Benchmarking of Deep Learning Methods for Generic MRI Multi-Organ Abdominal Segmentation

TL;DR

This paper benchmarks three deep learning models for MRI abdominal segmentation, introduces a CT-trained model as an alternative, and evaluates their accuracy and generalizability across diverse datasets.

Contribution

It provides a comprehensive comparison of state-of-the-art MRI segmentation models and introduces ABDSynth, a CT-trained model that reduces annotation effort.

Findings

MRSegmentator achieves the highest accuracy and generalizability.

ABDSynth offers a viable alternative with less training data requirements.

All models are evaluated on diverse public datasets to assess robustness.

Abstract

Recent advances in deep learning have led to robust automated tools for segmentation of abdominal computed tomography (CT). Meanwhile, segmentation of magnetic resonance imaging (MRI) is substantially more challenging due to the inherent signal variability and the increased effort required for annotating training datasets. Hence, existing approaches are trained on limited sets of MRI sequences, which might limit their generalizability. To characterize the landscape of MRI abdominal segmentation tools, we present here a comprehensive benchmarking of the three state-of-the-art and open-source models: MRSegmentator, MRISegmentator-Abdomen, and TotalSegmentator MRI. Since these models are trained using labor-intensive manual annotation cycles, we also introduce and evaluate ABDSynth, a SynthSeg-based model purely trained on widely available CT segmentations (no real images). More generally, we assess accuracy and generalizability by leveraging three public datasets (not seen by any of the evaluated methods during their training), which span all major manufacturers, five MRI sequences, as well as a variety of subject conditions, voxel resolutions, and fields-of-view. Our results reveal that MRSegmentator achieves the best performance and is most generalizable. In contrast, ABDSynth yields slightly less accurate results, but its relaxed requirements in training data make it an alternative when the annotation budget is limited. The evaluation code and datasets are given for future benchmarking at https://github.com/deepakri201/AbdoBench, along with inference code and weights for ABDSynth.