Improved prediction of MAPKi response duration in melanoma patients using genomic data and machine learning

Baseline genomic data have not demonstrated significant value for predicting the response duration to MAPK inhibitors (MAPKi) in patients with advanced BRAFV600-mutated melanoma. We used machine learning algorithms and pre-processed genomic data to test whether they could contain useful information to improve the progression-free survival (PFS) prediction. This exploratory analysis compared the predictive performance of a dataset that contained clinical features alone and supplemented with baseline genomic data. In the evaluation set (two cohorts, n = 111), the cross-validated model performance improved when pre-processed genomic data, such as mutation rates, were added to the clinical features. In the validation dataset (two cohorts, n = 73), the best model with genomic data outperformed the best model with clinical features alone. Finally, our best model outperformed with baseline genomic data, increasing the number of patients with a correctly predicted relapse by between +12% and +28%. In our models, baseline genomic data improved the prediction of response duration and could be incorporated into the development of predictive models of MAPKi treatment in melanoma.