Fig. 2

Least absolute shrinkage and selection operator was used to select features and construct the SRRM. A CT images, cluster images, and three-dimensional-cluster results are presented. B, C Optimal sub-regional radiomics (n = 11) were selected in patients with non-small cell lung cancer who underwent immunotherapy. D ROC curves for immunotherapy response prediction using the developed SRRM in the training cohort. E Precision–recall curves for response prediction using the developed SRRM in the training cohort. SRRM sub-regional radiomics model, ROC receiver operating characteristic, CT computed tomography, AUC area under the curve