Fig. 2From: Radiomic signatures from T2W and DWI MRI are predictive of tumour hypoxia in colorectal liver metastasesAn illustrative diagram highlighting the (a) study design as well as the (b) radiomic/machine learning workflows. Expert readers delineated pre-resection MRI images, and from these regions of interest, handcrafted radiomic features are extracted. Machine learning classifier selection, training, and hyperparameter optimisation were performed using an evolutionary genetic approach. The candidate model is then evaluated on its predictive performance with an unseen validation cohortBack to article page