Deep Learning in Vertebral Fracture Detection: Systematic Review and Meta-analysis of Subject- vs. Vertebra-Level Approaches.

Journal: Academic radiology
Published Date: Dec 6, 2025

Abstract

RATIONALE AND OBJECTIVES: To provide a context-aware evaluation of deep learning algorithms for vertebral fracture detection by disentangling subject-level from vertebra-level approaches, quantifying the influence of key technical and methodological factors, and generating evidence to guide task-specific clinical use and standardized reporting. MATERIALS AND METHODS: In this PRISMA‑compliant review (PROSPERO CRD42024523301), five databases were searched to February 2025 for English‑language studies reporting accuracy metrics. Risk of bias was assessed with QUADAS‑AI. Hierarchical summary ROC models pooled sensitivity, specificity, and AUC for each analytical level; subgroup analysis and meta‑regression explored heterogeneity by test‑set origin, imaging modality, and scanner vendor. RESULTS: 36 studies (96,956 patients; 171,552 images) were eligible; 28 provided 113 contingency tables. Pooled subject‑level sensitivity/specificity was 84%/91% (AUC 0.94); vertebra‑level 80%/97% (AUC 0.96). Subject-level models prioritized sensitivity, whereas vertebra-level models achieved higher specificity with precise localization. External validation lowered sensitivity, yet retained high specificity. Radiographs favored subject‑level screening, whereas CT supported vertebra‑level precision. Multi‑vendor datasets improved subject‑level sensitivity, and single‑vendor datasets enhanced vertebra‑level specificity. Methodological quality varied across studies; QUADAS-AI identified high risk of patient selection bias-the most commonly identified bias source-in 61% of studies. CONCLUSION: Deep learning models demonstrate high accuracy for vertebral fracture detection; subject-level approaches are suited to screening/triage due to higher sensitivity, whereas vertebra-level approaches offer higher specificity and precise localization for confirmatory diagnosis and treatment planning. Given performance variability across imaging modality and data sources, clinical use should align model granularity with the intended task and context.

Authors

  • Mohammad-Reza Hosseini-Siyanaki
    Department of Radiology, University of Florida, Gainesville, Florida (M.H., H.S.S., A.R., S.M., A.R., A.D., K.R.P.).
  • Babak Ahmadi
    Department of Industrial and Systems Engineering, University of Florida, Gainesville, Florida (B.A.); Department of Neurology, University of Florida, Gainesville, Florida (B.A., A.B.F.); Magnetoencephalography (MEG) Lab, The Norman Fixel Institute of Neurological Diseases, University of Florida Health, Gainesville, Florida (B.A., A.B.).
  • Hakki Serdar Sagdic
    Radiomics and Augmented Intelligence Laboratory (RAIL), Department of Radiology and the Norman Fixel Institute for Neurological Diseases, University of Florida College of Medicine, Gainesville, FL (M.H-S., H.S.S., A.G.R., S.E.M., J.C.P., E.Y.A., B.H., R.F.); Department of Radiology, University of Florida College of Medicine, Gainesville, FL (M.H-S., H.S.S., A.G.R., J.C.P., E.Y.A., B.H., R.F.).
  • Abheek Raviprasad
    Radiomics and Augmented Intelligence Laboratory (RAIL), Department of Radiology and the Norman Fixel Institute for Neurological Diseases, University of Florida College of Medicine, Gainesville, FL, USA.
  • Sefat Munjerin
    Radiomics and Augmented Intelligence Laboratory (RAIL), Department of Radiology and the Norman Fixel Institute for Neurological Diseases, University of Florida College of Medicine, Gainesville, FL, USA.
  • Antika Roy
    Radiomics and Augmented Intelligence Laboratory (RAIL), Department of Radiology and the Norman Fixel Institute for Neurological Diseases, University of Florida College of Medicine, 1600 SW Archer Road, Gainesville, FL, 32610-0374, USA.
  • Aysha Dogan
    Department of Radiology, University of Florida, Gainesville, Florida (M.H., H.S.S., A.R., S.M., A.R., A.D., K.R.P.).
  • Abbas Babajani-Feremi
    Department of Pediatrics, Division of Clinical Neurosciences, University of Tennessee Health Science Center, Memphis, TN, USA; Neuroscience Institute, Le Bonheur Children's Hospital, Memphis, TN, USA; Department of Anatomy and Neurobiology, University of Tennessee Health Science Center, Memphis, TN, USA. Electronic address: [email protected].
  • Keith R Peters
    Radiomics and Augmented Intelligence Laboratory (RAIL), Department of Radiology and the Norman Fixel Institute for Neurological Diseases, University of Florida College of Medicine, Gainesville, FL; Department of Radiology, University of Florida College of Medicine, Gainesville, FL.

Keywords

No keywords available for this article.

External Resources

Popular Topics
Recent Journals