Letter to the editor: radiomics analysis for predicting pembrolizumab response in patients with advanced rare cancers

Abstract

A commentary on the original research article: ‘Radiomics analysis for predicting pembrolizumab response in patients with advanced rare cancers’. Of note, the predictor selection process, the cross-validation method, along with the lack of final testing of the developed model with a separated data set may mask overfitting, overestimating performance metrics.