Quantification of Intracranial Aneurym Shape Irregularity

Introduction: Irregular shape of intracranial aneurysms (IAs) has shown a strong correlation with rupture status, however subjective methodology has been historically utilized to define an aneurysm as irregular. Conceptually, a holistic 3D measure of an aneurysm’s shape could provide a quantification of irregularity. The objective of our study was to assess correlation of various 3D morphometric measurements of IAs with a consensus definition of irregular shape.

Methods: IAs were defined as irregular if they exhibited blebs, daughter sacs, or gross surface irregularity. A consensus agreement of irregular shape for each IA was reached through a multiphase process between three independent reviewers with agreement assessed via intraclass correlation coefficient (ICC). Multivariate regression was performed to assess for correlation of 3D morphometric features with irregular shape. For 3D shape measurements found to be significantly associated with irregular shape, the optimal cut-point for each measurement was determined through binary regression.

Results: The dataset consisted of 65 individual IAs each with 58 clinical and radiographic features for analysis. Anterior (77%) and posterior (23%) circulation aneurysms were represented with a median size of 5.23 mm (3.87, 7.74). Twenty-six (40%) IAs were classified as irregular following consensus classification, with improvement in ICC between reviewers pre- (0.73) and post-training (0.82). A lower fractal dimension (p = 0.003) and surface area to volume ratio (SA:V) (p = 0.033), as well as a higher undulation index (p = 0.010) were independently associated with irregular shape. Optimal cut-point analysis for each measure showed that SA:V was the most sensitive (96.1%) and fractal dimension the most specific (73.5%) for irregular shape. Undulation index yielded the highest AUC (0.78) with a balanced sensitivity (87.1%) and specificity (61.7%).

Conclusion : 3D morphometric measures, specifically fractal dimension, undulation index, and surface area to volume ratio, show promise at defining and quantifying irregular shape of IAs.