Accuracy and reliability of MRI quantitative measurements to assess spinal cord compression in cervical spondylotic myelopathy: a prospective study.
Accuracy and reliability of MRI quantitative measurements to assess spinal cord compression in cervical spondylotic myelopathy: a prospective study.
复制标题
DOI：
10.1055/s-0028-1100916
复制
发表时间：
2010-08-01
期刊：
影响因子：
--
通讯作者：
Fehlings, Michael G
中科院分区：
文献类型：
--
作者：
Karpova, Alina;Craciunas, Sorin;Fehlings, Michael G
研究类型：可靠性研究简介：脊髓型颈椎病（CSM）是55岁以上人群中最常见的脊髓疾病。尽管提出了多种神经影像学方法来量化CSM患者的脊髓损伤，磁共振成像（MRI）仍然是首选的程序，为临床决策提供有用的信息，确定最佳的治疗亚群，并选择最佳的治疗策略。然而，有效性，可靠性和准确性的MRI定量measurements尚未addressed.OBJECTIVE：评估内和观察者间的可靠性MRI定量测量脊髓compromise in CSM patients.METHODS：17例CSM患者（13名男性）的平均年龄54.5岁，从AOSpine北美数据库中选择。患者的狭窄节段（1-4个节段）和临床严重程度（范围mJOA基线：8-18）组合不同。排除无症状或既往手术治疗的CSM、活动性感染、肿瘤性疾病、类风湿性关节炎、强直性脊柱炎、创伤或伴随腰椎管狭窄。患者术前均接受了1.5T（15例患者）和3 T（2例患者）扫描仪的MRI检查，包括正中矢状位T1加权、轴位和正中矢状位T2加权序列。MRI数据由四名盲法评估者在三个不同阶段进行分析（Mango 2.0软件;多图像分析GUI）。分析了四个测量值：横截面积（TA）（图1）、压缩比（CR）（图2）、最大椎管损害（MCC）和最大脊髓压缩（MSCC）（图3）。使用混合效应ANOVA模型（大鼠、会话、大鼠x会话）评价每次测量的差异。使用组内相关系数（ICC）评价评价者内和评价者间可靠性（图4）。图1横截面积（TA）图2压缩比（CR=AP/W）图3最大椎管损害（MCC）和最大脊髓压迫（MSCC）。MCC（%）=1-[Dx/（Da+Db）/2] * 100%; MSCC（%）=1-[dx/（da+db）/2] * 100%图4组内相关系数（ICC）结果：主要发现是：（i）对于TA（71.48 ± 12.99mm 2），评定者内一致率为0.97（95%CI，范围0.94-0.99），评定者间一致性为0.76（95% CI，范围0.49-0.90）;（ii）CR（0.35 ± 0.04%），0.94（95% CI，范围0.88-0.98）和0.79（95% CI，范围0.57-0.91）;（iii）对于MCC（83.21 ± 2.08%），0.95（95% CI，范围0.89-0.98）和0.64（95% CI，范围0.28-0.85）;和（iv）对于MSCC（82.87 ± 1.52%），0.93（95% CI，范围0.86-0.97）和0.84（95% CI，范围0.65-0.93）。我们的数据表明，四个测量中的三个（TA，CR和MSCC）具有可接受的内部和内部可靠性系数（ICC > 0.75）。然而，对于最大根管损害测量，虽然内部可靠性是可以接受的，但评价者间可靠性是不可接受的（0.64）。在此基础上，我们建议在CSM患者的脊髓影像学评估中应使用三种MRI测量方法：横截面积、压缩比和脊髓最大压缩。
STUDY TYPE: Reliability study Introduction: Cervical spondylotic myelopathy (CSM) is the most common spinal cord disorder in persons more than 55 years old. Despite multiple neuroimaging approaches proposed to quantify the spinal cord compromise in CSM patients, magnetic resonance imaging (MRI) remains the procedure of choice by providing helpful information for clinical decision making, determining optimal subpopulations for treatment, and selecting the optimal treatment strategies. However, the validity, reliability, and accuracy of the MRI quantitative measurements have not yet been addressed.OBJECTIVE: To assess the intra- and inter-observer reliability of MRI quantitative measurements of the spinal cord compromise in CSM patients.METHODS: Seventeen CSM patients (13 male) of mean age 54.5 years old were selected from the AOSpine North America database. The patients had different combinations of stenotic levels (1-4 levels) and the clinical severity (range mJOA baseline: 8-18). Asymptomatic or previous surgically treated CSM, active infection, neoplastic disease, rheumatoid arthritis, ankylosing spondylitis, trauma, or concomitant lumbar stenosis were excluded. The patients underwent preoperative MRI using 1.5T (15 patients) and 3T (two patients) scanner, including mid-sagittal T1-weighted, axial and mid-sagittal T2-weighted series. MRI data were analyzed (Mango 2.0 software; Multi-Image Analysis GUI) by four blind raters in three different sessions. Four measurements were analysed: transverse area (TA) (Figure 1), compression ratio (CR) (Figure 2), maximal canal compromise (MCC), and maximal spinal cord compression (MSCC) (Figure 3). The differences for each measurement were evaluated using mixed-effect ANOVA models (ratter, session, ratter x session). The intra- and inter-rater reliability was evaluated with intraclass correlation coefficients (ICC) (Figure 4). Figure 1 Transverse area (TA)Figure 2 Compression ratio (CR=AP/W)Figure 3 Maximal canal compromise (MCC), and maximal spinal cord compression (MSCC). MCC(%)=1-[Dx/(Da+Db)/2] * 100%; MSCC(%)=1-[dx/(da+db)/2] * 100%Figure 4 Intraclass correlation coefficients (ICC) Results: The principal findings were: (i) for TA (71.48 ± 12.99mm2), the intra-rater agreement was 0.97 (95% CI, range 0.94-0.99) and the inter-rater agreement was 0.76 (95% CI, range 0.49-0.90); (ii) for CR (0.35 ± 0.04%), 0.94 (95% CI, range 0.88-0.98), and 0.79 (95% CI, range 0.57-0.91) respectively; (iii) for MCC (83.21 ± 2.08%), 0.95 (95% CI, range 0.89-0.98), and 0.64 (95% CI, range 0.28-0.85) respectively; and (iv) for MSCC (82.87 ± 1.52%), 0.93 (95% CI, range 0.86-0.97), and 0.84 (95% CI, range 0.65-0.93) respectively.CONCLUSIONS: Our data suggest that three out of four measurements (TA, CR and MSCC) have acceptable intra- and interreliability coefficients (ICC > 0.75). However, for the maximal canal compromise measure, although the intrareliability was acceptable, the inter-rater reliability was not acceptable (0.64). Based on this study, we recommend that three MRI measures: transverse area, compression ratio and maximal spinal cord compression should be used in the imaging assessment of the spinal cord in CSM patients.