Validation of automated computed tomography segmentation software to assess body composition among cancer patients

IF 2.6 Q3 NUTRITION & DIETETICS

Clinical nutrition ESPEN Pub Date : 2025-08-25 DOI:10.1016/j.clnesp.2025.08.024

Mushfiqus Salehin , Vincent Tze Yang Chow , Hyunwoo Lee , Erin K. Weltzien , Long Nguyen , Jia Ming Li , Varun Akella , Bette J. Caan , Elizabeth M. Cespedes Feliciano , Da Ma , Mirza Faisal Beg , Karteek Popuri

{"title":"Validation of automated computed tomography segmentation software to assess body composition among cancer patients","authors":"Mushfiqus Salehin , Vincent Tze Yang Chow , Hyunwoo Lee , Erin K. Weltzien , Long Nguyen , Jia Ming Li , Varun Akella , Bette J. Caan , Elizabeth M. Cespedes Feliciano , Da Ma , Mirza Faisal Beg , Karteek Popuri","doi":"10.1016/j.clnesp.2025.08.024","DOIUrl":null,"url":null,"abstract":"<div><h3>Background & aims</h3><div>Assessing body composition using computed tomography (CT) can help predict the clinical outcomes of cancer patients, including surgical complications, chemotherapy toxicity, and survival. However, manual segmentation of CT images is labor-intensive and can lead to significant inter-observer variability. In this study, we validate the accuracy and reliability of automatic CT-based segmentation using the Data Analysis Facilitation Suite (DAFS) Express software package, which rapidly segments single CT slices.</div></div><div><h3>Methods</h3><div>The study analyzed single-slice images at the third lumbar vertebra (L3) level (n = 5973) of patients diagnosed with non-metastatic colorectal (n = 3098) and breast cancer (n = 2875) at Kaiser Permanente Northern California. Manual segmentation used SliceOmatic with Alberta protocol HU ranges; automated segmentation used DAFS Express with identical HU limits. The accuracy of the automated segmentation was evaluated using the DICE index, the reliability was assessed by intra-class correlation coefficients (ICC) with 95 % CI, and the agreement between automatic and manual segmentations was assessed by Bland–Altman analysis. DICE scores below 20 % and 70 % were considered failed and poor segmentations, respectively, and underwent additional review. The mortality risk associated with each tissue's area was generated using Cox proportional hazard ratios (HR) with 95 % CI, adjusted for patient-specific variables including age, sex, race/ethnicity, cancer stage and grade, treatment receipt, and smoking status. A blinded review process categorized images with various characteristics for sensitivity analysis.</div></div><div><h3>Results</h3><div>The mean (standard deviation, SD) ages of the colorectal and breast cancer patients were 62.6 (11.4) and 56 (11.8), respectively. Automatic segmentation showed high accuracy vs. manual segmentation, with mean DICE scores above 96 % for skeletal muscle (SKM), visceral adipose tissue (VAT), and subcutaneous adipose tissue (SAT), and above 77 % for intermuscular adipose tissue (IMAT), with three failures, representing 0.05 % of the cohort. Bland–Altman analysis of 5973 measurements showed mean cross-sectional area differences of −5.73, −0.84, −2.82, and −1.02 cm<sup>2</sup> for SKM, VAT, SAT and IMAT, respectively, indicating good agreement, with slight underestimation in SKM and SAT. Reliability Coefficients ranged from 0.88 to 1.00 for colorectal and 0.95–1.00 for breast cancer, with Simple Kappa values of 0.65–0.99 and 0.67–0.97, respectively. Additionally, mortality associations for automated and manual segmentations were similar, with comparable hazard ratios, confidence intervals, and p-values. Kaplan–Meier survival estimates showed mortality differences below 2.14 %.</div></div><div><h3>Conclusion</h3><div>DAFS Express enables rapid, accurate body composition analysis by automating segmentation, reducing expert time and computational burden. This rapid analysis of body composition is a prerequisite to large-scale research that could potentially enable use in the clinical setting. Automated CT segmentations may be utilized to assess markers of sarcopenia, muscle loss, and adiposity and predict clinical outcomes.</div></div>","PeriodicalId":10352,"journal":{"name":"Clinical nutrition ESPEN","volume":"69 ","pages":"Pages 686-695"},"PeriodicalIF":2.6000,"publicationDate":"2025-08-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Clinical nutrition ESPEN","FirstCategoryId":"1085","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S2405457725029109","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q3","JCRName":"NUTRITION & DIETETICS","Score":null,"Total":0}

引用次数: 0

Abstract

Background & aims

Assessing body composition using computed tomography (CT) can help predict the clinical outcomes of cancer patients, including surgical complications, chemotherapy toxicity, and survival. However, manual segmentation of CT images is labor-intensive and can lead to significant inter-observer variability. In this study, we validate the accuracy and reliability of automatic CT-based segmentation using the Data Analysis Facilitation Suite (DAFS) Express software package, which rapidly segments single CT slices.

Methods

The study analyzed single-slice images at the third lumbar vertebra (L3) level (n = 5973) of patients diagnosed with non-metastatic colorectal (n = 3098) and breast cancer (n = 2875) at Kaiser Permanente Northern California. Manual segmentation used SliceOmatic with Alberta protocol HU ranges; automated segmentation used DAFS Express with identical HU limits. The accuracy of the automated segmentation was evaluated using the DICE index, the reliability was assessed by intra-class correlation coefficients (ICC) with 95 % CI, and the agreement between automatic and manual segmentations was assessed by Bland–Altman analysis. DICE scores below 20 % and 70 % were considered failed and poor segmentations, respectively, and underwent additional review. The mortality risk associated with each tissue's area was generated using Cox proportional hazard ratios (HR) with 95 % CI, adjusted for patient-specific variables including age, sex, race/ethnicity, cancer stage and grade, treatment receipt, and smoking status. A blinded review process categorized images with various characteristics for sensitivity analysis.

Results

The mean (standard deviation, SD) ages of the colorectal and breast cancer patients were 62.6 (11.4) and 56 (11.8), respectively. Automatic segmentation showed high accuracy vs. manual segmentation, with mean DICE scores above 96 % for skeletal muscle (SKM), visceral adipose tissue (VAT), and subcutaneous adipose tissue (SAT), and above 77 % for intermuscular adipose tissue (IMAT), with three failures, representing 0.05 % of the cohort. Bland–Altman analysis of 5973 measurements showed mean cross-sectional area differences of −5.73, −0.84, −2.82, and −1.02 cm² for SKM, VAT, SAT and IMAT, respectively, indicating good agreement, with slight underestimation in SKM and SAT. Reliability Coefficients ranged from 0.88 to 1.00 for colorectal and 0.95–1.00 for breast cancer, with Simple Kappa values of 0.65–0.99 and 0.67–0.97, respectively. Additionally, mortality associations for automated and manual segmentations were similar, with comparable hazard ratios, confidence intervals, and p-values. Kaplan–Meier survival estimates showed mortality differences below 2.14 %.

Conclusion

DAFS Express enables rapid, accurate body composition analysis by automating segmentation, reducing expert time and computational burden. This rapid analysis of body composition is a prerequisite to large-scale research that could potentially enable use in the clinical setting. Automated CT segmentations may be utilized to assess markers of sarcopenia, muscle loss, and adiposity and predict clinical outcomes.

查看原文本刊更多论文

评估癌症患者身体成分的自动计算机断层分割软件的验证

背景和目的利用计算机断层扫描（CT）评估身体成分可以帮助预测癌症患者的临床结果，包括手术并发症、化疗毒性和生存率。然而，CT图像的人工分割是劳动密集型的，并且会导致观察者之间的显著差异。在这项研究中，我们使用数据分析促进套件（DAFS） Express软件包验证了基于CT的自动分割的准确性和可靠性，该软件包可以快速分割单个CT切片。方法：研究分析了北加州Kaiser Permanente诊断为非转移性结直肠癌（n = 3098）和乳腺癌（n = 2875）的患者在第三腰椎（L3）水平（n = 5973）的单片图像。手动分割采用了带有Alberta协议HU范围的SliceOmatic；自动分割使用具有相同HU限制的DAFS Express。采用DICE指数评价自动分割的准确性，采用95% CI的类内相关系数（ICC）评价可靠性，采用Bland-Altman分析评价自动分割与人工分割的一致性。DICE分数低于20%和70%分别被认为是失败和分割不良，并进行额外的审查。使用95% CI的Cox比例风险比（HR）生成与每个组织面积相关的死亡风险，并根据患者特定变量（包括年龄、性别、种族/民族、癌症分期和分级、接受治疗和吸烟状况）进行调整。盲法评价过程对具有不同特征的图像进行分类以进行敏感性分析。结果结直肠癌和乳腺癌患者的平均（标准差，SD）年龄分别为62.6岁（11.4岁）和56岁（11.8岁）。与人工分割相比，自动分割显示出更高的准确性，骨骼肌（SKM）、内脏脂肪组织（VAT）和皮下脂肪组织（SAT）的平均DICE评分在96%以上，肌间脂肪组织（IMAT）的平均DICE评分在77%以上，有3例失败，占队列的0.05%。Bland-Altman分析5973个测量结果显示，SKM、VAT、SAT和IMAT的平均横截面面积差异分别为- 5.73、- 0.84、- 2.82和- 1.02 cm2，表明一致性良好，SKM和SAT略有低估。结直肠癌的信度系数范围为0.88至1.00，乳腺癌的信度系数范围为0.95至1.00，简单Kappa值分别为0.65至0.99和0.67至0.97。此外，自动和人工分割的死亡率关联相似，具有可比的风险比、置信区间和p值。Kaplan-Meier生存估计显示死亡率差异低于2.14%。结论dafs Express可实现快速、准确的人体成分分析，减少专家时间和计算量。这种对人体成分的快速分析是大规模研究的先决条件，有可能在临床环境中使用。自动CT分割可用于评估肌肉减少、肌肉损失和肥胖的标志物，并预测临床结果。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文求助全文

来源期刊

Clinical nutrition ESPEN NUTRITION & DIETETICS-

CiteScore

4.90

自引率

3.30%

发文量

512

期刊介绍： Clinical Nutrition ESPEN is an electronic-only journal and is an official publication of the European Society for Clinical Nutrition and Metabolism (ESPEN). Nutrition and nutritional care have gained wide clinical and scientific interest during the past decades. The increasing knowledge of metabolic disturbances and nutritional assessment in chronic and acute diseases has stimulated rapid advances in design, development and clinical application of nutritional support. The aims of ESPEN are to encourage the rapid diffusion of knowledge and its application in the field of clinical nutrition and metabolism. Published bimonthly, Clinical Nutrition ESPEN focuses on publishing articles on the relationship between nutrition and disease in the setting of basic science and clinical practice. Clinical Nutrition ESPEN is available to all members of ESPEN and to all subscribers of Clinical Nutrition.