Incremental Contribution of Size Ratio as a Discriminant for Rupture Status in Cerebral Aneurysms

Abstract

Aneurysm size ratio (SR), variably defined as the ratio of dome height (H) or maximal dimension (D(max)) over average parent vessel diameter (PV) diameter, has been proposed as a promising aneurysm rupture status predictor. To evaluate the incremental contribution of SR to retrospective rupture status determination in a large high-resolution aneurysm database. Measurements were performed on catheter 3D-rotational angiographic volumetric datasets for 267 aneurysms (98 ruptured). SR was computed both as H/PV (SR1) and as D(max)/PV (SR2), and its discriminant performance was evaluated on the whole dataset, on aneurysm-type subsets (bifurcation [BIF] vs sidewall [SW]), and at specific aneurysm locations. Univariate and multivariate statistical analyses were performed by the use of area under the curve (AUC) of the receiver-operating characteristics. Neither SR1 nor SR2 were statistically correlated to rupture status in the BIF group, where only PV (AUC = 0.61) achieved significance. All parameters were statistically significant in the combined group, but with modest performance (AUC range, 0.62-0.74). SR1 (AUC = 0.84) and SR2 (AUC = 0.78) were strong predictors in the SW group, similar to H (AUC = 0.83) and D(max) (AUC = 0.77). Multivariate statistics failed to support SR as an incremental independent parameter from PV, D(max), and H. SR provides an uneven performance that depends strongly on the BIF/SW distribution of the data and is not useful for bifurcation lesions. In the SW subset, the incremental contribution of the SR over its H or D(max) individual component measurements could not be validated, suggesting prior findings of its utility to be the result of aneurysm-type selection bias.