Overview

Schema Version

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

Name

Open-access Cardiac Implantable Electronic Devices (CIED) chest radiograph dataset

Link

https://github.com/AndreasZhukov/cied

Indexing

Keywords: cardiac implantable electronic device, CIED, pacemaker, implantable cardioverter defibrillator, ICD, chest radiograph, DICOM, smartphone, segmentation, classification, U-Net, ResNet-50

Content: CH, IN

RadLex: RID35727, RID10784, RID49475, RID5434, RID5433, RID11231

Author(s)

Felix Busch

Keno K. Bressem

Phillip Suwalski

Lena Hoffmann

Stefan M. Niehues

Denis Poddubnyy

Marcus R. Makowski

Hugo J. W. L. Aerts

Andrei Zhukov

Lisa C. Adams

Organization(s)

Charité–Universitätsmedizin Berlin

German Heart Center, Technical University of Munich

Department of Diagnostic and Interventional Radiology, Technical University of Munich

Mass General Brigham, Harvard Medical School (AIM Program)

Dana-Farber Cancer Institute and Brigham and Women’s Hospital

Maastricht University (CARIM & GROW)

Funding

Authors declared no funding for this work.

Ethical review

Single-center retrospective study; IRB approval Charité–Berlin University Medicine (EA4/042/20); informed consent waived due to retrospective design.

Comments

Public dataset and code for CIED segmentation and classification on chest radiographs (DICOM and smartphone photographs); authors state the entire collection will be made accessible on PhysioNet.

Date

Published: 2024

References

[1] Busch F, Bressem KK, Suwalski P, Hoffmann L, Niehues SM, Poddubnyy D, Makowski MR, Aerts HJWL, Zhukov A, Adams LC. "Open Access Data and Deep Learning for Cardiac Device Identification on Standard DICOM and Smartphone-based Chest Radiographs". Radiology: Artificial Intelligence. 2024 Sep;6(5):e230502. 2024-07-17. doi:10.1148/ryai.230502. PMID: 39017033. PMCID: PMC11427927.

[2] . "Heart-devices. GitHub repository for CIED project". GitHub. 2024. Available from: https://github.com/AndreasZhukov/cied

Dataset

Motivation

Provide a publicly available dataset and code for robust CIED segmentation and exact model/manufacturer classification on standard and smartphone-acquired chest radiographs to support clinical use and research reproducibility.

Sampling

Retrospective single-center cohort of adults (≥18 years) with CIEDs who underwent AP bedside or standing PA chest radiography between Jan 2012 and Jan 2022; exclusions: poor quality, incomplete device capture, multiple devices in one image; only distinct time points per patient included.

Partitioning scheme

Segmentation: 70% training, 30% validation. Classification: intended 70%/20%/10% train/validation/test; final split 70.03%/18.52%/11.45% due to class imbalance and variable image counts per patient; mixes DICOM and smartphone images.

Missing information

Exact license and final PhysioNet accession/link not specified in the paper; detailed counts per split not fully enumerated.

Relationships between instances

Multiple images per patient; smartphone images are photographs of the corresponding chest radiographs from monitors; to avoid overrepresentation, patients were divided into three groups and photographed by different authors so each patient appears in only one group.

Noise

Smartphone photographs captured from clinical monitors at varying angles and with five different smartphone models to introduce real-world variability.

External data

External testing performed using datasets/applications from Howard et al. and Weinreich et al.; not included in this dataset.

Confidentiality

All DICOM images were anonymized by deleting header information and removing burned-in text via OCR; images were converted to PNG for distribution.

Re-identification

DICOM headers removed and burned-in text removed with OCR prior to conversion to PNG; residual re-identification risk minimized.

Sensitive data

Clinical images with PHI removed; no direct patient identifiers retained.