Assessing ventricular size: is subjective evaluation accurate enough? New MRI-based normative standards for 19-year-olds

To create new standards for radiological indices of dilated ventricles and to compare these with subjectively assessed ventricular size. One hundred healthy controls (54 females), birth weight above 3,000 g, were followed throughout childhood as part of a longitudinal study of ex-prematures. All had a 3 Tesla brain magnetic resonance scan at age 17-20, and the following measurements were performed: biparietal and occipitofrontal diameters, width and depth of the frontal and occipital horns, diameter of the third ventricle and the frontal sub-arachnoid space. Ventricular size was judged subjectively by two neuroradiologists as being normal, or mildly, moderately or severely dilated. Head circumference was 31 mm higher for males than for females (95% confidence interval (CI) 25-28, p < 0.001). Similar, ventricular size except for the depth of the right frontal horn was larger for male; however, the observed differences were partly accounted for by the larger head circumference. Normative sex specific standards for different cerebral measurements were presented as mean and ranges and additional 2.5, 10, 50, 90, 97.5 percentiles. The mean depth of the left ventricle was larger than the right for males, with an observed difference of 0.6 mm in male (95% CI 0.2-0.9, p = 0.005). The mean width of the left ventricle was larger than the right for females, with an observed difference of 0.4 mm in male (95% CI 0.1-0.7, p = 0.018). Two subjects were judged to have moderately and 36 to have mildly dilated ventricles by observer one, while figures for observer two were one and 14. Overall, the two observers agreed on 15 having either mild or moderate dilatation (kappa 0.43). For both sexes, the mean depth of the frontal horns as well as of the larger occipital horns differed significantly between the no dilatation and the mild/moderate dilatation groups. In our unselected cohort of healthy 19-year-olds, a high total of 14% was diagnosed to have dilated cerebral ventricles when subjectively assessed by an experienced neuroradiologist, underscoring the need for our new normative standards. (orig.)