Fig. 3From: Delta radiomics analysis of Magnetic Resonance guided radiotherapy imaging data can enable treatment response prediction in pancreatic cancerRobust analysis requires high spatial stability of features. Heat map (A) shows the distribution of Concordance Correlation Coefficient (CCC) for all quantified radiomic features, describing the robustness of the features to small changes in positions of the Region of Interest (ROI), calculated through translation of the ROI 1.5 mm in x and y. Results of this quantification for representative features with high CCC (Skewness, B) and lower CCC (C) are presented, showing a tighter distribution for repeated measurements for higher CCC. Bars denote distance from min to max value of the feature when ROI is shifted, with one bar for each patientBack to article page