Partial Information Decomposition as a Multi-Contrast 3D MRI Selection Strategy for Resource-Constrained Deep Neural Network Training in Brain Tumor Segmentation

A pre-training Partial Information Decomposition framework that ranks multi-contrast 3D MRI input pairs by their redundant, unique, and synergistic information about regional tumor burden selected T1c+T2-FLAIR as the optimal pair for brain tumor segmentation. When eleven lightweight 3D U-Nets were trained with different input configurations, T1c+T2-FLAIR achieved a mean Dice of 0.676, ranking second overall compared to 0.687 for all four inputs, demonstrating the framework's value for identifying compact, informative MRI sets before costly model development.

arXiv:2607.15396v1 Announce Type: new Abstract: Multi-contrast 3D MRI segmentation can be computationally demanding when all available sequences are used. We evaluate a pre-training Partial Information Decomposition framework that ranks input pairs according to their redundant, unique, and synergistic information about regional tumor burden and selects the highest-ranked pair for downstream training. Applied to T1n, T1c, T2w, and T2-FLAIR MRI, the framework selected T1c+T2-FLAIR. We then trained eleven architecturally identical lightweight 3D U-Nets using different input configurations. On an independent test cohort, T1c+T2-FLAIR was the strongest two-input configuration and ranked second overall in mean Dice 0.676 versus 0.687 for all four inputs . Independent Shapley analysis on the full-input model also identified T2-FLAIR and T1c as the most influential inputs and their pairwise interaction as the strongest. These findings demonstrate the practical value of PID based pre-training selection for identifying compact, informative MRI input sets before costly 3D model development.