as immunotherapy are often recommended.
Immunotherapy is a new type of cancer treatment which stimulates the body’s immune system to kill cancer cells. These drugs typically benefit ~50% of patients, and can be successful when other therapies fail. Despite these benefits, immunotherapy infrequently causes severe and/or permanent side effects. In this light, it is very important to be able to predict which patients will respond to immunotherapy, although our ability to do this is currently limited.
Our aim is to develop a genetic test of a patient’s melanoma to determine which patients will respond to immunotherapy. This will enable a more personalised approach to treatment, with immunotherapy only prescribed in patients predicted to respond. Our preliminary work demonstrates our ability to predict patient outcomes using advanced machine learning techniques, and our goal in this project is to refine these methods to increase the accuracy of our predictions. The best method to increase accuracy of predictions is to train machine learning models on large, well-curated datasets, such as are available through Melanoma Research Victoria
This project is a collaboration between Mark Shackleton (Director of Oncology, The Alfred Hospital) and the GMDx Group, a biotechnology company that has patented a method to predict immunotherapy response from easily-obtained genetic information about a patient’s cancer. We plan to undertake genetic sequencing of melanomas from 250 patients that either did or didn’t respond to immunotherapy.
We expect that by bringing together the GMDx Group’s patented therapy response prediction machine learning methods and data from the MRV cohort, we will develop a test to predict accurately a patient’s response to immunotherapy.