Fast and Accurate Semi-Automatic Segmentation Tool for Brain Tumor MRIs

TL;DR

This paper introduces MITKats, an open-source semi-automatic segmentation tool for brain tumor MRIs that is both fast and accurate, significantly reducing segmentation time while maintaining high precision.

Contribution

MITKats is a new open-source tool that offers rapid and accurate semi-automatic brain tumor segmentation, outperforming existing algorithms in speed with comparable accuracy.

Findings

Segmentation accuracy approaches expert-level ground truths.

MITKats reduces segmentation time to 5 minutes per case.

Performance is 4 to 11 times faster than competing methods.

Abstract

Segmentation, the process of delineating tumor apart from healthy tissue, is a vital part of both the clinical assessment and the quantitative analysis of brain cancers. Here, we provide an open-source algorithm (MITKats), built on the Medical Imaging Interaction Toolkit, to provide user-friendly and expedient tools for semi-automatic segmentation. To evaluate its performance against competing algorithms, we applied MITKats to 38 high-grade glioma cases from publicly available benchmarks. The similarity of the segmentations to expert-delineated ground truths approached the discrepancies among different manual raters, the theoretically maximal precision. The average time spent on each segmentation was 5 minutes, making MITKats between 4 and 11 times faster than competing semi-automatic algorithms, while retaining similar accuracy.