Artificial intelligence-guided clavicle fracture detection on plain radiographs: A retrospective diagnostic accuracy study.

Clinical Snapshot

Key Findings

• P Value: Not reported in abstract

• Effect Size: Cohen kappa 0.95 (AP radiographs) — indicating near-perfect agreement between AI and senior radiologist reference standard

• Primary Outcome: AI-guided clavicle fracture detection on AP radiographs: accuracy 97.31%, sensitivity 95.83%, specificity 98.89%, Cohen kappa 0.95, F1 score 0.97

• Nnt Or Sensitivity: Sensitivity 95.83%, Specificity 98.89% (AP radiographs); positive and negative likelihood ratios not explicitly reported. Derived positive LR ≈ 86.2, derived negative LR ≈ 0.042 based on reported figures, though these cannot be confirmed without raw cell counts

• Confidence Interval: Not reported for any metric — a significant methodological limitation

Clinical Application

AI-assisted fracture detection tools are increasingly available on commercial PACS platforms and radiology workflow systems. The study demonstrates post-implementation real-world data collection, suggesting feasibility of integration. However, the unnamed AI system limits direct clinical translation. Workflow integration, radiologist oversight requirements, and failure mode characterisation remain undefined. No AI-specific fracture detection tool for clavicle fractures is currently listed on the Australian Register of Therapeutic Goods (ARTG) as a standalone Class IIb or III medical device, though several AI radiology platforms with broader musculoskeletal modules have received TGA clearance. RACGP and RACS guidelines do not currently reference AI-assisted fracture detection for clavicle injuries. In Australian emergency departments and rural/remote settings, where after-hours radiologist coverage may be limited, AI decision-support tools could theoretically reduce diagnostic delays — however, this study's single-centre European data and absence of confidence intervals do not provide sufficient evidence to support PBS-funded or TGA-mandated adoption at this stage. Australian clinicians should await prospective multicentre validation, ideally including Australian patient populations, before integrating such tools into standard care pathways. Adult and paediatric patients (mean age 44.6 years, range implied by SD of 26.2) presenting with suspected clavicle fracture requiring plain radiographic assessment in an orthopaedic or emergency setting

Abstract

Extensive research exists on artificial intelligence (AI)-assisted fracture detection in various anatomical regions, but clavicle fractures remain comparatively understudied. This study aimed to compare the diagnostic performance of AI-guided clavicle fracture detection on plain radiographs with that of an experienced orthopedic surgeon. We retrospectively analyzed all clavicle radiographs obtained at our institution between September 11,2023 and October 31,2024 following the clinical implementation of AI-guided fracture detection. Radiological findings confirmed by a senior radiologist served as the reference standard. AI performance was additionally compared with the assessments of an experienced, blinded orthopedic surgeon. Diagnostic performance metrics included accuracy, sensitivity, specificity, Cohen kappa, F1 score, and Youden index. A total of 367 clavicle radiographs were included, comprising 186 anteroposterior (AP) and 181 second-plane views. The mean patient age was 44.6 ± 26.2 years, and 59.4% were male. According to the reference standard, clavicle fractures were present in 96 AP radiographs (51.6%) and 95 second-plane radiographs (52.5%). On AP radiographs, the AI system achieved an accuracy of 97.31%, sensitivity of 95.83%, specificity of 98.89%, Cohen kappa of 0.95, and F1 score of 0.97. Diagnostic performance on second-plane radiographs was similarly high and overall comparable to that of the experienced orthopedic surgeon. This retrospective diagnostic accuracy study shows that AI-guided clavicle fracture detection is highly accurate and reliable, with performance approaching that of an experienced orthopedic surgeon. These findings support the potential role of AI as a diagnostic support tool in clavicle fracture detection, while prospective and multicenter validation remains warranted.

References

1. 1.Ramadanov, N., Hable, R., Voss, M., Becker, R., Schreyer, A. G., Prill, R., & Salzmann, M. (2026). Artificial intelligence-guided clavicle fracture detection on plain radiographs: A retrospective diagnostic accuracy study. Medicine. https://doi.org/10.1097/MD.0000000000049488