Assessing quantitative performance and expert review of multiple deep learning-based frameworks for computed tomography-based abdominal organ auto-segmentation

Udbhav S. Ram; Joel A. Pogue; Michael Soike; Neil T. Pfister; Rojymon Jacob; Carlos E. Cardenas

doi:10.1016/j.intonc.2025.03.003

您当前的位置：

首页 >

文章列表页 >

Assessing quantitative performance and expert review of multiple deep learning-based frameworks for computed tomography-based abdominal organ auto-segmentation

Research article | 更新时间：2026-06-03

- Assessing quantitative performance and expert review of multiple deep learning-based frameworks for computed tomography-based abdominal organ auto-segmentation
- Assessing quantitative performance and expert review of multiple deep learning-based frameworks for computed tomography-based abdominal organ auto-segmentation
- 智能肿瘤学 2025年1卷第2期页码：160-171
- 作者机构：
  
  Department of Radiation Oncology, The University of Alabama at Birmingham,AL,Birmingham,United States,35233
- 作者简介：
- 基金信息：
- DOI：10.1016/j.intonc.2025.03.003
  中图分类号：
- 收稿：2025-01-29，
  
  网络首发：2025-03-19，
  
  纸质出版：2025-04-25
- 稿件说明：
移动端阅览
Udbhav S. Ram, Joel A. Pogue, Michael Soike, 等. Assessing quantitative performance and expert review of multiple deep learning-based frameworks for computed tomography-based abdominal organ auto-segmentation[J]. 智能肿瘤学, 2025,1(2):160-171.

Udbhav S. Ram, Joel A. Pogue, Michael Soike, et al. Assessing quantitative performance and expert review of multiple deep learning-based frameworks for computed tomography-based abdominal organ auto-segmentation[J]. Intelligent Oncology, 2025, 1(2): 160-171.
Udbhav S. Ram, Joel A. Pogue, Michael Soike, 等. Assessing quantitative performance and expert review of multiple deep learning-based frameworks for computed tomography-based abdominal organ auto-segmentation[J]. 智能肿瘤学, 2025,1(2):160-171. DOI： 10.1016/j.intonc.2025.03.003.

Udbhav S. Ram, Joel A. Pogue, Michael Soike, et al. Assessing quantitative performance and expert review of multiple deep learning-based frameworks for computed tomography-based abdominal organ auto-segmentation[J]. Intelligent Oncology, 2025, 1(2): 160-171. DOI： 10.1016/j.intonc.2025.03.003.

摘要

BackgroundSegmentation of abdominal organs in computed tomography (CT) images within clinical oncological workflows is crucial for ensuring effective treatment planning and follow-up. However

manually generated segmentations are time-consuming and labor-intensive in addition to being subject to inter-observer variability. Many deep learning and automated machine learning (AutoML) frameworks have emerged as a solution to this challenge and show promise in clinical workflows.ObjectiveThis study aims to provide a comprehensive evaluation of existing AutoML frameworks (Auto3DSeg

nnU-Net) against a state-of-the-art non-AutoML framework

the Shifted Window U-Net Transformer (SwinUNETR).MethodsEach framework was trained on the same 122 training images

taken from the Abdominal Multi-Organ Segmentation (AMOS) Grand Challenge. Frameworks were compared using dice similarity coefficient (DSC)

surface DSC (sDSC)

and 95th percentile Hausdorff distances (HD95) on

an additional 72 holdout-validation images. The perceived clinical viability of 30 auto-contoured test cases was assessed by three physicians in a blinded evaluation.ResultsComparisons show significantly better performance by AutoML methods: nnU-Net (average DSC: 0.924

average sDSC: 0.938

average HD95: 4.26

median Likert: 4.57)

Auto3DSeg (average DSC: 0.902

average sDSC: 0.919

average HD95: 8.76

median Likert: 4.49)

and SwinUNETR (average DSC: 0.837

average sDSC: 0.844

average HD95: 13.93). AutoML frameworks were quantitatively preferred (13/13 organs at risks [OARs

]

  0.05 in DSC and sDSC

12/13 OARs

  0.05 in HD95

comparing Auto3DSeg to SwinUNETR

and all OARs

  0.05 in all metrics comparing SwinUNETR to nnU-Net). Qualitatively

nnU-Net was preferred over Auto3DSeg (

 = 0.0027).ConclusionThe findings suggest that AutoML frameworks offer a significant advantage in the segmentation of abdominal organs

and underscores the potential of AutoML methods to enhance the efficiency of oncological workflows.

Abstract

BackgroundSegmentation of abdominal organs in computed tomography (CT) images within clinical oncological workflows is crucial for ensuring effective treatment planning and follow-up. However

nnU-Net) against a state-of-the-art non-AutoML framework

the Shifted Window U-Net Transformer (SwinUNETR).MethodsEach framework was trained on the same 122 training images

taken from the Abdominal Multi-Organ Segmentation (AMOS) Grand Challenge. Frameworks were compared using dice similarity coefficient (DSC)

surf

ace DSC (sDSC)

and 95th percentile Hausdorff distances (HD95) on an additional 72 holdout-validation images. The perceived clinical viability of 30 auto-contoured test cases was assessed by three physicians in a blinded evaluation.ResultsComparisons show significantly better performance by AutoML methods: nnU-Net (average DSC: 0.924

average sDSC: 0.938

average HD95: 4.26

median Likert: 4.57)

Auto3DSeg (average DSC: 0.902

average sDSC: 0.919

average HD95: 8.76

median Likert: 4.49)

and SwinUNETR (average DSC: 0.837

average sDSC: 0.844

average HD95: 13.93). AutoML frameworks were quantitatively preferred (13/13 organs at risks [OARs

]

  0.05 in DSC and sDSC

12/13 OARs

  0.05 in HD95

comparing Auto3DSeg to SwinUNETR

and all OARs

  0.05 in all metrics comparing SwinUNETR to nnU-Net). Qualitatively

nnU-Net was preferred over Auto3DSeg (

 = 0.0027).ConclusionThe findings suggest that AutoML frameworks offer a significant advantage in the segmentation of abdominal organs

and underscores the potential of AutoML methods to enhance the efficiency of oncological workflows.

关键词

Keywords

references

浏览量

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

Deep learning-based segmentation of small-volume brain metastases in lung cancer patients

A multiscale residual dense fusion network for nuclear medical image fusion

Integrative machine learning-driven prioritization of ceRNA networks in adrenocortical carcinoma

From morphology to function: Advances in multimodal imaging for pulmonary function prediction

TSMIL: Transformer-based structured low-rank end-to-end multi-instance learning network for renal cell carcinoma classification in whole-slide images