Prediction of lung cancer immunotherapy response via machine learning of immune cell characteristics
What is it about?
Cancer immunotherapy targets the immune cells so that they can more efficiently attack tumor cells. Not every patient benefits, however, because some tumors can turn off the activity of the immune cells against them. It is not well understood why some tumors respond better than others to immunotherapy. This study analyzed immune cells from lung cancer patients who were treated with immunotherapy, to see what the differences might be between patients with tumors which responded better than for the other patients. The analysis applied a comprehensive set of machine learning methods that could automatically detect these differences.
Why is it important?
The study found that there exist combinations and proportions of immune cell types that can accurately indicate whether a patient’s tumor will respond well or not to the immunotherapy. In particular, markers associated with B-cells were identified as key features that were different between the patient groups. Next steps will be to evaluate the study findings in a larger group of patients, with the ultimate goal to help select the best immunotherapy for each and every patient.