OBJECTIVES:To evaluate the utility of the splenial angle (SA), an axial angular index of lateral ventriculomegaly measured on diffusion tensor MRI color fractional anisotropy maps, in differentiating NPH from Alzheimer's disease (AD), Parkinson's disease (PD), and healthy controls (HC), and post-shunt changes in NPH, compared to Evans' index and callosal angle. METHODS:Evans' index, callosal angle, and SA were measured on brain MRI of 76 subjects comprising equal numbers of age- and sex-matched subjects from each cohort of NPH, AD, PD, and HC by two raters. Receiver operating characteristics (ROC) and multivariable analysis were used to assess the screening performance of each measure in differentiating and predicting NPH from non-NPH groups respectively. Temporal changes in the measures on 1-year follow-up MRI in 11 NPH patients (with or without ventriculoperitoneal shunting) were also assessed. RESULTS:Inter-rater and intra-rater reliability were excellent for all measurements (intraclass correlation coefficients > 0.9). Pairwise comparison showed that SA was statistically different between NPH and AD/PD/HC subjects (p < 0.0001). SA performed the best in predicting NPH, with an area under the ROC curve of > 0.98, and was the only measure left in the final model of the multivariable analysis. Significant (p < 0.01) change in SA was seen at follow-up MRI of NPH patients who were shunted compared to those who were not. CONCLUSIONS:The SA is readily measured on axial DTI color FA maps compared to the callosal angle and shows superior performance differentiating NPH from neurodegenerative disorders and sensitivity to ventricular changes in NPH after surgical intervention. KEY POINTS:• The splenial angle is a novel simple angular radiological index proposed for idiopathic normal pressure hydrocephalus, measured in the ubiquitous axial plane on DTI color fractional anisotropy maps. • The splenial angle quantitates the compression and stretching of the posterior callosal commissural fibers alongside the distended lateral ventricles in idiopathic normal pressure hydrocephalus (NPH) using tools readily accessible in clinical practice and shows excellent test-retest reliability. • Splenial angle outperforms Evans' index and callosal angle in predicting NPH from healthy, Parkinson's disease, and Alzheimer's disease subjects on ROC analysis with an area under the curve of > 0.98 and is sensitive to morphological ventricular changes in NPH patients after ventricular shunting.

The callosal angle (CA) is a useful biomarker in the diagnosis and management of idiopathic normal pressure hydrocephalus (NPH). Used incorrectly, CA measurements are variable, affecting its reliability as a clinical tool. Our objectives are to evaluate (i) reproducibility of established CA measurements between trained raters and (ii) impact of minor angular mal-rotations of the true coronal plane on CA measurements. CAs were measured by two independent raters on three-dimensional isovolumetric T1-weighted brain MRI of NPH patients and healthy controls using the established true coronal plane reformatted orthogonal to the plane containing the anterior-posterior commissural (AC-PC) line at the level of the posterior commissure. CA changes were subsequently evaluated when the coronal plane was mal-rotated by ± 5° and ± 10° in anterior-posterior and clockwise-anticlockwise directions. Inter-rater reliability of CA measurements was assessed using the intraclass correlation coefficient (ICC). On the true coronal plane, inter-rater ICC was excellent (0.973) for NPH patients and good (0.875) for controls. On mal-rotated coronal plane setups, ICC for CA was worse in controls (0.484–0.886) than NPH (0.879–0.981) groups and in clockwise-anticlockwise (0.484–0.956) than anterior-posterior (0.503–0.981) mal-rotations. CA changes secondary to mal-rotations from the true coronal plane were significant in NPH patients (P < 0.0001 to 0.0378) but not in controls (P > 0.1). This is the first demonstration of how small angular mal-rotations of the coronal plane used for CA measurement affect its value and inter-rater reliability, highlighting the importance of a standardized protocol when measuring the CA in NPH workup.