The clinical correlation of a new practical MRI method for grading cervical neural foraminal stenosis based on oblique sagittal images.

Abstract

The purpose of this study was to determine whether each grade in a new system suggested by Park et al. (Park system) to assess cervical neural foraminal stenosis validly correlates with the associated clinical findings and to evaluate the interobserver agreement in grading between two MRI readers. We evaluated 166 patients (98 men and 68 women; mean age, 46 years) at our institution who underwent oblique sagittal MRI of the cervical spine. Using the new Park grading system, two radiologists evaluated the MRI findings for the presence and grade of cervical neural foraminal stenosis at the most narrow point. A neurosurgeon assessed the associated clinical manifestations. A positive neurologic manifestation of the cervical neural foraminal stenosis was defined as more than one positive neurologic clinical manifestation combined with more than one positive neurologic sign. Interobserver agreements between the two radiologists were analyzed using kappa statistics. Correlation coefficients (R) to assess the relationship between the grade and neurologic manifestations were calculated with nonparametric correlation analysis (Spearman correlation). The relationship between the assigned grade and the clinical manifestations was analyzed several ways: vertebrae level (C4-5, C5-6, or C6-7) and by age group (< 46 years and ≥ 46 years). Among patients who were evaluated by each reader to be grade 0, only 19 (17%) and 20 patients (18%) showed positive neurologic manifestations, respectively, with most patients showing negative neurologic manifestations. Among the patients who were grade 2 and 3, one reader found all patients and the second reader found all but one patient (100% and 93%, respectively) to have positive neurologic manifestations. According to the correlation coefficients, each Park grade was moderately correlated with the associated neurologic manifestations, such that higher grades were associated with more severe clinical manifestations. If we consider grade 2 or 3 MRI findings positive for identifying positive neurologic manifestations, the sensitivities and specificities were 39.7% and 99.0% (reader 1) and 39.7% and 99.0% (reader 2), respectively. The Park system, based on oblique sagittal MRI sections, provides a reliable and reproducible assessment of the severity of cervical neural foraminal stenosis. According to the Park system, grades 2 and 3 are associated with positive neurologic manifestations, and the Park system successfully predicts positive neurologic manifestations at these grades.