MRI定义的膝关节组织损伤对膝骨关节炎发生与进展的预测作用：一项系统综述
简介
本系统综述纳入99项高质量研究，评估MRI检测的关节组织损伤对膝骨关节炎（KOA）发生与进展的预测价值。关键发现：基线骨髓病变（BMLs）、特定骨形态（OR最高达12.5）、半月板撕裂和滑膜炎是KOA发生的强预测因子；半月板挤压、滑膜炎和BMLs则预测KOA进展（以软骨损伤加重为标志）。值得注意的是，基线软骨T2信号异常是未来新发结构性软骨缺损的强力预测因子…
英文摘要
BACKGROUND: MRI is increasingly recognized not only for visualization of knee joint structures in knee osteoarthritis (KOA), but also for its potential to predict KOA incidence and progression. OBJECTIVE: We aim to provide a comprehensive overview of how different types of MRI-detected joint tissue pathology perform in predicting radiographic progression and longitudinal evolution of clinical outcomes and functional decline. METHODS: The study protocol was registered with the International Prospective Register of Systematic Reviews (PROSPERO; registration number CRD420251132451). A systematic literature search was performed in PubMed, Scopus, and Web of Science. After removal of duplicates, 4810 studies underwent a multi-step screening process, of which 99 were included in the qualitative synthesis. The quality of included studies was evaluated using the Newcastle-Ottawa Scale and the Downs and Black checklist. RESULTS: Most of the included studies were of good quality. Strong predictors of KOA incidence included baseline bone marrow lesions (BMLs), specific bone shape patterns (ORs up to 12.5 (95% CI, 4.0-39.3)), meniscal tears, and synovitis. Predictors of KOA progression, characterized by increasing cartilage damage, were meniscal extrusion, synovitis, and BMLs. Notably, baseline cartilage T2 signal abnormalities were a powerful predictor of the future development of new structural cartilage defects (OR = 21.3 (95% CI, 11.1, 40.6)), highlighting a pathway from compositional to structural deterioration in knees with and without pre-existing disease. CONCLUSION: Several MRI-detected joint tissue pathologies longitudinally associated with structural progression and clinically relevant outcomes, such as total knee arthroplasty, allowing patient stratification for disease-modifying osteoarthritis drug (DMOAD) trials. These associations may be further strengthened using compositional and multi-featured MRI models as well as AI-based feature extraction.