The Potential Clinical Utility of an Artificial Intelligence Model for Identification of Vertebral Compression Fractures in Chest Radiographs

Ankita Ghatak; James M Hillis; Sarah F Mercaldo; Isabella Newbury-Chaet; John K Chin; Subba R Digumarthy; Karen Rodriguez; Victorine V Muse; Katherine P Andriole; Keith J Dreyer; Mannudeep K Kalra; Bernardo C Bizzo

doi:10.1016/j.jacr.2024.08.026

The Potential Clinical Utility of an Artificial Intelligence Model for Identification of Vertebral Compression Fractures in Chest Radiographs

J Am Coll Radiol. 2024 Sep 17:S1546-1440(24)00766-X. doi: 10.1016/j.jacr.2024.08.026. Online ahead of print.

Affiliations

¹ Mass General Brigham AI, Boston, Massachusetts.
² Director of Clinical Operations, Mass General Brigham AI, Boston, MA; Assistant in Neurology, Department of Neurology, Massachusetts General Hospital, Boston, Massachusetts; and Instructor in Neurology, Harvard Medical School, Boston, Massachusetts.
³ Lead Biostatistician, Mass General Brigham AI, Boston, Massachusetts; Instructor in Radiology/Statistics, Harvard Medical School, Boston, Massachusetts; and Statistician, Department of Radiology, Massachusetts General Hospital, Boston, Massachusetts.
⁴ Associate Professor of Radiology, Harvard Medical School, Boston, Massachusetts, and Department of Radiology, Massachusetts General Hospital, Boston, Massachusetts.
⁵ Harvard Medical School, Boston, Massachusetts, and Thoracic Fellowship Program Director, Department of Radiology, Massachusetts General Hospital, Boston, Massachusetts.
⁶ Assistant Professor of Radiology, Harvard Medical School, Boston, Massachusetts, and Department of Radiology, Massachusetts General Hospital, Boston, Massachusetts.
⁷ Director of Academic Research and Education, Mass General Brigham AI, Boston, Massachusetts; Associate Professor of Radiology, Harvard Medical School, Boston, Massachusetts; Associate Professor of Radiology, Department of Radiology, Brigham and Women's Hospital, Boston, Massachusetts; Senior Scientist for Education, ACR Data Science Institute; and ACR Committee on Practice Parameters and Technical Standards.
⁸ Chief Data Science Officer, Mass General Brigham AI, Boston, Massachusetts; Associate Professor of Radiology, Harvard Medical School, Boston, Massachusetts; Department of Radiology, Massachusetts General Hospital, Boston, Massachusetts; Vice Chair of Informatics at Mass General Brigham Radiology; and Chief Science Officer at ACR Data Science Institute.
⁹ Scientific Director, Mass General Brigham AI, Boston, Massachusetts; Professor of Radiology, Harvard Medical School, Boston, Massachusetts; Department of Radiology, Massachusetts General Hospital, Boston, Massachusetts; and Director at Webster Center for Quality and Safety.
¹⁰ Senior Director, Mass General Brigham AI, Boston, Massachusetts; Assistant Professor of Radiology, Harvard Medical School, Boston, Massachusetts; Department of Radiology, Massachusetts General Hospital, Boston, Massachusetts; and Associate Chief Science Officer at ACR Data Science Institute. Electronic address: bbizzo@mgh.harvard.edu.

PMID: 39299617
DOI: 10.1016/j.jacr.2024.08.026

Abstract

Purpose: To assess the ability of the Annalise Enterprise CXR Triage Trauma (Annalise AI Pty Ltd, Sydney, NSW, Australia) artificial intelligence model to identify vertebral compression fractures on chest radiographs and its potential to address undiagnosed osteoporosis and its treatment.

Materials and methods: This retrospective study used a consecutive cohort of 596 chest radiographs from four US hospitals between 2015 and 2021. Each radiograph included both frontal (anteroposterior or posteroanterior) and lateral projections. These radiographs were assessed for the presence of vertebral compression fracture in a consensus manner by up to three thoracic radiologists. The model then performed inference on the cases. A chart review was also performed for the presence of osteoporosis-related International Classification of Diseases, 10th revision diagnostic codes and medication use for the study period and an additional year of follow-up.

Results: The model successfully completed inference on 595 cases (99.8%); these cases included 272 positive cases and 323 negative cases. The model performed with area under the receiver operating characteristic curve of 0.955 (95% confidence interval [CI]: 0.939-0.968), sensitivity 89.3% (95% CI: 85.7%-92.7%) and specificity 89.2% (95% CI: 85.4%-92.3%). Out of the 236 true-positive cases (ie, correctly identified vertebral compression fractures by the model) with available chart information, only 86 (36.4%) had a diagnosis of vertebral compression fracture and 140 (59.3%) had a diagnosis of either osteoporosis or osteopenia; only 78 (33.1%) were receiving a disease-modifying medication for osteoporosis.

Conclusion: The model identified vertebral compression fracture accurately with a sensitivity 89.3% (95% CI: 85.7%-92.7%) and specificity of 89.2% (95% CI: 85.4%-92.3%). Its automated use could help identify patients who have undiagnosed osteoporosis and who may benefit from taking disease-modifying medications.

Keywords: Artificial intelligence; chest radiography; medical code review; radiology; vertebral compression fracture.