---
title: 'MRISegmentator-Abdomen: Automated MRI Segmentation'
url: https://www.emergentmind.com/topics/mrisegmentator-abdomen
type: topic
---

# MRISegmentator-Abdomen: Automated MRI Segmentation

Searching arXiv for MRISegmentator-Abdomen and closely related abdominal MRI segmentation/benchmark papers.
MRISegmentator-Abdomen is a fully automated multi-organ and structure segmentation tool for T1-weighted abdominal MRI that was introduced to delineate 62 abdominal organs and structures from routine MRI acquisitions [2405.05944]. It was developed alongside a voxel-level annotated dataset of 780 T1-weighted abdominal MRI series from 195 patients, with four axial phases per patient—pre-contrast, arterial, venous, and delayed—and was evaluated on an internal held-out test set as well as the external AMOS22 and Duke Liver datasets [2405.05944]. In the originating report, the method is implemented as a 3D nnU-Net model and is referred to as “MRISegmentator” in short [2405.05944]. Subsequent benchmarking positioned it among the state-of-the-art open-source abdominal MRI segmentation tools, while also highlighting sequence-dependent generalization limits outside its training distribution [2507.17971].

## 1. Definition and scope

MRISegmentator-Abdomen addresses automated segmentation of abdominal MRI at a granularity substantially broader than earlier abdominal MRI approaches that focused on a limited number of structures [2405.05944]. The reported target set comprises 62 abdominal organs and structures, and the accompanying dataset contains voxel-level annotations for each of those targets across all four T1-weighted contrast phases collected for each patient [2405.05944].

The training cohort was assembled from 195 patients imaged at the National Institutes of Health Clinical Center between January 2019 and October 2021, with same-day abdominal MRI and CT available for each patient [2405.05944]. For each patient, four axial T1-weighted MRI series were collected—pre-contrast, arterial, portal venous, and delayed—yielding 780 volumes and 69,248 individual slices [2405.05944]. The MRI spatial resolution varied within in-plane pixel spacing of 0.94–1.47 mm and slice thickness of 3.0–3.3 mm, and both 1.5 T and 3 T scanners were used [2405.05944].

The motivation given for the tool is the lack, at the time, of a publicly available abdominal MRI dataset with voxel-level annotations of multiple organs and structures, and correspondingly the absence of a publicly available multi-structure segmentation tool for abdominal MRI [2405.05944]. A plausible implication is that MRISegmentator-Abdomen was designed not only as a model but also as an annotation and benchmarking resource intended to close a dataset bottleneck in abdominal MRI segmentation.

## 2. Dataset construction and annotation workflow

The annotation strategy was explicitly semi-automated and proceeded in two stages [2405.05944]. In the first, “cross-domain” stage, each MRI was translated synthetically into CT contrast using a deep MR-to-CT network; the synthetic CT was then segmented with TotalSegmentator to produce pseudo-annotations for 15 major organs and 7 large vessels [2405.05944]. In parallel, the patient’s true CT series, acquired the same day, were deformably registered to the MRI, segmented by TotalSegmentator to yield masks for 8 muscle groups and 32 osseous structures, and then manually refined by an expert [2405.05944].

In the second, “iterative learning” stage, the initial pseudo-labels were corrected and augmented over four rounds, denoted R1–R4, of “train-refine-retrain” [2405.05944]. At each round, an nnU-Net model was trained on the current annotations, deployed to predict on unlabeled volumes, and its predictions were refined by a radiologist [2405.05944]. The final result was fully corrected voxel-wise maps for all 62 structures in every volume [2405.05944].

The patient-level split used 135 patients and 540 volumes for training and 60 patients and 240 volumes for a held-out internal test set [2405.05944]. All annotations in the test set were reviewed by both a radiology resident and a board-certified radiologist prior to final correction [2405.05944]. This review protocol is notable because it constrains label quality on the evaluation set more tightly than would be expected from a purely single-rater workflow. A plausible implication is that the reported internal performance reflects both algorithmic behavior and a relatively stringent reference standard.

A later benchmark summarized the same training regime somewhat differently but consistently in substance: it reported 780 T1-weighted MRI volumes from a proprietary in-house cohort, 62 annotated anatomical structures, and labels transferred from CT segmentations via a learned MR-to-pseudo-CT translation plus iterative expert refinement using MONAI Label [2507.17971]. That benchmark also noted the presence of pathologies such as liver tumors, pancreatic cysts, and assorted abdominal abnormalities in the training data [2507.17971].

## 3. Model architecture, optimization, and inference

MRISegmentator-Abdomen is implemented using the 3D full-resolution instantiation of nnU-Net [2405.05944]. In the originating description, the network is automatically configured as a U-shaped encoder-decoder with five resolution stages, with each stage consisting of two blocks of convolution, instance normalization, and leaky ReLU activation, and with down- and up-sampling by stride-2 operations [2405.05944]. The training objective is the sum of Dice loss and binary cross-entropy loss,
$$
L = DSC_{\text{loss}} + BCE_{\text{loss}},
$$
optimized with Adam [2405.05944].

The reported training schedule used 2,000 epochs, an initial learning rate of $10^{-2}$ decayed by a polynomial schedule, a batch size of 2 volumes, and five-fold cross-validation [2405.05944]. At inference, the five models were ensembled by soft voting [2405.05944]. All experiments ran on an NVIDIA A100 80 GB GPU in the NIH Biowulf cluster [2405.05944]. Input volumes were z-score normalized per channel and cropped to nonzero bounding boxes under the nnU-Net pipeline, which also supplied on-the-fly 3D augmentations including random affine deformations, elastic deformations, intensity scaling or noise, gamma correction, and mirroring [2405.05944].

A subsequent benchmark characterized the same system as a standard 3D U-Net configured by nnU-Net with four down-sampling and up-sampling levels, two successive $3 \times 3 \times 3$ convolutions per level, instance normalization, LeakyReLU, a final $1 \times 1 \times 1$ convolution to $C=62$ output channels, and softmax, with no additional attention or transformer modules reported [2507.17971]. That benchmark also described nnU-Net-style preprocessing—resampling to the median voxel spacing, clipping intensities at the 0.5 and 99.5 percentiles, z-score normalization, and automatic body cropping—and patch-based optimization with patch size $48 \times 160 \times 192$, batch size 2, stochastic gradient descent with Nesterov momentum 0.99, weight decay $3 \times 10^{-5}$, and a polynomial learning-rate decay over approximately 1,000 epochs [2507.17971]. For inference, it reported sliding-window prediction with one-quarter overlap and post-processing by removal of isolated components smaller than 10,000 voxels for each label [2507.17971].

These two descriptions are not identical in implementation detail. The common denominator is that MRISegmentator-Abdomen is framed as an nnU-Net-based 3D multi-class segmentation system whose novelty lies primarily in the training data and annotation pipeline rather than in a bespoke architectural departure [2405.05944; 2507.17971].

## 4. Segmentation targets and quantitative performance

The primary evaluation metrics in the originating study were the Dice Similarity Coefficient (DSC) and the Normalized Surface Distance (NSD) [2405.05944]. On the internal test set, comprising 240 volumes and all 62 structures, MRISegmentator achieved a mean DSC of $0.861 \pm 0.170$ and a mean NSD of $0.924 \pm 0.163$ [2405.05944].

Performance was further stratified into four structure groups [2405.05944]:

| Group | Structures | Performance |
|---|---:|---|
| G1 | 15 major organs | DSC $= 0.918 \pm 0.069$, NSD $= 0.959 \pm 0.067$ |
| G2 | 7 vessels | DSC $= 0.913 \pm 0.105$, NSD $= 0.972 \pm 0.080$ |
| G3 | 8 muscles | DSC $= 0.929 \pm 0.093$, NSD $= 0.971 \pm 0.084$ |
| G4 | 32 bones | DSC $= 0.806 \pm 0.250$, NSD $= 0.886 \pm 0.245$ |

Phase-stratified internal performance was reported as follows [2405.05944]:

| Phase | DSC | NSD |
|---|---|---|
| PRE | $0.863 \pm 0.231$ | $0.922 \pm 0.208$ |
| ART | $0.874 \pm 0.217$ | $0.932 \pm 0.196$ |
| VEN | $0.880 \pm 0.212$ | $0.937 \pm 0.191$ |
| DEL | $0.876 \pm 0.220$ | $0.933 \pm 0.197$ |

A mixed-effects model confirmed that post-contrast phases significantly outperformed PRE with $p<0.05$, while ART, VEN, and DEL did not differ significantly from one another [2405.05944]. This suggests that contrast enhancement systematically improves separability for the target classes in this dataset, but that the three post-contrast phases provide broadly similar segmentation conditions.

Per-structure rankings followed expected size and contrast trends [2405.05944]. Large, high-contrast organs such as the liver, spleen, kidneys, and back muscles were among the highest-performing classes, with reported DSCs including liver $0.993 \pm 0.002$, spleen $0.977 \pm 0.015$, right kidney $0.984 \pm 0.010$, and left kidney $0.970 \pm 0.093$ [2405.05944]. The smallest or morphologically thin structures were more challenging: adrenal glands were approximately 0.66 Dice, upper ribs fell to approximately 0.48, and the transitional vertebra T7 was 0.682 [2405.05944].

External validation showed mean DSC and NSD of $0.829 \pm 0.133$ and $0.908 \pm 0.067$ on AMOS22, and mean DSC and NSD of $0.933 \pm 0.015$ and $0.929 \pm 0.021$ on Duke Liver [2405.05944]. On the 20-volume AMOS22 validation subset, the model, trained only on its internal data, achieved DSC $= 0.844$ and ranked second among methods trained directly on AMOS22 [2405.05944]. On the Duke Liver set, PRE, ART, and VEN phases yielded liver DSCs of 0.912, 0.946, and 0.942 and NSDs of 0.903, 0.929, and 0.956, respectively [2405.05944].

## 5. Generalization, robustness, and benchmark positioning

MRISegmentator-Abdomen was later benchmarked against MRSegmentator, TotalSegmentator MRI, and ABDSynth on public datasets not seen during training [2507.17971]. In that study, it was described as one of three state-of-the-art abdominal MRI segmentation tools, but its generalization profile was more sequence-sensitive than that of MRSegmentator [2507.17971].

On AMOS, the benchmark reported strong Dice values for major organs, including liver $0.97 \pm 0.02$, spleen $0.96 \pm 0.04$, left kidney $0.95 \pm 0.10$, right kidney $0.94 \pm 0.06$, pancreas $0.85 \pm 0.11$, and stomach $0.88 \pm 0.13$ [2507.17971]. However, it also reported very large HD95 outliers, such as liver $18.86 \pm 40.15$ mm and right kidney $31.10 \pm 58.37$ mm, which it interpreted as evidence of fragmented or anatomically implausible predictions in some volumes [2507.17971].

On the CHAOS dataset, which contained sequences absent from the model’s training set, the benchmark reported marked performance degradation [2507.17971]. For T1 dual in-phase, Dice values were 0.81 for liver, 0.34 for spleen, 0.63 for left kidney, and 0.78 for right kidney; for T1 dual out-phase, 0.83, 0.45, 0.77, and 0.74; and for T2 SPIR, 0.88, 0.87, 0.92, and 0.91, respectively [2507.17971]. The same benchmark concluded that MRISegmentator-Abdomen performed well on AMOS T1 sequences but lacked cross-sequence robustness on unseen T1 dual in/out-phase acquisitions and showed high HD95 values on several settings [2507.17971].

For LiverHCCSeg, the benchmark reported liver Dice $= 0.93 \pm 0.07$ and HD95 $= 10.20 \pm 21.90$ mm against one rater [2507.17971]. It noted that this Dice approached inter-rater agreement, although HD95 remained elevated compared with MRSegmentator [2507.17971].

The benchmark’s comparative interpretation was that MRSegmentator achieved the best performance and was most generalizable overall [2507.17971]. In contrast, MRISegmentator-Abdomen was portrayed as highly accurate on T1 acquisitions similar to its training data but less robust to commonly used MRI protocols outside that distribution [2507.17971]. This suggests that the model’s dense annotation scope and strong in-domain performance do not automatically translate into protocol-agnostic behavior.

## 6. Failure modes, deployment characteristics, and applications

The originating study identified several failure patterns concentrated in small, thin, and low-contrast targets, as well as in orientations or sequences not represented in training [2405.05944]. Specifically, adrenal glands, duodenum, thin ribs, boundary regions with abrupt tissue transitions, and coronal T1-weighted volumes including chest structures were reported as challenging [2405.05944]. When the model encountered coronal or sagittal volumes incorporating portions of the lower chest, lung parenchyma and upper ribs were occasionally mis-segmented as bony structures [2405.05944]. The study also stated that the tool was robust to a variety of pathologies, including liver metastases and renal lesions, and contrasted this with a comparator nnUNet-AMOS model trained only on “healthy” AMOS22 scans, which often failed to adhere to lesion boundaries in tumorous cases [2405.05944].

The benchmark paper emphasized somewhat different failure characteristics: sequence sensitivity, spatial outliers reflected by large HD95, and the operational cost of the annotation pipeline [2507.17971]. It identified spurious island-like predictions, missed anatomy, and sensitivity to training-sequence scope as salient limitations, and it characterized the cross-modality pseudo-CT plus iterative manual refinement workflow on nearly 800 MRI scans as time- and expert-intensive [2507.17971].

For deployment, the model is used through the nnU-Net inference pipeline, accepts NIfTI inputs of any one of the four T1-weighted phases, performs its own normalization and cropping, and outputs a multi-label NIfTI segmentation [2405.05944]. Dependencies include Python $\geq 3.7$, PyTorch, nnU-Net, and standard Python libraries, and a Docker container is provided [2405.05944]. Reported runtime for a single 80–96 slice volume is approximately 1–2 minutes on a single GPU such as an A100, or roughly 5–10 minutes on a high-end CPU [2405.05944].

The intended application areas named in the study include radiotherapy planning, opportunistic body composition analysis, longitudinal disease monitoring, and radiomics and AI research [2405.05944]. Because the segmenter covers 62 organs and structures rather than a narrow organ subset, a plausible implication is that it is particularly suited to retrospective cohort analysis and multi-task downstream pipelines in which anatomical masks are reused for several quantitative tasks.

## 7. Relation to adjacent abdominal MRI segmentation methods

MRISegmentator-Abdomen emerged within a rapidly developing ecosystem of abdominal MRI segmentation tools. Earlier MRI-focused systems generally addressed fewer targets. ALAMO, for example, was a fully automated abdominal MR segmentation framework on 3 T T1-VIBE images for 10 organs-at-risk, using a 2D U-Net with dense connections, deep supervision, and multi-view aggregation, and reported overall mean DSC $= 0.909 \pm 0.048$ on its 20-subject test set [1912.11000]. ProRSeg targeted gastrointestinal organs at risk on MRI and CBCT, using joint registration-segmentation and reporting MRI DSC values of $0.94 \pm 0.02$ for liver, $0.88 \pm 0.04$ for large bowel, $0.78 \pm 0.03$ for small bowel, and $0.82 \pm 0.04$ for stomach-duodenum [2210.14297]. These methods addressed narrower label sets than MRISegmentator-Abdomen.

Among contemporary open-source tools, MRSegmentator and TotalVibeSegmentator provide relevant comparators. MRSegmentator segments 40 classes in MRI and CT, is also built on nnU-Net, and was trained on 1,200 UK Biobank MRI scans, 221 in-house MRI scans, and 1,228 CT scans with a human-in-the-loop annotation workflow [2405.06463]. In a benchmark on multi-parametric T1-weighted abdominal MRI, MRSegmentator achieved a pooled Dice score of $80.7 \pm 18.6$ and Hausdorff Distance error of $8.9 \pm 10.4$ mm on 40 curated Duke Liver volumes, outperforming TotalSegmentator MRI and TotalVibeSegmentator across sequence types with $p < .05$ [2504.07729]. TotalVibeSegmentator, by contrast, focuses on VIBE images, produces 71 semantic torso labels, and reported an average Dice score of $0.89 \pm 0.07$ overall 71 labels, with Dice-score above 0.90 on abdominal organs except for the pancreas at 0.70 [2406.00125].

Within this comparative landscape, MRISegmentator-Abdomen is distinguished primarily by its 62-class T1-weighted abdominal MRI scope and its paired dataset-and-tool release model [2405.05944]. The later benchmark’s conclusion that MRSegmentator was more generalizable does not negate MRISegmentator-Abdomen’s importance; rather, it indicates a trade-off between dense in-domain annotation and broader sequence coverage in training [2507.17971]. This suggests that MRISegmentator-Abdomen is especially strong as a resource and tool for T1-weighted abdominal MRI resembling its NIH acquisition regime, while broader protocol robustness remains an open area for further work.

Source: https://www.emergentmind.com/topics/mrisegmentator-abdomen