Towards Improved Identification of Vertebral Fractures in Routine Computed Tomography (CT) Scans - Development and External Validation of a Machine Learning Algorithm

Aug 5, 2026

Vertebral fractures (VFs) are a key hallmark of osteoporosis, linked to substantial morbidity and mortality, yet they are often missed in routine patient CT scans. A machine learning algorithm was developed to identify VFs in abdominal/chest CT images and tested with external validation. Two independent datasets were used: a training set (n=1011) and an external validation cohort (n=2000) with Genant semiquantitative reference grading. An ensemble model achieved strong subject-level discrimination (AUROC 0.88) with high accuracy and specificity.

LanguageEnglish

CategoryHealthcare

FormatPDF

Pages12

Reading Time30 min

AccessFree to download

Abstract

Study objective and problem context

Data sources and reference standard

Model development and validation

Performance outcomes and clinical implications

University of Southern Denmark Towards Improved Identification of Vertebral Fractures in Routine Computed Tomography (CT) Scans Development and External Validation of a Machine Learning Algorithm Nicolaes, Joeri; Skjødt, Michael Kriegbaum; Raeymaeckers, Steven; Smith, Christopher Dyer; Abrahamsen, Bo; Fuerst, Thomas; Debois, Marc; Vandermeulen, Dirk; Libanati, Cesar Published in: Journal of Bone and Mineral Research DOI: 10.1002/jbmr.4916 Publication date: 2023 Document version: Final published version Document license: CC BY-NC-ND Citation for pulished version (APA): Nicolaes, J. , Skjødt, M. K. , Raeymaeckers, S. , Smith, C. D. , Abrahamsen, B. , Fuerst, T. , Debois, M. , Vandermeulen, D. , & Libanati, C. (2023) . Towards Improved Identification of Vertebral Fractures in Routine Computed Tomography (CT) Scans: Development and External Validation of a Machine Learning Algorithm. Journal of Bone and Mineral Research, 38(12), 1856-1866 . [https://doi.org/10.1002/jbmr.4916](https://doi.org/10.1002/jbmr.4916) Go to publication entry in University of Southern Denmark's Research Portal Terms of use This work is brought to you by the University of Southern Denmark. Unless otherwise specified it has been shared according to the terms for self-archiving. If no other license is stated, these terms apply: • You may download this work for personal use only. • You may not further distribute the material or use it for any profit-making activity or commercial gain • You may freely distribute the URL identifying this open access version If you believe that this document breaches copyright please contact us providing details and we will investigate your claim. Please direct all enquiries to [puresupport@bib.sdu.dk](puresupport@bib.sdu.dk) Download date: 05. Aug. 2026 RESEARCH ARTICLE Towards Improved Identiﬁcation of Vertebral Fracturesin Routine Computed Tomography (CT) Scans: Development and External Validation of a Machine Learning Algorithm Joeri Nicolaes, 1,2 Michael Kriegbaum Skjødt,3,4 Steven Raeymaeckers,5 Christopher Dyer Smith,4 Bo Abrahamsen,3,4,6 Thomas Fuerst,7 Marc Debois,2 Dirk Vandermeulen, 1 and Cesar Libanati2 1Department of Electrical Engineering (ESAT), Center for Processing Speech and Images, KU Leuven, Leuven, Belgium 2UCB Pharma, Brussels, Belgium 3Department of Medicine, Hospital of Holbæk, Holbæk, Denmark 4OPEN–Open Patient Data Explorative Network, Department of Clinical Research, University of Southern Denmark and Odense University Hospital, Odense, Denmark 5Department of Radiology, Universitair Ziekenhuis Brussel, Brussels, Belgium 6NDORMS, Nufﬁeld Department of Orthopaedics, Rheumatology and Musculoskeletal Sciences, Oxford University Hospitals, Oxford, UK 7Clario, Princeton, NJ, USA ABSTRACT Vertebral fractures (VFs) are the hallmark of osteoporosis, being one of the most frequent types of fragility fracture and an early sign of the disease. They are associated with signiﬁcant morbidity and mortality. VFs are incidentally found in one out of ﬁve imaging studies, however, more than half of the VFs are not identiﬁed nor reported in patient computed tomography (CT) scans. Our study aimed to develop a machine learning algorithm to identify VFs in abdominal/chest CT scans and evaluate its performance. We acquired two independent data sets of routine abdominal/chest CT scans of patients aged 50 years or older: a training set of 1011 scans from a non-interventional, prospective proof-of-concept study at the Universitair Ziekenhuis (UZ) Brussel and a validation set of 2000 subjects from an observational cohort study at the Hospital of Holbæ k. Both data sets were externally reevaluated to identify reference standard VF readings using the Genant semiquantitative (SQ) grading. Four independent models have been trained in a cross-validation experiment using the training set and an ensemble of four models has been applied to the external validation set.

FAQs

What problem does the study address?

Vertebral fractures are frequently overlooked in routine abdominal/chest CT scans, leading to missed early signs of osteoporosis.

How were vertebral fractures labeled for model training and validation?

Reference standard readings were generated using the Genant semiquantitative (SQ) grading.

What datasets and modelling approach were used?

The study used a training set of 1011 routine CT scans and an external validation set of 2000 subjects, training four models with cross-validation and applying an ensemble of four models for validation.

How well did the machine learning algorithm perform on external validation?

The ensemble model showed subject-level discrimination with an AUROC of 0.88, along with accuracy 0.92, sensitivity 0.81, and specificity 0.95.

Aug 5, 2026

Upload by Evangeline