Magnetic resonance imaging evaluation of common peroneal nerve injury in acute and subacute posterolateral corner lesion: a retrospective study

Abstract Objective: To evaluate qualitative and quantitative magnetic resonance imaging (MRI) criteria for injury of the common peroneal nerve (CPN) in patients with acute or subacute injuries in the posterolateral corner (PLC) of the knee, as well as to evaluate the reproducibility of MRI evaluation of CPN alterations. Materials and Methods: This was a retrospective study of 38 consecutive patients submitted to MRI and diagnosed with acute or subacute injury to the PLC of the knee (patient group) and 38 patients with normal MRI results (control group). Two musculoskeletal radiologists (designated radiologist A and radiologist B, respectively) evaluated the images. Nerve injury was classified as neurapraxia, axonotmesis, or neurotmesis. Signal strength was measured at the CPN, the tibial nerve (TN), and a superficial vein (SV). The CPN/TN and CPN/SV signal ratios were calculated. The status of each PLC structure, including the popliteal tendon, arcuate ligament, lateral collateral ligament, and biceps tendon, was classified as normal, partially torn, or completely torn, as was that of the cruciate ligaments. For the semiquantitative analysis of interobserver agreement, the kappa statistic was calculated, whereas a receiver operating characteristic (ROC) curve was used for the quantitative analysis. Results: In the patient group, radiologist A found CPN abnormalities in 15 cases (39.4%)-neurapraxia in eight and axonotmesis in seven-whereas radiologist B found CPN abnormalities in 14 (36.8%)-neurapraxia in nine and axonotmesis in five. The kappa statistic showed excellent interobserver agreement. In the control group, the CPN/TN signal ratio ranged from 0.63 to 1.1 and the CPN/SV signal ratio ranged from 0.16 to 0.41, compared with 1.30-4.02 and 0.27-1.08, respectively, in the patient group. The ROC curve analysis demonstrated that the CPN/TN signal ratio at a cutoff value of 1.39 had high (93.3%) specificity for the identification of nerve damage, compared with 81.3% for the CPN/SV signal ratio at a cutoff value of 0.41. Conclusion: CPN alterations are common in patients with PLC injury detected on MRI, and the level of interobserver agreement for such alterations was excellent. Calculating the CPN/TN and CPN/SV signal ratios may increase diagnostic confidence. We recommend systematic analysis of the CPN in cases of PLC injury.

摘要
目的：评估针对膝关节后外侧角（posterolateral corner，PLC）急性或亚急性损伤患者腓总神经（common peroneal nerve，CPN）损伤的定性与定量磁共振成像（magnetic resonance imaging，MRI）标准，并评估MRI评估CPN异常的可重复性。
材料与方法：本研究为回顾性研究，纳入38例接受MRI检查且确诊为膝关节PLC急性或亚急性损伤的连续患者（患者组），以及38例MRI结果正常的患者（对照组）。由2名肌肉骨骼放射科医师（分别记为医师A与医师B）对图像进行阅片。将神经损伤分为神经失用症、轴突断伤及神经断伤三类。于CPN、胫神经（tibial nerve，TN）及浅静脉（superficial vein，SV）处测量信号强度，并计算CPN/TN及CPN/SV信号比值。对包括腘肌腱、弓状韧带、外侧副韧带及股二头肌腱在内的各PLC结构，以及交叉韧带的状态，均分为正常、部分撕裂及完全撕裂三类。对于观察者间一致性的半定量分析，采用kappa统计量；定量分析则采用受试者工作特征（receiver operating characteristic，ROC）曲线。
结果：患者组中，医师A检出15例（39.4%）CPN异常，其中神经失用症8例、轴突断伤7例；医师B检出14例（36.8%）CPN异常，其中神经失用症9例、轴突断伤5例。kappa统计量结果显示观察者间一致性极佳。对照组中，CPN/TN信号比值范围为0.63~1.1，CPN/SV信号比值范围为0.16~0.41；而患者组对应比值范围分别为1.30~4.02与0.27~1.08。ROC曲线分析显示，当CPN/TN信号比值截断值取1.39时，识别神经损伤的特异性高达93.3%；而CPN/SV信号比值截断值取0.41时，特异性为81.3%。
结论：MRI检出的PLC损伤患者中，CPN异常较为常见，且此类异常的观察者间一致性极佳。计算CPN/TN与CPN/SV信号比值可提升诊断置信度。我们建议对PLC损伤患者常规开展CPN系统评估。