Overview

Schema Version

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

Name

Cancer-associated Incidental Pulmonary Embolism dataset

Link

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10698599

Indexing

Keywords: Incidental pulmonary embolism, Pulmonary embolism, Artificial intelligence, Cancer, CT imaging

Content: CH, CT, OI

RadLex: RID4247, RID12362, RID13060, RID4834

SNOMED: 59282003, 363346000

Author(s)

Peder Wiklund

Koshiar Medson

Organization(s)

Region Halland, Halland Hospital Halmstad

Karolinska University Hospital

Karolinska Institutet

License

URL: https://pubs.rsna.org/journal/rai

Funding

Authors declared no funding for this work.

Ethical review

Approved by the Swedish Ethical Review Authority; informed consent was waived.

Date

Published: 2023-10-18

References

[1] Wiklund P, Medson K. "Use of a Deep Learning Algorithm for Detection and Triage of Cancer-associated Incidental Pulmonary Embolism". Radiology: Artificial Intelligence. 2023-11-01. doi:10.1148/ryai.220286. PMID: 38074784. PMCID: PMC10698599.

[2] Wiklund P, Medson K, Elf J. "Incidental pulmonary embolism in patients with cancer: prevalence, underdiagnosis and evaluation of an AI algorithm for automatic detection of pulmonary embolism". European Radiology. 2023-01-01. PMID: 36002759. PMCID: PMC9889421. Available from: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9889421

Dataset

Motivation

Evaluate the effect of implementing an AI algorithm for iPE detection/triage on detection rate, report turnaround time, and time to treatment in cancer patients.

Sampling

Adult patients (≥18 years) with confirmed cancer and at least one contrast-enhanced chest CT for cancer evaluation within the defined time windows.

Partitioning scheme

Two time-based cohorts: before AI implementation (2018-07-01 to 2019-06-30) and after AI implementation (2020-11-01 to 2021-04-30).

Missing information

Patient counts per period and image/series counts not reported; file formats and image resolution not specified.

Relationships between instances

Multiple examinations included; iPE analyzed at study level; AI outputs reviewed by radiologists.

External data

Commercial AI algorithm (Aidoc BriefCase; Aidoc Medical) used for detection/triage; training/validation details reported elsewhere.

Confidentiality

Retrospective clinical imaging and EHR timestamps from a single center; not publicly shared.

Re-identification

Not discussed; clinical data handled within hospital systems.

Sensitive data

Contains clinical imaging and treatment timing information.