Reproducibility in medical image radiomic studies: contribution of dynamic histogram binning
This addresses reproducibility issues in medical image radiomics, which is crucial for reliable machine learning applications in healthcare, though it is incremental as it focuses on a specific preprocessing step.
The study identified that dynamic histogram binning in radiomic feature extraction increases sensitivity to annotation fluctuations, potentially affecting reproducibility in many recent radiomic studies, and suggested static binning as a remedy.
The de facto standard of dynamic histogram binning for radiomic feature extraction leads to an elevated sensitivity to fluctuations in annotated regions. This may impact the majority of radiomic studies published recently and contribute to issues regarding poor reproducibility of radiomic-based machine learning that has led to significant efforts for data harmonization; however, we believe the issues highlighted here are comparatively neglected, but often remedied by choosing static binning. The field of radiomics has improved through the development of community standards and open-source libraries such as PyRadiomics. But differences in image acquisition, systematic differences between observers' annotations, and preprocessing steps still pose challenges. These can change the distribution of voxels altering extracted features and can be exacerbated with dynamic binning.