, and other clinical factors. The models are designed to input routine patient electronic health record (EHR) data, which ensures they are adaptable and scalable for broad usage.
Dr. Travis Osterman of VUMC highlighted the importance of these AI models in enhancing patient selection for immunotherapy, aiming to reduce associated risks and costs. Moreover, Dr. Jens Kleesiek of UME underscored the successful application of these models across continents, marking a significant step towards their practical implementation for cancer patients undergoing immunotherapy.
Julia Casey, at the helm of GE HealthCare’s Molecular Imaging in the Pharmaceutical Diagnostics division, expressed optimism about partnering with stakeholders in the healthcare and pharmaceutical sectors. She emphasized the long-term goal of refining and applying AI models to advance therapy development and clinical applications.
These AI models are central to GE HealthCare’s burgeoning immuno-oncology program, which includes the innovation of new PET tracers. The company is at the forefront of a Phase I clinical trial for a novel fluorine-18 PET radiotracer, which is particularly geared towards identifying CD8+ T cells, the white blood cell subset critical to the success of most immunotherapies.
A trailblazer in the field of Pharmaceutical Diagnostics, GE HealthCare’s segment is responsible for the global supply of imaging agents pivotal for about 100 million medical procedures annually. Their Molecular Imaging portfolio is a blend of established and pioneering products designed to support accurate diagnoses and monitoring, thereby enriching therapy decisions and clinical outcomes across multiple medical specializations.
