Overview

Schema Version

https://atlas.rsna.org/schemas/2025-11/dataset.json

Name

Liver Segmental Volume Ratio and Spleen Segmentation Datasets 1 and 2

Link

https://dx.doi.org/10.1148/ryai.210268

Indexing

Keywords: LSVR, Couinaud segments, Spleen volume, Cirrhosis, Advanced fibrosis, Contrast-enhanced CT, HCV, Biopsy reference

Content: GI, CT, RS

RadLex: RID11580, RID65, RID39534

SNOMED: 50711007, 16294009

Author(s)

Sungwon Lee

Daniel C. Elton

Alexander H. Yang

Christopher Koh

David E. Kleiner

Meghan G. Lubner

Perry J. Pickhardt

Ronald M. Summers

Organization(s)

Imaging Biomarkers and Computer-Aided Diagnosis Laboratory, Department of Radiology and Imaging Sciences, NIH Clinical Center

Liver Diseases Branch, Intramural Research Program, National Institute of Diabetes and Digestive and Kidney Diseases

Laboratory of Pathology, National Cancer Institute, National Institutes of Health

Department of Radiology, University of Wisconsin School of Medicine & Public Health

License

URL: https://pubs.rsna.org/journals/ryai

Contact

Ronald M. Summers, vog.hin@smr

Funding

Supported in part by the Intramural Research Program of the National Institutes of Health, Clinical Center, National Institute of Diabetes and Digestive and Kidney Diseases, and National Cancer Institute.

Ethical review

HIPAA-compliant, retrospective, multi-institutional study with IRB approval at each institution; need for additional signed informed consent was waived.

Comments

Retrospective, multi-institutional study evaluating automated CT-based liver Couinaud segment and spleen segmentation to compute LSVR and spleen volume for predicting fibrosis/cirrhosis. Two biopsy-referenced cohorts were used for evaluation: Dataset 1 (HCV cohort, METAVIR) and Dataset 2 (mixed etiologies, Ishak/HAI).

Date

Published: 2022-08-24

References

[1] Lee S, Elton DC, Yang AH, Koh C, Kleiner DE, Lubner MG, Pickhardt PJ, Summers RM. "Fully Automated and Explainable Liver Segmental Volume Ratio and Spleen Segmentation at CT for Diagnosing Cirrhosis". Radiology: Artificial Intelligence. 2022-09-01. doi:10.1148/ryai.210268. PMID: 36204530. PMCID: PMC9530761.

Dataset

Motivation

Automate explainable liver and spleen volumetry (including LSVR) on contrast-enhanced CT to aid noninvasive diagnosis of cirrhosis/advanced fibrosis.

Sampling

Retrospective inclusion of consecutive/eligible patients meeting criteria at two institutions over long time spans (2000–2016 and 2001–2021).

Partitioning scheme

Multivariable models built on dataset 1 and tested on dataset 1 (hold-out) and dataset 2; segmentation DL model training details in supplement; 70-case subset (35 per dataset) used for manual vs automated comparison.

Missing information

Image file formats, exact scanner models/protocol parameters per site, and exact partition sizes for training/hold-out are only in supplement; series/image counts not reported.

Relationships between instances

Each patient has contrast-enhanced portal venous phase abdominal CT and a liver biopsy within 1 year; automated outputs include per-segment volumes and attenuation; derived LSVR and spleen volume relate to biopsy stage.

Noise

Multi-scanner, multi-year acquisition variability (16–20 years) with varying protocols and contrast parameters.

External data

Authors note the DL model training used public data referenced in Tian et al. 2019; evaluation datasets are institutional clinical cohorts.

Confidentiality

Retrospective, HIPAA-compliant clinical data with IRB approval; specific de-identification procedures not detailed.

Sensitive data

Includes clinical biopsy results and potentially PHI in source CT; publication does not specify release of de-identified images.