18F]-Fluorodeoxyglucose–Positron Emission Tomography in Patients With Active Myelopathy

Abstract

To report and compare spinal cord [(18)F]-fluorodeoxyglucose-positron emission tomography (FDG-PET) metabolism in 51 patients with active myelopathy. We retrospectively identified patients from January 1, 2001, through December 31, 2011, with active myelopathy in whom FDG-PET was performed. Inclusion criteria were (1) intramedullary myelopathy, (2) neoplastic/inflammatory etiology, and (3) FDG-PET performed after myelopathy onset. Exclusion criteria were (1) extramedullary myelopathy, (2) radiation-associated myelopathy, (3) no pathological confirmation of neoplasm, and (4) inactive myelopathy. Diagnostic categories of nonsarcoid inflammatory, neoplastic, and neurosarcoid were based on their final myelopathic diagnosis. Two radiologists who independently assessed FDG-PET for spinal cord hypermetabolism and maximum standardized uptake value (SUVmax) were blinded to the underlying etiology. Fifty-one patients (53% women) with a median age of 60 years (range, 20-82 years) were included. Inflammatory myelopathic diagnoses (n=24) were as follows: paraneoplastic (n=13), autoimmune/other (n=5), inflammatory demyelinating (n=4), and transverse myelitis (n=2). Neoplastic diagnoses (n=21) were as follows: intramedullary metastases (n=12), intramedullary lymphoma/leukemia (n=7), and primary intramedullary neoplasm (n=2). Six patients had neurosarcoid myelopathy. Spinal cord hypermetabolism was more common with neoplastic myelopathy than with nonsarcoid inflammatory myelopathy (17 of 21 [81%] vs 6 of 24 [25%]; P<.001). Agreement between radiologist's assessments was excellent (κ=0.88). Median SUVmax was greater in neoplastic than in nonsarcoid inflammatory causes of myelopathy (3.3 g/mL vs 1.9 g/mL; P<.001). The FDG-PET hypermetabolism was seen in 3 of the 6 patients (50%) with neurosarcoid myelopathy (median SUVmax, 2.6 g/mL; range, 1.8-12.2 g/mL). Spinal cord FDG-PET hypermetabolism in patients with active myelopathy may be reliably detected and was more common in neoplastic than in inflammatory myelopathies in this study. Future investigation of spinal cord FDG-PET is indicated to assess its potential contributions in evaluating active myelopathies.