Overview

Schema Version

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

Name

ASIST-TBI Traumatic Brain Injury Head CT datasets (Unity Health Toronto)

Link

https://pubs.rsna.org/doi/10.1148/ryai.230088

Indexing

Keywords: CT, Brain/Brain Stem, Surgery, Trauma, Prognosis, Classification, Traumatic Brain Injury, Triage, Machine Learning, Decision Support

Content: ER, NR, CT

RadLex: RID35976, RID11069, RID45946, RID4630, RID12362, RID15717, RID4710, RID10568, RID45700, RID6476

SNOMED: 127295002, 703861005, 1386000

Author(s)

Christopher W. Smith

Armaan K. Malhotra

Christopher Hammill

Derek Beaton

Erin M. Harrington

Yingshi He

Husain Shakil

Amanda McFarlan

Blair Jones

Hui Ming Lin

François Mathieu

Avery B. Nathens

Alun D. Ackery

Garrick Mok

Muhammad Mamdani

Shobhit Mathur

Jefferson R. Wilson

Robert Moreland

Errol Colak

Christopher D. Witiw

Organization(s)

St Michael's Hospital, Unity Health Toronto

Li Ka Shing Knowledge Institute, Unity Health Toronto

Data Science and Advanced Analytics, Unity Health Toronto

University of Toronto (Temerty Faculty of Medicine; Institute for Health Policy, Management and Evaluation; Interdepartmental Division of Critical Care; Leslie Dan Faculty of Pharmacy)

Sunnybrook Health Sciences Centre (Division of Trauma Surgery)

Contact

Corresponding author: Christopher D. Witiw (email shown in article: ot.htlaehytinu@witiW.rehpotsirhC)

Funding

Supported by St Michael's Hospital Medical Services Association Innovation Fund; E.C. supported by the Odette Professorship in Artificial Intelligence for Medical Imaging, St Michael's Hospital, Unity Health Toronto.

Ethical review

Institutional research ethics board approval obtained; informed consent waived due to impracticality, minimal risk, and safeguards.

Date

Published: 2024-01-10

Created: 2005-01-01

References

[1] Smith CW, Malhotra AK, Hammill C, Beaton D, Harrington EM, He Y, Shakil H, McFarlan A, Jones B, Lin HM, Mathieu F, Nathens AB, Ackery AD, Mok G, Mamdani M, Mathur S, Wilson JR, Moreland R, Colak E, Witiw CD. "Vision Transformer–based Decision Support for Neurosurgical Intervention in Acute Traumatic Brain Injury: Automated Surgical Intervention Support Tool". Radiology: Artificial Intelligence. 2024-03-01. doi:10.1148/ryai.230088. PMID: 38197796. PMCID: PMC10982820.

Dataset

Motivation

Develop and evaluate a deep learning model (ASIST-TBI) to predict need for neurosurgical intervention from acute TBI head CT scans.

Sampling

Single specialized tertiary trauma center; patients identified from Ontario Trauma Registry, 2005–2022. Inclusion/exclusion detailed in article.

Partitioning scheme

Dataset 1 (development) split into training/validation/testing; testing cohort n=562. Dataset 2 used as a held-out consecutive external test set (n=612).

Missing information

Public release not described; contact third-party providers as per Acknowledgments.

Relationships between instances

Patient-level labels correspond to whether neurosurgical intervention occurred within 72 hours of scan acquisition.

Noise

Postoperative, duplicate, contrast-enhanced, and nondiagnostic motion-degraded scans were excluded; scans resampled and registered.

External data

Data analyzed were provided by Data Science and Advanced Analytics at Unity Health Toronto and the Ontario Trauma Registry (third-party).

Confidentiality

Retrospective clinical imaging linked to trauma registry data; REB approved with consent waived.

Sensitive data

Clinical variables (age, sex, ED GCS, in-hospital mortality) linked to imaging; procedural codes (ICD-10-CCI) used for labels.