MELANO-PREDICT: development of a clinically applicable algorithm for prediction of checkpoint inhibitor response in melanoma

Principal Investigator: Ian Watson, Hamed S. Najafabadi, John Stagg
Theme : Artificial Intelligence (AI) / Health
Competition : Omics Data Against Cancer (ODAC) Competition
Status : Completed
Start : Oct. 1, 2020
End: Sept. 30, 2022
Budget : $300,000.00



Melanoma is the deadliest form of skin cancer. Prior to 2010, very few people survived melanoma when it spread to distant parts of the body. But the research community developed therapies that boost the patient’s immune system to combat melanoma. This therapy can produce cures for some patients. Unfortunately, many melanoma patients do not respond to these immune therapies or have severe side effects. Currently, there is no effective clinical test to predict whether a patient will benefit from immune therapy. Such a test would be valuable to help doctors prescribe therapy that is most likely to benefit patients and one that will have the least number of side effects.



More and more, cancer patients are having their DNA sequenced to determine treatment options. However, information from these tests is rarely used to help guide prescriptions of immune therapy. The main reason for this is because the scientific community does not know which DNA sequences actually have information that is useful to predict an immune therapy response. Our proposal will address this problem, by using artificial intelligence research techniques to figure out the combination of DNA sequences that will predict a response to immune therapy. We believe that the research results from this project will not only help melanoma patients, but also other cancer patients that are currently not receiving immune therapy that could benefit from such treatment.


Lead Genome Centre: Génome Québec 

Partners: IVADO, Oncopole


Kam Kafi Imagia