Fourth ventricle roof angle as a measure of fourth ventricle bowing and a radiographic predictor of brainstem dysfunction in Chiari malformation type I

Abstract

OBJECTIVE Chiari malformation type I (CM-I) is a congenital and developmental abnormality that results in tonsillar descent 5 mm below the foramen magnum. However, this cutoff value has poor specificity as a predictor of clinical severity. Therefore, the authors sought to identify a novel radiographic marker predictive of clinical severity to assist in the management of patients with CM-I. METHODS The authors retrospectively reviewed 102 symptomatic CM-I (sCM-I) patients and compared them to 60 age-matched normal healthy controls and 30 asymptomatic CM-I (aCM-I) patients. The authors used the fourth ventricle roof angle (FVRA) to identify fourth ventricle “bowing,” a configuration change suggestive of fourth ventricle outlet obstruction, and compared these results across all three cohorts. A receiver operating characteristic (ROC) curve was used to identify a predictive cutoff for brainstem dysfunction. Binary logistic regression was used to determine whether bowing of the fourth ventricle was more predictive of brainstem dysfunction than tonsillar descent, clival canal angle, or obex position in aCM-I and sCM-I patients. RESULTS The FVRA had excellent interrater reliability (intraclass correlation 0.930, 95% CI 0.905–0.949, Spearman r2 = 0.766, p < 0.0001). The FVRA was significantly greater in the sCM-I group than the aCM-I and healthy control groups (59.3° vs 41.8° vs 45.2°, p < 0.0001). No difference was observed between aCM-I patients and healthy controls (p = 0.347). ROC analysis indicated that an FVRA of 65° had a specificity of 93% and a sensitivity of 50%, with a positive predictive value of 76% for brainstem dysfunction. FVRA > 65° was more predictive of brainstem dysfunction (OR 5.058, 95% CI 1.845–13.865, p = 0.002) than tonsillar herniation > 10 mm (OR 2.564, 95% CI 1.050–6.258, p = 0.039), although increasing age was also associated with brainstem dysfunction (OR 1.045, 95% CI 1.011–1.080, p = 0.009). A clival canal angle < 140° (p = 0.793) and obex below the foramen magnum (p = 0.563) had no association with brainstem dysfunction. CONCLUSIONS The authors identified a novel radiographic measure, the FVRA, that can be used to assess fourth ventricular bowing in CM-I and is more predictive of brainstem dysfunction than tonsillar herniation. The FVRA is easy to measure, has excellent interrater variability, and can be a reliable universal radiographic measure. The FVRA will be useful in further describing CM-I radiographically and clinically by identifying patients more likely to be symptomatic as a result of brainstem dysfunction.