Analyzing Deep Learning Based Brain Tumor Segmentation with Missing MRI Modalities

TL;DR

This paper compares deep learning methods for brain tumor segmentation with missing MRI modalities, highlighting ACN's superior performance when T1c is absent and exploring combinations of mmGAN and DeepMedic.

Contribution

It provides a comparative analysis of existing approaches and introduces a more stable version of mmGAN, along with discussions on future research directions.

Findings

ACN outperforms others when T1c is missing

Combination of mmGAN and DeepMedic is effective with single missing modality

Open-sourced a more stable version of mmGAN

Abstract

This technical report presents a comparative analysis of existing deep learning (DL) based approaches for brain tumor segmentation with missing MRI modalities. Approaches evaluated include the Adversarial Co-training Network (ACN) and a combination of mmGAN and DeepMedic. A more stable and easy-to-use version of mmGAN is also open-sourced at a GitHub repository. Using the BraTS2018 dataset, this work demonstrates that the state-of-the-art ACN performs better especially when T1c is missing. While a simple combination of mmGAN and DeepMedic also shows strong potentials when only one MRI modality is missing. Additionally, this work initiated discussions with future research directions for brain tumor segmentation with missing MRI modalities.