Assessment of interobserver agreement and use of selected magnetic resonance imaging variables for differentiation of acute noncompressive nucleus pulposus extrusion and ischemic myelopathy in dogs.

Abstract

OBJECTIVE To evaluate interobserver agreement for features used in presumptive diagnosis of acute noncompressive nucleus pulposus extrusion (ANNPE) or ischemic myelopathy by MRI, compare findings on postcontrast T1-weighted (T1W) MRI sequences with fat saturation (FS) for the 2 conditions, and determine whether length and directional patterns of hyperintensity of the intramedullary spinal cord on T2-weighted (T2W) fast spin echo (FSE) MRI sequences differ between dogs with these diseases. DESIGN Retrospective, observational study. ANIMALS 20 dogs with clinical signs compatible with ANNPE (n = 14) or ischemic myelopathy (6). PROCEDURES 3 observers evaluated MRI data (including T2W FSE, T2W single-shot FSE, and T1W FS sequences) for dogs with a presumptive diagnosis of ischemic myelopathy or ANNPE. Interobserver agreement for variables of interest including presumptive diagnosis was assessed by κ statistic calculations. Associations between diagnosis and variables of interest were assessed with Fisher exact or Cochran-Mantel-Haenszel tests. RESULTS Perfect interobserver agreement (κ = 1 for all comparisons) was found for the presumptive diagnosis of ischemic myelopathy versus ANNPE. Meningeal enhancement on postcontrast T1W FS MRI images and nonlongitudinal directional pattern of intramedullary hyperintensity on T2W FSE images were significantly associated with a diagnosis of ANNPE. Greater length of intramedullary hyperintensity was significantly associated with a diagnosis of ischemic myelopathy. CONCLUSIONS AND CLINICAL RELEVANCE Directional pattern and length of intramedullary hyperintensity on T2W FSE MRI images and enhancement patterns in postcontrast T1W FS sequences may provide important contributions to the criteria currently used in the presumptive diagnosis of ischemic myelopathy versus ANNPE.