【作者】 李辉；

【导师】 唐德秋；

【作者基本信息】 南华大学 ， 临床医学（专业学位）， 2020， 硕士

【摘要】 目的:探讨膝关节骨关节炎(knee osteoarthritis,KOA)患者通过双能量计算机断层成像(dual-energy computed tomography,DECT)后处理虚拟去钙(virtual noncalcium,VNCa)技术评估骨髓病变(bone marrow lesions,BMLs)的诊断效能,与磁共振成像(magnetic resonance imaging,MRI)进行比较;然后联合西安大略和麦克马斯特大学骨关节炎指数(Western Ontario and Mcmaster umirerosteoarthritis Index,WOMAC)、膝关节损伤和骨关节炎结果评分(Knee Injury and Osteoarthritis Outcome Score,KOOS)疼痛量表评估患者疼痛与BMLs的相关性。材料与方法:本研究于2017年8月至2018年11月最终纳入24名受试者(平均年龄59.2±7.4岁;年龄范围:43-76岁),共计28个膝关节,420个区域。所有受试者进行WOMAC、KOOS疼痛量表评估患者疼痛程度,在WOMAC量表进行疼痛分级处理。受试者均采用3.0 T高场磁共振和双能量CT扫描(包括图像后处理),由骨肌放射组医生根据膝关节骨关节炎评分系统(Knee Osteoarthritis Scoring System,KOSS)将膝关节分为15个区域进行BMLs分级,并对全膝关节进行总评分及分级。本研究以MRI为参考标准,一方面评估DECT-VNCa图像视觉定性、CT值定量诊断BMLs的效能及其优势,另一方面结合WOMAC、KOOS膝关节疼痛量表分析疼痛与DECT-VNCa图像上显示BMLs的相关性。结果:在DECT-VNCa图像视觉定性分析上,对BMLs诊断的灵敏度、特异度、阳性预测值和阴性预测值分别为83.7%、99.5%、95.3%、97.9%和89.8%、99.5%、95.7%、98.7%,观察者间有较好的一致性(Kappa=0.92);其次在CT值定量分析上表明VNCa图像的CT值在BMLs组和无BMLs组有显著性差异(p<0.001),通过ROC曲线得出cut-off值分别为-41HU和-47HU,进一步计算灵敏度、特异度分别为87.8%、99.5%和98.0%、99.5%。本研究中也发现不同程度BMLs与疼痛间存在相关性(p<0.001),在WOMAC量表中,MRI、DECT-VNCa显示BMLs与疼痛的相关系数r分别为0.803、0.779、0.718;在KOOS量表中,MRI、DECT-VNCa显示BMLs与疼痛的相关系数r分别为-0.711、-0.797、-0.691。结论:DECT-VNCa对BMLs的诊断及结合两种膝关节疼痛量表去预测关节疼痛具有一定可行性,能够作为MRI检查的重要补充手段,对于指导临床决策及治疗后评估有重要作用。更多还原

【Abstract】 Purpose: To evaluate the diagnostic efficacy of bone marrow lesions(BMLs)in patients with knee osteoarthritis(KOA)by dual energy computed tomography(DECT)post-processing technology virtual noncalcium(VNCa),Compared with magnetic resonance imaging(MRI),and then evaluated the correlation between pain and BMLs with Western Ontario and McMaster University Osteoarthritis Index(WOMAC)and knee injury and Osteoarthritis Outcome Score(KOOS)pain scale.Materials and Methods: From August 2017 to November 2018,24subjects(average age 59.2 ± 7.4 years;age range: 43-76 years)were included in this study,with a total of 28 knee joints and 420 regions.All the subjects were assessed by WOMAC and KOOS pain scale,and the pain was graded in WOMAC scale.All the subjects were scanned with3.0 T high field MRI and DECT(including image post-processing).The Knee Osteoarthritis Scoring System(KOSS)was used to divide the knee joint into 15 areas for BMLs classification,and the total score andclassification of the whole knee joint were carried out.In this study,MRI was used as a reference standard.On the one hand,the effectiveness and advantages of DECT-VNCa image visual qualitative and CT number quantitative diagnosis of BMLs were evaluated.On the other hand,the correlation between pain and the display of BMLs on DECT-VNCa image was analyzed by combining WOMAC and KOOS knee joint pain scale.Results: In the visual qualitative analysis of DECT-VNCa images,the sensitivity,specificity,positive predictive value and negative predictive value for BMLs diagnosis were 83.7%,99.5%,95.3%,97.9%and 89.8%,99.5%,95.7% and 98.7%,respectively.There was good consistency between the observers(Kappa = 0.92).Secondly,in terms of quantitative analysis of CT number,our study showed that the CT number of VNCa images were significantly different between the BMLs group and the non-BMLs group(p < 0.001),the cut-off values were-41 HU and-47 HU,respectively,according to ROC curve,and the further calculation sensitivity and specificity were 87.8%,99.5%,98.0% and 99.5%,respectively.In our experimental study,we also found an association between BMLs of different degrees and pain(p < 0.001).In WOMAC scale,the correlation coefficient between BMLs and pain was 0.803,0.779 and 0.718,respectively,according to MRI and DECT-VNCa.In the KOOS scale,the correlation coefficient between BMLs and pain was-0.711,-0.797 and-0.691,respectively,according to MRI and DECT-VNCa.Conclusion: DECT-VNCa is feasible for the diagnosis of BMLs and the combination of two knee pain scales to predict joint pain.It can be used as an important supplement to MRI,and plays an important role in guiding clinical decision-making and post treatment evaluation.更多还原