Literature Review of Automated Grading Systems Utilizing MRI for Neuroforaminal Stenosis

Abstract

Background: Cervical neural foraminal stenosis is a common and debilitating condition affecting people 40-60 years old. Although it is established that MRI is the best method of scanning the neural foramen, the question remains whether there is a role for three-dimensional MRIs and subsequently if it is possible to develop a computer aided automated grading system to establish the degree of clinically relevant cervical foraminal stenosis. Objective: The aim of the study is to review the literature for current or emerging automated grading systems of the cervical neural foramen, also including volumetric assessments of the neural foramen using MRI. Methods: A systematic search of Cochrane Library, Cochrane Clinical Trials, Ovid MEDLINE, EMBASE, CINAHL, ACM Digital Library and Institute of Electrical and Electronics Engineers (IEEE) and Web of Science was performed for reports examining automated systems and volumetric scanning foraminal stenosis published before 31.07.2021. Results: 3971 articles were identified with 8 included. The automated grading systems of the neural foramen focus largely of the lumbar spine with elements that may be applicable to the cervical spine. Although there are established studies for the automated grading of the lumbar spine, it is uncertain whether any of these are reproducible in the cervical spine. Visual grading systems for the cervical spine demonstrate good inter-reader reliability between radiologists and clinicians. Conclusion: The Park visual grading system although has limited data on the correlation with neurological symptoms or surgical outcome does demonstrate good inter-reader reliability between radiologists and clinicians. There is scope for further development of an automated grading system for cervical foraminal stenosis to improve the speed and consistency of image interpretation.