Overview
For patients with recurrent or metastatic head and neck cancer, doctors often choose between treating with the immunotherapy drug pembrolizumab alone or combining it with chemotherapy. While both approaches are standard options, there is currently no direct evidence from clinical trials to help clinicians decide which patients need the added intensity of chemotherapy. Because chemotherapy can provide rapid relief but also carries a higher risk of side effects, finding a way to identify which individuals will derive the most benefit from the combination is essential for personalizing care.
Researchers used the Flatiron Health Research Database to analyze data from over 1,700 patients and developed a machine learning model to predict a patient’s initial prognosis. The model evaluated various clinical factors, such as weight loss, laboratory results, and where the cancer had spread. The study found that patients identified as being at the highest risk saw a significant survival benefit when chemotherapy was added to their immunotherapy. However, for the majority of patients who had a more favorable initial outlook, adding chemotherapy provided no extra survival benefit.
Why this matters
This research demonstrates that artificial intelligence can be used to potentially help clinicians make informed treatment decisions for head and neck cancer. By pinpointing the specific subgroup of patients who gain the most from combination therapy, clinicians can ensure patients receive the appropriate level of care. Ultimately, this approach supports a more precise model of cancer care that better balances treatment effectiveness with a patient's quality of life.