Pyradiomics is a Python library that analyses the histogram and grey scale patterns in the areas of scans which have been described as a Region of Interest. The radiomics output is a large spreadsheet where each ROI has been analysed and numbers recorded for particular features. The current spreadsheet has ~132 columns, but pyradiomics can produce >4000 columns if configuration parameters are changed.
Machine learning is required for the systematic correlation of these numerous feature values with clinical data.
OnkoDICOM executes Pyradiomics analysis with the click of a button, or automatically during the anonymisation process.