Fast virtual coiling algorithm for intracranial aneurysms using pre-shape path planning

To aid in predicting and improving treatment outcome of endovascular coiling of intracranial aneurysms, simulation of patient-specific coil deployment should be both accurate and fast. We developed a fast virtual coiling algorithm called Pre-shape Path Planning (P3). It captures the mechanical propensity of a released coil to restore its pre-shape for bending energy minimization, producing coils without unrealistic kinks and bends. A coil is discretized into finite-length segments and extruded from the delivery catheter segment-by-segment following a generic coil pre-shape. With the release of each segment, coil-wall and coil-coil collisions are detected and resolved. Modeling of each case took seconds to minutes. To test the algorithm, we evaluated its output against the literature, experiments, and patient angiograms. The periphery-to-core ratio of coils deployed by P3 decreased with increasing coil packing density, consistent with observations in the literature. Coils deployed by P3 compared well with in vitro experiments, free from unphysical kinks and loops that arose from previous virtual coiling algorithms. Simulations of coiling in four patient-specific aneurysms agreed well with the patient angiograms. To test the influence of coil pre-shape on P3, we performed hemodynamic simulations in aneurysms with coils deployed by P3 using the generic pre-shape, P3 using a coil-specific pre-shape, and full finite-element-method simulation. We found that the generic pre-shape was sufficient to produce results comparable to virtual coiling by finite element modeling. Based on these findings, P3 can rapidly simulate coiling in patient-specific aneurysms with good accuracy and is thus a potential candidate for clinical treatment planning.