# Stratify or Inject: Two Simple Training Strategies to Improve Brain Tumor Segmentation

## TL;DR

This paper proposes two simple training strategies that incorporate tumor grade information to improve brain tumor segmentation accuracy in deep learning models, demonstrated on the BRATS 2018 dataset.

## Contribution

Introduction of two novel training strategies that utilize tumor grade information to enhance brain tumor segmentation performance.

## Key findings

- Both strategies improve segmentation accuracy over baseline methods.
- Incorporating tumor grade reduces heterogeneity impact on model training.
- Strategies are validated on the BRATS 2018 dataset.

## Abstract

Deep learning methods for brain tumor segmentation are typically trained in an ad hoc fashion on all available data. Brain tumors are tremendously heterogeneous in image appearance and labeled training data is limited. We argue that incorporation of additional prior information, specifically tumor grade, associated with tumor imaging phenotypes during model training can significantly improve segmentation performance. Two strategies for incorporation of tumor grade during model training are proposed and their impact on segmentation performance is demonstrated on the BRATS 2018 dataset.