Figure 2

Visualization of the association between pathway alterations and clusters based on FDG PET features in SQCC. (A) Patient clustering according to PET image features. The pathways on the right shows different incidence of alterations according to the clusters. (B) Heatmaps showing z-score of representative PET image features that were chosen based on statistical difference between clusters. CNV, copy number variation, MUT, mutation; NA, non applicable; T, texture feature; TFCCM, texture feature coding co-occurrence matrix; GLCM, gray level co-occurrence matrix; SUV, standardized uptake value; SD, standard deviation; GLSZM, gray level size zone matrix; NGLD, neighboring gray level dependence.