METHODS The authors evaluated preoperative modifications of the cervical spinal canal in flexion and extension in 50 patients with cervical spondylotic myelopathy (CSM) and looked for impingement of the spinal cord not diagnosed in the neutral position. OBJECTIVE To evaluate the usefulness of preoperative flexion-extension magnetic resonance imaging (MRI) for patients with CSM. BACKGROUND Dynamic factors contribute to CSM. Although the clinical manifestations and spinal or spinal cord morphology in patients with myelopathy have been reported, to our knowledge, there are no studies that include the cervical spinal cord length, sagittal diameter, and available space in patients with CSM in flexion, extension, and the neutral position. METHODS Dynamic MRI changes in canal stenosis during flexion-extension were evaluated in 50 patients with CSM in the supine position. The authors determined length of the cervical cord (LCC, C1-C7), cervical cord sagittal diameter (CCSD, C3-T1), cervical cord available space (CCAS, C3-T1), intramedullary high-intensity signal (IHIS) changes, number of stenosis, and severity of cord impingement in flexion, extension, and the neutral positions. RESULTS On both the anterior and posterior edges of the cord, mean LCC in flexion was longer than in extension or the neutral position and longer in the neutral position than in extension (P < 0.05). In all three positions, the average length of the anterior edge of the cervical cord was longer than the posterior edge (P < 0.05). The mean value of CCSD at each level in extension was greater than in flexion or the neutral position (P < 0.05). In the neutral position, CCSDs were greater than in flexion from C4 to C7 (P < 0.05), but this difference failed to reach significance at levels C3 and T1. In the neutral position, CCAS was greater than in either extension or flexion (P < 0.05), and CCAS was greater in flexion than in extension (P < 0.05) at all levels except C6, at which CCAS was greater in flexion than in either extension or the neutral position (P < 0.05). MRI demonstrated functional cord impingement (grade 3 of Mühle) in 6 of the 50 (12%) patients in flexion, in 17 patients (34%) in the neutral position, and in 37 patients (74%) in extension. IHIS was observed in flexion in 20 patients (40%), in the neutral position in 13 patients (26%), and in extension in 7 patients (14%). CONCLUSIONS Cervical spondylotic myelopathy results from the synergistic action of static and dynamic factors, the latter of which play an important role. In some patients, IHIS on T2 images is only visible with the neck in flexion. That might explain why IHIS is first detected after surgery in some patients in whom MRI was obtained before surgery only in the neutral position. Dynamic MRI is useful to determine more accurately the number of levels where the spinal cord is compromised, and to better evaluate narrowing of the canal and IHIS. New information provided by flexion-extension MRI might change our strategy for CSM management.