Study design: A cross-sectional analysis.
Objective: To elucidate the accuracy of neurologic level diagnosis of cervical stenotic myelopathy.
Summary of background data: Neurologic level diagnosis in cervical myelopathy has not been well established.
Methods: A total of 106 patients with cervical stenotic myelopathy, with a single-level intramedullary high-intensity area confirmed on both preoperative and postoperative T2-weighted magnetic resonance imaging (MRI), were included in this study. We performed a level diagnosis on the basis of neurologic signs (the uppermost muscle with weakness, diminished or exaggerated deep tendon reflex, the uppermost level of sensory disturbance of the upper extremities) and compared it with a level diagnosis made by T2-weighted MRI. The sensitivity, specificity, and accuracy of neurologic signs on our index corresponding to each intervertebral level were calculated.
Results: The averages of sensitivity, specificity, and accuracy were 42%, 80%, and 70%, respectively, in the uppermost muscle with weakness, 66%, 89%, and 83% in deep tendon reflex, and 74%, 91%, and 87% in the sensory disturbance area. The positive and negative predictive values were 40% and 91%, respectively, in the uppermost muscle with weakness, 66% and 89% in deep tendon reflex, and 74% and 91% in the sensory disturbance area. Accuracy of a diagnosis based on muscle weakness was less high, the reason being that in many patients, the uppermost muscle with weakness was extensor digiti communis or the intrinsic muscles of the hands, and this led to a lower sensitivity.
Conclusions: The average accuracy of neurologic level diagnosis based on the index we proposed was > or =70%. The level diagnosis by a sensory disturbance area showed the highest accuracy (87%).