Analysis and interrater reliability of pB-C2 using MRI and CT: data from the Park-Reeves Syringomyelia Research Consortium on behalf of the Pediatric Craniocervical Society.

Abstract

OBJECTIVE The distance to the ventral dura, perpendicular to the basion to C2 line (pB-C2), is commonly employed as a measure describing the anatomy of the craniovertebral junction. However, both the reliability among observers and the clinical utility of this measurement in the context of Chiari malformation Type I (CM-I) have been incompletely determined. METHODS Data were reviewed from the first 600 patients enrolled in the Park-Reeves Syringomyelia Research Consortium with CM-I and syringomyelia. Thirty-one cases were identified in which both CT and MRI studies were available for review. Three pediatric neurosurgeons independently determined pB-C2 values using common imaging sequences: MRI (T1-weighted and T2-weighted with and without the inclusion of retro-odontoid soft tissue) and CT. Values were compared and intraclass correlations were calculated among imaging modalities and observers. RESULTS Intraclass correlation of pB-C2 demonstrated strong agreement between observers (intraclass correlation coefficient [ICC] range 0.72-0.76). Measurement using T2-weighted MRI with the inclusion of retro-odontoid soft tissue showed no significant difference with measurement using T1-weighted MRI. Measurements using CT or T2-weighted MRI without retro-odontoid soft tissue differed by 1.6 mm (4.69 and 3.09 mm, respectively, p < 0.05) and were significantly shorter than those using the other 2 sequences. Conclusions pB-C2 can be measured reliably by multiple observers in the context of pediatric CM-I with syringomeyelia. Measurement using T2-weighted MRI excluding retro-odontoid soft tissue closely approximates the value obtained using CT, which may allow for the less frequent use of CT in this patient population. Measurement using T2-weighted MRI including retro-odontoid soft tissue or using T1-weighted MRI yields a more complete assessment of the extent of ventral brainstem compression, but its association with clinical outcomes requires further study.