Interreader Reliability and Clinical Validity of a Magnetic Resonance Imaging Grading System for Cervical Foraminal Stenosis. Issue 6 (November 2017)

• Record Type: Journal Article
• Authors: Lee, Jung Eun; Park, Hee Jin; Lee, So Yeon; Lee, Yong Taek; Kim, Yong Bum; Lee, Kyu Hong; Shin, Hyunchul; Kwon, Young Joon
• Is Part Of: Journal of computer assisted tomography. Volume 41:Issue 6(2017)
• Publication Date: 2017-11
• Subjects: MR -- cervical foraminal stenosis -- neural foramen -- grade; Tomography -- Periodicals; Tomography -- Periodicals; Tomography; Periodicals; 616.0757
• DOI: 10.1097/RCT.0000000000000628
• Languages: English
• ISSNs: 0363-8715

Abstract

Abstract : Objective: Park system is a magnetic resonance imaging (MRI) grading system for cervical neural foraminal stenosis (CNFS) and consists of 4 grades (0–3) based on the foraminal shape of 45-degree T2-weighted oblique sagittal images. The objective of this study was to evaluate the interreader reliability of the Park system among radiologists, residents, and clinicians. This study also assessed the correlations between radiologic and clinical findings. Material and Methods: A total of 289 patients (men:women = 155:134, mean age = 50 years) who underwent oblique sagittal MRI of the cervical spine at our hospital were included. According to the MR grading system suggested by Park et al ( Br J Radiol 2013;86:20120515), 2 radiologists, 2 trainees, and 2 clinicians measured CNFS grade at the most narrow point. A neurosurgeon assessed the associated clinical manifestations. κ statistics were used to analyze the interreader agreement among the radiologists and clinicians. The clinical correlations between grade and positive clinical manifestations were assessed with R using nonparametric correlation analysis (Spearman correlation). Results: The overall interreader agreements between radiologists, between trainees, between clinicians, and between radiologists and clinicians were almost perfect (κ = 0.80–0.96). There were moderate correlations between grade and clinical manifestations in each group ( R = 0.562–0.669). There were moderate to relatively high correlations between grade and neurologic manifestations based on cervical level ( R = 0.570–0.715) (all P < 0.05). Conclusions: Regardless of reader experience, there was substantial to almost perfect interreader reliability with the Park system for CNFS based on oblique sagittal MRI.