Background Proper sizing of intraarterial devices for aneurysm treatment is crucial to provide safety and ease of deployment while limiting torque, coverage of perforator branches and thrombogenicity. The behavior of an intraarterial device (wall apposition and foreshortening, for example) depends on several factors, including its original length, the relationship between the diameters of the device and recipient vessel as well as on the target vessel’s anatomy. Usually, the choice of device dimensions is made based on manual two- dimensional measurements from the 3D rotational angiography images and the operator’s individual experience. However, proper device dimensions and landing zonesare poorly predictable. A numerical computer-based simulation model (Sim and Cure; Grabels, France) has be shown to provide accurate and fast prediction of endovascular distention, wall apposition and final length of different sized devices based on 3D rotational angiography DICOM data.

Purpose The aim of this study was to evaluate whether use of a computer based simulation model results in selection of different device dimensions than the ones chosen by neurointerventionalists based on conventional methods.

Material and methods In a retrospective multi-center cohort study of 41 cases undergoing aneurysm treatment using the Pipeline Embolization Device (PED), device dimensions selected by experienced neurointerventionalists based on manual 2D measurements taken from rotational angiography were compared to PED dimensions calculated by the simulation model. Agreement between the different calculation methods wasevaluated by calculating Cohen’s Kappa.

Results Software based measurements resulted in different device dimension suggestions in 92.7% (38/41 cases). In 56% (23/41), a shorter length was suggested by the algorithm, in 20% (8/41) a longer length and in 24% the same length, whereas a shorter diameter was suggested in 37% (15/41), a longer diameter in 31.5% (13/41) and the same diameter in 31.5% (13/41). Agreement between conventional and computer based measurements was low (Cohen’s K=0.125 for length; K=0.239 for diameter, p<0.05).

Conclusions The low agreement between conventional and software based calculations confirms that the choice of proper device dimensions is challenging. Since the software based solution allows virtual simulation of multiple device sizes and prediction of their endovascular behavior easily within a few seconds, it potentially allows a decrease in procedure time and cost. Furthermore, it may remove uncertainty related to proper device sizing by accelerating the neurointerventionalist’s learning curve and confidence. This work is part of continuing evaluation of the simulation and its translation into clinical practice.

Disclosures K. Blackham: None. D. Zumofen: None. J. Ospel: None. V. Costalat: None. G. Gascou: None.