Diagnosis Of Magnetic Resonance Imaging In Cervical Ligament Injury And Analysis Of Correlative Factors For Anterior Longitudinal Ligament Injury

ObjectiveTo investigate the reliability of different magnetic resonance imaging (MRI) features in detecting subaxial cervical anterior longitudinal ligament(ALL) and posterior longitudinal ligament(PLL) injury, and analysis the possible correlative factors of ALL injury, to conduct clinical diagnosis and avoid missed diagnosis.Methods87patients with subaxial cervical spinal injury who were no fracture or dislocation of certebral body were operated by anterior approach or anterior approach combined with posterior approach. X-ray.Computed tomography(CT) and MRI were performed in all patients pre-operation. MRI feature of discontinuity on sagittal T1-weighted image(TiD), longitudinal high signal intensity on sagittal T2-weighted image(T2L), transversal high or intermediate signal intensity on sagittal T2-weighted image (T2T) was used to assess the status of anterior longitudinal ligament(ALL) and posterior longitudinal ligament(PLL) by two observers. During the operation, injury of ALL and PLL was carefully examined. The MRI features were compared with surgical findings. Interobserver agreement for each assessment was determined using the kappa statistic. The sensitivity, specificity, accuracy, positive and negative predictive values of different MRI features in detection of ALL and PLL injury were calculated using operative finding as the gold standard. Sex, age, American Spinal Injury Association(ASIA)motor scores, facial injury,intramedullary MRI T2WI high signal, prevertebral osteophyte of observed segments, intervertebral disc degeneration were selected to be conducted with single factor analysis and multiple-factors Logistic regression to explore the related factors for ALL.ResultsInterobserver agreement of T1D for the assessment of ALL and PLL injury was not satisfactory (Kappa value=0.152,0.238), and the sensitivity, specificity of T1D by comparing with the operative finding was low. For T2L, the interobserver agreement was good(Kappa value=0.657,0.607), and there was a high specificity by comparing with the operative finding (78.4～80.4%.88.7～90.4%), however, the sensitivity was only54.1～65.8%and29.4～50%. By contrast, the interobserver agreement of T2T for the assessment of ALL and PLL injury was excellent (Kappa value=0.837,0.799), and there was high sensitivity (83.6～86.9%,82.4～88.2%), specificity (91.8～95.9%.90.31～91.9%), accuracy (89.9～91.1%,89.9-89.9%), positive predictive value (86.9～92.7%,71.4～73.7%) and negative predictive value (90.3～91.8%,95.0～ 96.6%) There were no differences in sex, age between ALL injury group and no ALL injury group(P=0.207,0.185). ASIA motor scores was significantly low in ALL injury group than no injury group (P<0.001). There were differences in facial injury, intramedullary high signal,and intervertebral disc degeneration(p<0.05) and was significantly difference in osteophyte before the vertebral body(P<0.001) between ALL injury group and no ALL injury group. Multiple-factors Logistic regression showed that facial injury, prevertebral osteophyte was associated with ALL injury but intervertebral disc degeneration had no relationship with ALL injury.Conclusion1、The T1D and T2L of MRI cannot be used as a reliable sign of ligament injury in subaxial cervical spine ALL and PLL injury. However, the T2T of MRI was a reliable indicator and would be helpful in assessing the stability of cervical spine.2、Facial injury, prevertebral osteophyte was correlated with ALL injury. The prevertebral osteophyte is high correlated with ALL injury.