Chemotherapy has long been the standard treatment for cancer, but in recent years, a revolutionary alternative called immunotherapy has gained momentum. Unlike chemotherapy, which attacks both healthy and cancerous cells, immunotherapy leverages the patient’s own immune system to target and fight cancer. While it doesn’t work for everyone, the results have been life-saving for 15% to 20% of patients who respond positively to the treatment.
However, as with any medication, immunotherapy comes with potential side effects. In fact, studies show that 10% to 15% of patients develop “significant toxicities.” To address this issue, GE HealthCare, in collaboration with Vanderbilt University Medical Center, has developed an artificial intelligence (AI) model to help predict patients' response to immunotherapy and identify potential side effects.
Over the course of five years, the AI model was trained using thousands of patients' electronic health records (EHRs). By analyzing patterns in how these patients responded to immunotherapy, the model focused on safety and effectiveness. The data pulled from the EHRs included demographic information, imaging scans, preexisting diagnoses, lifestyle habits, medication history, and more.
“The model predicts which patients are likely to derive the benefit from immunotherapy versus those patients who may not,” explained Jan Wolber, global digital product leader at GE HealthCare. “It also predicts which patients have a likelihood of developing one or more significant toxicities.”
In a study, the AI model demonstrated an accuracy rate of 70% to 80% in predicting patients' responses to immunotherapies, as reported in the Journal of Clinical Oncology Clinical Cancer Informatics. Wolber emphasized that while the models are not perfect, the results are promising, and implementation in the clinic requires minimal additional effort since the data points are already being collected during regular patient visits.
Dr. Travis Osterman, a medical oncologist and associate chief medical information officer at Vanderbilt University Medical Center, highlighted the natural progression of using AI in medicine. Instead of relying solely on patient surveys, the AI model leverages the entirety of a patient’s medical record to identify risk factors and potential treatment outcomes. Osterman expressed concern about missing patients who could have significantly benefited from immunotherapy, as well as avoiding harm to those who might experience severe side effects.
Dr. Marc Siegel, clinical professor of medicine at NYU Langone Medical Center, considers AI models like this one to be the future of personalized medicine. He believes that AI can enhance treatment options and provide more predictive outcomes, particularly when combined with genetic and protein analysis. However, he emphasized that physicians and scientists should remain in control of decision-making.
While the AI model shows promise, it does have limitations. Wolber acknowledged that the models do not provide 100% accuracy and may produce false positives or false negatives. Additionally, the dataset used for training is relatively small, and the team is exploring opportunities to incorporate larger datasets for more refined predictions. Integrating AI recommendations into the healthcare workflow also poses a challenge that the medical community is currently grappling with.
Looking ahead, after obtaining necessary regulatory approvals, GE HealthCare plans to make the AI model widely available for clinicians. It may also expand into other areas of care, such as neurology or cardiology, and could potentially be integrated into drug development. The hope is that by tailoring treatments to individual patients, researchers can increase efficacy while minimizing the risks associated with certain therapies.
Dr. Osterman describes this time as the golden age of AI recommendations in medicine. As AI becomes more prevalent, it is likely here to stay. The potential to enhance treatment options and improve patient outcomes is a cause for excitement in the medical community. As Dr. Osterman aptly puts it, “it’s a really exciting time to be in medicine.”
Use the share button below if you liked it.