The evolution of T2-weighted intramedullary signal changes following ventral decompressive surgery for cervical spondylotic myelopathy

OBJECT: T2-weighted intramedullary increased signal intensity (ISI) on MRI in patients with cervical spondylotic myelopathy (CSM) appears to represent a wide spectrum of pathological changes that determine reversibility of cord damage. Although sharp T2-weighted ISI on preoperative imaging may correlate with poorer surgical outcomes, there are limited data on how these changes progress following surgery. In this study, the authors characterized pre-and postoperative ISI changes in patients undergoing surgery for CSM and studied their postoperative evolution in an attempt to quantify their clinical significance. METHODS: The preoperative and postoperative MR images obtained in 56 patients who underwent oblique cervical corpectomy for CSM were reviewed, and the ISI was classified into 4 subtypes based on margins and intensity: Type 0 (none), Type 1 ("fuzzy"), Type 2 ("sharp"), and Type 3 ("mixed"). The locations of the ISI were further classified as focal if they represented single discrete lesions, multifocal if there were multiple lesions with intervening normal cord, and multisegmental if the lesions were continuous over more than 1 segment. The maximum craniocaudal length of the ISI was measured on each midsagittal MR image. The Nurick grade and Japanese Orthopaedic Association (JOA) score were used to assess clinical status. The mean duration of follow-up was 28 months. RESULTS: T2-weighted ISI changes were noted preoperatively in 54 patients (96%). Most preoperative ISI changes were Type 1 (41%) or Type 3 (34%), with a significant trend toward Type 2 (71%) changes at follow-up. Multi-segmental and Type 3 lesions tended to regress significantly after surgery (p = 0.000), reducing to Type 2 changes at follow-up. Clinical outcomes did not correlate with ISI subtype; however, there was a statistically significant trend toward improvement in postoperative Nurick Grade in patients with a > 50% regression in ISI size. In addition, patients with more than 18 months of follow-up showed significant regression in ISI size compared with patients imaged earlier. On logistic regression analysis, preoperative Nurick grade and duration of follow-up were the only significant predictors of postoperative improvement in functional status (OR 4.136, p = 0.003, 95% CI 1.623-10.539 and OR 6.402, p = 0.033, 95% CI 1.165-35.176, respectively). CONCLUSIONS: There is a distinct group of patients with multisegmental Type 3 intramedullary changes who show remarkable radiological regression after surgery but demonstrate a residual sharp focal ISI at follow-up. A regression of the ISI by > 50% predicts better functional outcomes. Patients with a good preoperative functional status remain the most likely to show improvement, and the improvement continues to occur even at remote follow-up. The clinical relevance of the quality of the T2-weighted ISI changes in patients with CSM remains uncertain; however, postoperative regression of the ISI change is possibly a more important correlate of patient outcome than the quality of the ISI change alone.