Association Between Fractal Dimension and Rupture Status in Intracranial Saccular Aneurysms: A CTA-Based Study

Abstract

Objective: To investigate the association between fractal dimension (FD) and rupture status in intracranial saccular aneurysms and to compare FD values between ruptured and unruptured aneurysms. Methods: This retrospective study included 145 patients with intracranial saccular aneurysms, comprising 72 ruptured and 73 unruptured cases. Aneurysms were segmented from computed tomography angiography images using semiautomated techniques. Two-dimensional (2D) fractal analysis was performed on binary and outline images generated from the segmented aneurysms using ImageJ with the FracLac plugin, and FD values were calculated using the box-counting method. Conventional morphologic parameters, including aneurysm size, aspect ratio (AR), and size ratio (SR), were also assessed. Group comparisons, multivariable logistic regression, and receiver operating characteristic (ROC) analyses were performed. Results: Binary FD values were significantly lower in ruptured than in unruptured aneurysms (1.742±0.018 vs. 1.753±0.016, P =0.015), whereas outline FD values showed no significant difference. In multivariable logistic regression analysis, standardized binary FD, AR, and SR were identified as independent predictors of aneurysm rupture. ROC analysis demonstrated modest discriminatory performance for binary FD, with an area under the curve of 0.620 (95% CI: 0.529-0.711). At an optimal cutoff value of 1.745, binary FD yielded a sensitivity of 65.8%, specificity of 47.2%, positive predictive value of 54.7%, and negative predictive value of 57.6%. Conclusion: Binary FD differed significantly between ruptured and unruptured intracranial saccular aneurysms and remained independently associated with rupture status after adjustment for conventional morphologic parameters. However, given the retrospective design and the modest discriminatory performance of FD, these findings should be interpreted as an association with rupture status rather than evidence of prospective rupture-risk prediction.