Resting-state Amplitude of Low-frequency Fluctuation is a Potentially Useful Prognostic Functional Biomarker in Cervical Myelopathy

Background Cervical MRI is the standard diagnostic imaging technique for patients with cervical myelopathy. However, the utility of conventional cervical MRI as a predictive biomarker for surgical recovery remains unclear, partly because of the limited information obtained from this anatomically small area. Brain resting-state functional MRI (rs-fMRI) may help identify candidate predictive biomarkers. Two analytical methods that assess local spontaneous brain activity are widely used for rs-fMRI: functional connectivity between two brain regions and amplitude of low-frequency fluctuation (ALFF). In our previous analysis of functional connectivity, we discovered that brain functional connectivity may be a predictive biomarker for neurologic recovery in patients with cervical myelopathy; however, the functional connectivity analysis identified a correlation with only one clinical outcome (the 10-second test). To establish a comprehensive prediction measure, we need to explore other brain biomarkers that can predict recovery of other clinical outcomes in patients with cervical myelopathy. Questions/purposes We aimed to (1) elucidate preoperative ALFF alterations in patients with cervical myelopathy and how ALFF changes after surgery, with a focus on postoperative normalization and (2) establish a predictive model using preoperative ALFF by investigating the correlation between preoperative ALFF and postoperative clinical recovery in patients with cervical myelopathy. Methods Between August 2015 and June 2017, we treated 40 patients with cervical myelopathy. Thirty patients met our prespecified inclusion criteria, all were invited to participate, and 28 patients opted to do so (93%; 14 men and 14 women; mean age: 67 years). The 28 patients and 28 age- and sex-matched controls underwent rs-fMRI (twice for patients with cervical myelopathy: before and 6 months after cervical decompression surgery). We analyzed the same study population that was used in our earlier study investigating functional connectivity. Controls had none of the following abnormalities: neck or arm pain, visual or auditory disorders, cognitive disorder, structural brain disorder, a history of brain surgery, mental and neurologic disorders, and medications for the central nervous system. We performed ALFF comparisons between preoperative patients with cervical myelopathy and controls, analyzed postoperative ALFF changes in patients with cervical myelopathy, and performed a correlation analysis between preoperative ALFF and clinical recovery in these patients. Clinical outcomes in the cervical myelopathy group were assessed using the 10-second test, the Japanese Orthopaedic Association upper-extremity motor (JOA-UEM) score, JOA upper-extremity sensory score (JOA-UES), and Japanese Orthopaedic Association Cervical Myelopathy Evaluation Questionnaire for upper-extremity function (JOACMEQ-UEF) score before and 6 months after surgery, which is when we believe these scores generally reach a plateau. A total of 93% of those enrolled (26 of 28 patients) were analyzed both preoperatively and postoperatively; the other two were lost to follow-up. Results The cervical myelopathy group had an increase in ALFF in the bilateral primary sensorimotor cortices (right, cluster size = 850 voxels, t-value = 6.10; left, cluster size = 370 voxels, t-value = 4.84) and left visual cortex (cluster size = 556 voxels, t-value = 4.21) compared with the control group. The cervical myelopathy group had a decrease in ALFF in the bilateral posterior supramarginal gyrus (right, cluster size = 222 voxels, t-value = 5. 09; left, cluster size = 436 voxels, t-value = 5.28). After surgery, the bilateral sensorimotor cortices (right, cluster size = 468 voxels, t-value = 6.74; left, cluster size = 167 voxels, t-value = 5.40) and left visual cortex (cluster size = 3748 voxels, t-value = 6.66) showed decreased ALFF compared with preoperative ALFF, indicating postoperative normalization of spontaneous brain activities in these regions. However, the bilateral posterior supramarginal gyrus did not show an increase in ALFF postoperatively, although ALFF in this region decreased preoperatively. Greater levels of ALFF at the left and right frontal pole and left pars opercularis of the inferior frontal gyrus before surgery in the cervical myelopathy group were correlated with larger improvements in the JOACMEQ-UEF score 6 months after surgery (r = 0.784; p < 0.001, r = 0.734; p < 0.001 and r = 0.770, respectively; p < 0.001). The prediction formula, based on preoperative ALFF values in the left frontal pole, was as follows: the predicted postoperative improvement in the JOACMEQ-UEF score = 34.6 x preoperative ALFF value - 7.0 (r(2)= 0.614; p < 0.001). Conclusions Our findings suggest that preoperative ALFF may be a biomarker for postoperative recovery in that it predicted postoperative JOACMEQ-UEF scores. To establish a comprehensive prediction measure for neurologic recovery in patients with cervical myelopathy, a multicenter study is underway.