A Prior Knowledge Based Tumor and Tumoral Subregion Segmentation Tool for Pediatric Brain Tumors

TL;DR

This paper introduces a knowledge-based segmentation tool for pediatric brain tumors that outperforms deep learning models when pediatric data is scarce, especially for heterogeneous tumors like ATRT.

Contribution

The paper presents a novel knowledge-driven segmentation method tailored for pediatric brain tumors, addressing data scarcity and heterogeneity challenges.

Findings

Outperforms deep learning models on pediatric tumor segmentation tasks.

Effective in segmenting multiple tumor subregions, including ATRT.

Demonstrates superior performance with limited or no training data.

Abstract

In the past few years, deep learning (DL) models have drawn great attention and shown superior performance on brain tumor and subregion segmentation tasks. However, the success is limited to segmentation of adult gliomas, where sufficient data have been collected, manually labeled, and published for training DL models. It is still challenging to segment pediatric tumors, because the appearances are different from adult gliomas. Hence, directly applying a pretained DL model on pediatric data usually generates unacceptable results. Because pediatric data is very limited, both labeled and unlabeled, we present a brain tumor segmentation model that is based on knowledge rather than learning from data. We also provide segmentation of more subregions for super heterogeneous tumor like atypical teratoid rhabdoid tumor (ATRT). Our proposed approach showed superior performance on both whole tumor and subregion segmentation tasks to DL based models on our pediatric data when training data is not available for transfer learning.