Scholarlink

Background <p>Musculoskeletal (MSK) disorders affect approximately 1.71 billion people worldwide and are a leading cause of disability, with care dependent on imaging interpretation. Diagnostic variability, workload, and limited quantitative biomarkers challenge current practice. Artificial intelligence (AI) can address these issues by improving diagnostic accuracy and efficiency, but clinical readiness and real-world impact are not yet established. This review critically evaluates AI applications in MSK imaging, emphasizing clinical validation and workflow integration, and proposes a framework for translating AI into practice.</p> Methods <p>A systematic search was conducted across PubMed/MEDLINE, Scopus, Web of Science, IEEE Xplore, and arXiv for studies published between January 2022 and December 2025. A total of 1,245 records were identified, 892 records were screened after duplicate removal, and 123 full-text reports were assessed for eligibility. Study quality, risk of bias, and implementation readiness were evaluated using criteria adapted from CLAIM, CONSORT-AI, QUADAS-2, PROBAST, and FDA AI/ML SaMD guidance.</p> Results <p>Of 123 full-text reports assessed, only three studies met the strict criteria for formal imaging-based AI synthesis requiring quantitative, imaging-based validation. All three formally included studies were exclusively focused on MRI, with formal evidence limited to MRI-based segmentation and quantitative biomarker extraction. Although these systems demonstrated promising technical performance under controlled conditions, clinical evidence remained limited by single-center designs, lack of external validation, absence of cost-effectiveness analysis, and lack of patient-centered outcome assessment.</p> Conclusions <p>AI has strong potential to improve diagnostic consistency, quantitative assessment, and clinical decision support in MSK care. However, the current formal evidence base remains technically promising but clinically under-validated and does not yet support widespread clinical adoption. Future research should prioritize multi-center prospective validation, workflow integration, patient-centered outcomes, cost-effectiveness, and post-deployment monitoring.</p> Trial registration <p>The review protocol was retrospectively registered on OSF (<a href="https://doi.org/10.17605/OSF.IO/FPAEJ">https://doi.org/10.17605/OSF.IO/FPAEJ</a>).</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Clinical translation of artificial intelligence in musculoskeletal care: a systematic review of current applications, evidence gaps, and implementation readiness

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年7月30日

查看全文
  • Zainab Ghazanfar,
  • Areej Fatima,
  • Saba Ghazanfar Ali,
  • Younhyun Jung

摘要

Background

Musculoskeletal (MSK) disorders affect approximately 1.71 billion people worldwide and are a leading cause of disability, with care dependent on imaging interpretation. Diagnostic variability, workload, and limited quantitative biomarkers challenge current practice. Artificial intelligence (AI) can address these issues by improving diagnostic accuracy and efficiency, but clinical readiness and real-world impact are not yet established. This review critically evaluates AI applications in MSK imaging, emphasizing clinical validation and workflow integration, and proposes a framework for translating AI into practice.

Methods

A systematic search was conducted across PubMed/MEDLINE, Scopus, Web of Science, IEEE Xplore, and arXiv for studies published between January 2022 and December 2025. A total of 1,245 records were identified, 892 records were screened after duplicate removal, and 123 full-text reports were assessed for eligibility. Study quality, risk of bias, and implementation readiness were evaluated using criteria adapted from CLAIM, CONSORT-AI, QUADAS-2, PROBAST, and FDA AI/ML SaMD guidance.

Results

Of 123 full-text reports assessed, only three studies met the strict criteria for formal imaging-based AI synthesis requiring quantitative, imaging-based validation. All three formally included studies were exclusively focused on MRI, with formal evidence limited to MRI-based segmentation and quantitative biomarker extraction. Although these systems demonstrated promising technical performance under controlled conditions, clinical evidence remained limited by single-center designs, lack of external validation, absence of cost-effectiveness analysis, and lack of patient-centered outcome assessment.

Conclusions

AI has strong potential to improve diagnostic consistency, quantitative assessment, and clinical decision support in MSK care. However, the current formal evidence base remains technically promising but clinically under-validated and does not yet support widespread clinical adoption. Future research should prioritize multi-center prospective validation, workflow integration, patient-centered outcomes, cost-effectiveness, and post-deployment monitoring.

Trial registration

The review protocol was retrospectively registered on OSF (https://doi.org/10.17605/OSF.IO/FPAEJ).