Objective: The purpose of this study was to examine the effect of aortoiliac tortuosity, as assessed by observers and 3-dimensional (3D) computer-based methods, on the conduct and outcome of endovascular repair of abdominal aortic aneurysms.
Methods: Infrarenal aortoiliac tortuosity was measured in 75 patients (mean follow-up, 14.8 +/- 10.4 months) who underwent endovascular repair of abdominal aortic aneurysms by using the following four methods: (1) grading by 2 experienced observers; (2) tortuosity index measured as the inverse radius of curvature (cm(-1)) at 1-mm intervals along the median luminal centerline (MLC) on 3D reconstructions of computed tomography (CT) angiograms and was calculated as the sum of values greater than 0.3 cm(-1); (3) MLC-straight line length ratio from renal to hypogastric arteries; (4) manual measurement of angles at points of angulation on anteroposterior and lateral projections of 3D CT reconstructions. In evaluating association between these measures, correlation between human observers was accepted as the gold standard.
Results: For rating of overall aortoiliac tortuosity, interobserver correlation (r = 0.67) was comparable with correlation of observers with tortuosity index (r = 0.67 and 0.56), whereas correlations of each observer with MLC-straight line ratio (r = 0.50 and 0.56) and cumulative angulation (r = 0.44 and 0.44) were significant but weaker. For determining the relative tortuosity of right and left aortoiliac access, agreement between observers and tortuosity index (54% and 58%; P < .05; kappa, 0.33 and 0.38) was not as good as between observers (68%; P < .001; kappa, 0.53). This difference was primarily related to evaluation of the aorta, where interobserver correlation (r = 0.71) was better than that between each observer and tortuosity index (r = 0.47 and 0.55), whereas correlations in the iliac arteries were comparable (r = 0.64 and 0.67) (all coefficients P < .01). Increased tortuosity was associated with a more complex endovascular repair, as reflected by longer fluoroscopy time (P = .05), use of more contrast material (P = .03), use of extender modules (P = .04), and more frequent use of arterial reconstruction (P = .01), but was not associated with a higher overall complication rate. Increased tortuosity, when it occurred in the aortic neck, was associated with predischarge endoleak (P = .03) but not with late endoleak, intervention, or aneurysm-related adverse events.
Conclusion: Aortoiliac tortuosity is associated with increased complexity of endovascular aneurysm repair and with predischarge endoleak but does not appear to affect intermediate-term results. Computer-based 3D measurement of aortoiliac tortuosity is feasible and clinically meaningful. Its ultimate role in relation to human assessment must be further defined in future studies.