Overview

Schema Version

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

Name

CXR-Age external testing cohort (Seoul National University Hospital Gangnam Center, 2004–2018)

Link

https://dx.doi.org/10.1148/ryai.230433

Indexing

Keywords: chest radiograph, biologic age, CXR-Age, mortality, cardiovascular, lung cancer, respiratory disease, asymptomatic, health checkup, external validation, Asian cohort

Content: CH

RadLex: RID5316, RID45686, RID28818, RID10345, RID3284

SNOMED: 363358000, 50043002, 49601007

Author(s)

Jong Hyuk Lee

Dongheon Lee

Michael T. Lu

Vineet K. Raghu

Jin Mo Goo

Yunhee Choi

Seung Ho Choi

Hyungjin Kim

Organization(s)

Seoul National University Hospital, Department of Radiology

Chungnam National University College of Medicine, Department of Biomedical Engineering

Massachusetts General Hospital Cardiovascular Imaging Research Center

Harvard Medical School

Institute of Radiation Medicine, Seoul National University Medical Research Center

Cancer Research Institute, Seoul National University

Medical Research Collaborating Center, Seoul National University Hospital

Healthcare Research Institute, Healthcare System Gangnam Center, Seoul National University Hospital

Funding

National Research Foundation of Korea (NRF) grant funded by the Korean government (MSIT), no. RS-2023-00207978.

Ethical review

Institutional Review Board approval obtained; consent waived. SNUH IRB approval no. H-2206-114-1335.

Comments

Single-center retrospective cohort of asymptomatic Asian individuals undergoing health checkups with chest radiography; used solely for external testing of a previously developed deep learning model (CXR-Age).

Date

Published: 2024-07-24

References

[1] Lee JH; Lee D; Lu MT; Raghu VK; Goo JM; Choi Y; Choi SH; Kim H. "External Testing of a Deep Learning Model to Estimate Biologic Age Using Chest Radiographs". Radiology: Artificial Intelligence. 2024-07-24. doi:10.1148/ryai.230433. PMID: 39046324. PMCID: PMC11427929.

Dataset

Motivation

To externally test generalizability and prognostic value of a previously developed deep learning model (CXR-Age) in a geographically, temporally, and racially distinct cohort.

Sampling

Consecutive asymptomatic individuals aged 50–80 years undergoing comprehensive health checkups between January 2004 and June 2018 at a single medical checkup center.

Partitioning scheme

Single external test cohort; analyses also performed on a subset with available comorbidity information.

Missing information

Comorbidity information available for a subset of individuals (24,804 of 36,924). Imaging counts per exam/series/images not reported.

Relationships between instances

If multiple chest radiographs per individual existed, the radiograph from the first visit with recorded smoking status/underlying disease information was selected (one exam per individual).

External data

Mortality status, date, and cause of death obtained from Statistics Korea database.