Duer-Jensen Anne, Hørslev-Petersen Kim, Hetland Merete Lund, Bak Lene, Ejbjerg Bo Jannik, Hansen Michael Sejer, Johansen Julia Sidenius, Lindegaard Hanne Merete, Vinterberg Henrik, Møller Jakob M, Østergaard Mikkel
Department of Rheumatology, Copenhagen University Hospital at Hvidovre, Hvidovre, Denmark.
Arthritis Rheum. 2011 Aug;63(8):2192-202. doi: 10.1002/art.30396.
To study magnetic resonance imaging (MRI) as a tool for early diagnosis of rheumatoid arthritis (RA) in patients with early undifferentiated arthritis (UA).
Patients (n = 116) without a specific rheumatologic diagnosis, but with ≥2 tender joints and/or ≥2 swollen joints among the metacarpophalangeal, proximal interphalangeal, wrist, or metatarsophalangeal (MTP) joints for >6 weeks but <24 months, underwent clinical, biochemical, conventional radiographic, and MRI examinations and were followed up for >12 months for the final diagnosis of RA or non-RA. Based on univariate analyses, clinical, biochemical, and imaging parameters were selected for inclusion as explanatory variables in multiple logistic regression analysis, with development of RA as the dependent variable. A prediction model was developed, and its performance was tested and compared with that of a previous model developed by van der Helm-van Mil et al (the vdHvM model).
Of the 116 patients with early UA, 27 (23.3%) developed RA. When the prediction model was applied, which included as explanatory variables presence of hand arthritis, positivity for rheumatoid factor (RF), morning stiffness lasting >1 hour, and the Outcome Measures in Rheumatology Clinical Trials MRI summary score for bone edema in the MTP and wrist joints, the outcome of RA or non-RA was correctly identified in 82% of the patients (sensitivity 81%, specificity 82%). Another cutoff value for the prediction index in the model would allow a higher specificity (98%) and higher accuracy (83%), but lower sensitivity (36%). With the vdHvM model, RA/non-RA was predicted in 60.2% of the population.
MRI evidence of bone edema in the MTP and wrist joints is an independent predictor of future RA in patients with early UA. A prediction model that includes the variables clinical hand arthritis, morning stiffness, positivity for RF, and bone edema on MRI in the MTP and wrist joints correctly identified the development or lack of development of RA in 82% of patients.
研究磁共振成像（MRI）作为早期未分化关节炎（UA）患者类风湿关节炎（RA）早期诊断工具的价值。
116例无明确风湿性诊断，但掌指关节、近端指间关节、腕关节或跖趾关节（MTP）中≥2个关节压痛和/或≥2个关节肿胀持续>6周但<24个月的患者，接受临床、生化、传统放射学和MRI检查，并随访>12个月以最终诊断为RA或非RA。基于单因素分析，选择临床、生化和影像参数作为多因素逻辑回归分析中的解释变量，以RA的发生作为因变量。建立预测模型，并测试其性能，与van der Helm-van Mil等人之前建立的模型（vdHvM模型）进行比较。
116例早期UA患者中，27例（23.3%）发展为RA。应用该预测模型（其解释变量包括手部关节炎的存在、类风湿因子（RF）阳性、晨僵持续>1小时以及MTP和腕关节骨水肿的风湿病临床试验MRI总结评分）时，82%的患者RA或非RA的结局被正确识别（敏感性81%，特异性82%）。模型中预测指数的另一个临界值可使特异性更高（98%）和准确性更高（83%），但敏感性更低（36%）。使用vdHvM模型时，60.2%的人群被预测为RA/非RA。
MTP和腕关节骨水肿的MRI证据是早期UA患者未来发生RA的独立预测因素。一个包含临床手部关节炎、晨僵、RF阳性以及MTP和腕关节MRI骨水肿这些变量的预测模型，在82%的患者中正确识别了RA的发生或未发生情况。
