Occipitocervical measurements: correlation and consistency between multi-positional magnetic resonance imaging and dynamic radiographs.

Abstract

To evaluate the reliability and validity of the multi-positional magnetic resonance imaging in measuring occipitocervical parameters using the standard cervical dynamic radiographs as a reference. Patients were included if they underwent both dynamic radiograph and cervical multi-positional MRI within a 2-week interval from January 2013 to December 2016. Twelve occipitocervical parameters were measured on both image modalities in all positions (neutral, flexion and extension): Posterior Atlanto-Dental Interval, Anterior Atlanto-Dental Interval (AADI), Dens-to-McRae distance, Dens-to-McGregor distance, Occipito-atlantal Cobb angle (C01 angle), Occipito-axis Cobb angle (C02 Cobb angle), Atlas-axis Cobb angle (C12 angle), Redlund-Johnell, Modified Ranawat, Clivus canal angle, Occiput inclination, and Occiput cervical distance. Pearson correlation and linear regression analysis were used to evaluate the correlation of both modalities for each parameter. A p value of < 0.05 was considered statistically significant. Cervical images of 70 patients were measured and analyzed. There was a significant positive correlation between dynamic X-ray and multi-positional MRI for all parameters (p < 0.05) except AADI. Dens-to-McGregor distance and Redlund-Johnell parameter demonstrated a very strong correlation in the neutral position (r = 0.72, r = 0.79 respectively) and moderate to very strong correlation(r > 0.4) for Modified Ranawat, Clivus canal angle, C02 Cobb angle and C02 distance in all neck position. The intra-class correlation (ICC) of intra- and inter-observer showed good to excellent reliability, and ICCs were 0.67-0.98. Multi-positional MRI can be a reliable imaging option for diagnosis of occipitocervical instability or basilar invagination compared to standard dynamic radiographs.