目的  探讨人工智能（AI）技术在骨质疏松性椎体骨折（OVF）影像评估中的准确性、效率及临床应用潜力，为优化骨折早期诊断和分级提供客观依据。方法回顾性分析100例OVF患者和300例正常对照的胸部CT影像数据，采用AI辅助诊断系统与影像医师双盲独立评估模式，分别对病例组和对照组胸部CT片第1胸椎（T1）至第1腰椎（L1）进行评估，以3位资深放射科医师共识为金标准，对两种方法进行一致性分析及效率对比。结果  AI系统与放射医师组在椎体骨折识别上具有高度一致性κ = 0.83（95% CI: 0.76~0.90）。其中，AI对椎体骨折的灵敏度达94.2%，与医师组（86.7%）差异有统计学意义（P<0.05），AI完成单例全胸椎CT分析耗时（16.6±3.2）s，显著短于医师组（89.7±21.4）s（P<0.001）。在AI预筛基础上进行针对性复核，使医师平均阅片时间减少81.5%。结论Genant半定量技术目测法存在重复性差的问题。将人工智能软件（如uAI Spine Analyzer）应用到椎体骨折评估中，可以帮助放射科医生提升读片效率，有效提高椎体骨折的诊断率，尤其在早期轻度骨折识别中表现突出，可作为临床辅助诊断的有效工具，借此密切监测骨健康，对实现“健康老龄化”战略目标具有积极意义。

ObjectiveTo evaluate the accuracy, efficiency, and clinical application potential of artificial intelligence (AI) technology in the imaging assessment of osteoporotic vertebral fractures (OVF), providing objective evidence for optimizing early fracture diagnosis and grading. MethodsA retrospective analysis was conducted on chest CT imaging data from 100 OVF patients and 300 normal controls. An AI-assisted diagnostic system and radiologists independently evaluated the thoracic 1(T1) to lumbar 1(L1) vertebrae in a double-blind manner. The consensus of three senior radiologists served as the gold standard. Agreement analysis and efficiency comparison were performed between the two methods. ResultsThe AI system demonstrated high agreement with radiologists in vertebral fracture detection (κ=0.83, 95% CI: 0.76-0.90). The AI system achieved a sensitivity of 94.2%, significantly higher than that of the radiologist group (86.7%, P<0.05). The AI system completed single-case whole-spine CT analysis in (16.6± 3.2) seconds, significantly faster than the radiologist group (89.7±21.4) seconds (P<0.001). When AI pre-screening was combined with targeted radiologist review, the average interpretation time was reduced by 81.5%. ConclusionThe conventional Genant semi-quantitative visual assessment method suffers from poor reproducibility. Integrating AI software (e.g., uAI Spine Analyzer) into vertebral fracture evaluation can enhance radiologists’ efficiency and significantly improve diagnostic accuracy, particularly in detecting early mild fractures. This AI-assisted approach serves as an effective clinical diagnostic tool, facilitating close monitoring of bone health and contributing to the strategic goal of “healthy aging.”