Deep Learning Models for CT Image Standardization

Multicentric CT imaging studies often encounter images acquired with scanners from different vendors or using different reconstruction algorithms. This leads to inconsistencies in noise level, sharpness, and edge enhancement, resulting in a lack of homogeneity in radiomic characteristics. These inconsistencies create significant variations in radiomic features and ambiguity in data sharing across different institutions. Therefore, normalizing CT images acquired using non-standardized protocols is vital for decision-making in cross-center large-scale data sharing and radiomics studies. To address this issue, we present four end-to-end deep-learning-based models for CT image standardization and normalization. The first two models require paired training data and can standardize images acquired from the same scanner but with different non-standardized protocols. The third model requires unpaired training data and can standardize images from one protocol to another. The final model is more robust and can utilize both paired and unpaired data during training. It can be used to standardize images within a scanner or between scanners. All the models' performances were evaluated based on the radiomic features. Our experimental results show that the proposed models can effectively reduce scanner-related radiomic feature variations and improve the reliability of CT imaging radiomic features.