Toward End-to-end Artificial Intelligence Systems for Radiographic Fracture Assessment: A Critical Review of Detection, Explainability, and Prognosis
Abstract
This review synthesizes recent advances in fracture intelligence through a unified seven-stage end-to-end pipeline that includes preprocessing and (region of interest) ROI extraction, fracture detection, classification, localization and segmentation, prognostic modeling, explainability, and clinical deployment. By shifting the focus from isolated algorithms to complete clinical workflows, we expose a critical gap: current systems excel at basic detection but struggle with complex prognostic modeling and real-world deployment. Within this framework, we critically analyze methodological progress, highlight enduring research challenges, and demonstrate that fracture artificial intelligence continues to be detection focused with relatively less advancement in quantitative fracture characterization, prognostic intelligence, intrinsically interpretable learning, and prospective clinical validation. Based on these findings, we outline future research directions toward unified, anatomically informed, clinically deployable fracture intelligence systems.
Related
- The role of radiographic knee alignment in total knee replacement outcomes and opportunities for artificial intelligence-driven assessment
- Reliable Detection, Unreliable Clinical Decision-making: Artificial Intelligence Models in Burn Assessment
- Artificial intelligence-based assessment of gastric mucosal exposure and its association with lesion detection: a multicenter study
- Development and validation of an artificial intelligence model for basal cell carcinoma detection and section completeness assessment in Mohs micrographic surgery in an Asian cohort
- Perceptions of artificial intelligence for radiographic image accuracy: a cross-sectional survey of radiology personnel
BibTeX
@article{sheikh2026toward,
title = {{Toward End-to-end Artificial Intelligence Systems for Radiographic Fracture Assessment: A Critical Review of Detection, Explainability, and Prognosis}},
author = {Arshi Sheikh and Shailendra Singh},
journal = {Journal of Medical Physics},
year = {2026},
month = jul,
doi = {10.4103/jmp.jmp_67_26},
url = {https://doi.org/10.4103/jmp.jmp_67_26},
note = {CC BY-NC-SA},
}
RIS
TY - JOUR
TI - Toward End-to-end Artificial Intelligence Systems for Radiographic Fracture Assessment: A Critical Review of Detection, Explainability, and Prognosis
AU - Arshi Sheikh
AU - Shailendra Singh
PY - 2026
DA - 2026/07/01
JO - Journal of Medical Physics
DO - 10.4103/jmp.jmp_67_26
UR - https://doi.org/10.4103/jmp.jmp_67_26
AB - This review synthesizes recent advances in fracture intelligence through a unified seven-stage end-to-end pipeline that includes preprocessing and (region of interest) ROI extraction, fracture detection, classification, localization and segmentation, prognostic modeling, explainability, and clinical deployment. By shifting the focus from isolated algorithms to complete clinical workflows, we expose a critical gap: current systems excel at basic detection but struggle with complex prognostic modeling and real-world deployment. Within this framework, we critically analyze methodological progress, highlight enduring research challenges, and demonstrate that fracture artificial intelligence continues to be detection focused with relatively less advancement in quantitative fracture characterization, prognostic intelligence, intrinsically interpretable learning, and prospective clinical validation. Based on these findings, we outline future research directions toward unified, anatomically informed, clinically deployable fracture intelligence systems.
SN - 0971-6203
SN - 1998-3913
ER -