Utilizing numerical simulations to prevent stent graft kinking during thoracic endovascular aortic repair

Numerical simulations of thoracic endovascular aortic repair (TEVAR) may be implemented in the preoperative workflow if credible and reliable. We present the application of a TEVAR simulation methodology to an 82-year-old woman with a penetrating atherosclerotic ulcer in the left hemiarch, that underwent a left common carotid artery to left subclavian artery bypass and consequent TEVAR in zone 2. During the intervention, kinking of the distal thoracic stent graft occurred and the simulation was able to reproduce this event. This report highlights the potential and reliability of TEVAR simulations to predict perioperative adverse events and short-term postoperative technical results.

Numerical simulations that virtually reproduce thoracic endovascular aortic repair (TEVAR) represent innovative computational adjuncts that may potentially aid the preprocedural planning phase in the future by predicting perioperative or short-term postoperative technical events and results. 1 Such tools could be further optimized regarding their credibility and reliability by providing evidence of their effectiveness and workflow, as illustrated in this study that applied a recently developed TEVAR simulation methodology to a patientspecific case with preoperative distal thoracic aortic stent graft kinking. The patient provided informed consent for the publication of this case report and related imaging.

CASE REPORT
An 82-year-old woman with hypertension and a history of heavy smoking (approximately 20 cigarettes per day) presented with a penetrating atherosclerotic ulcer (PAU) in the left hemiarch with maximum axial diameters of 38 Â 37 mm on computed tomography angiography (CTA). There were no further relevant cardiovascular diseases or interventions in her medical history. Additionally, an intraluminal floating thrombus located at the outer curvature of the proximal descending aorta (approximately 18 mm length, 15 Â 8 mm diameter) was identified on CTA, that seemed to be connected to the intraluminal thrombus of the PAU anteriorly (Fig 1, A). First  was a satisfactory overlap (Fig 2). In terms of quantitative assessment, the opening area at each nitinol stent ring (expressed as a percentage error between the simulation and CTA segmentation in square millimeters) remained <10%, with higher values in the region of the thrombus at the outer arch curvature (Fig 2).

DISCUSSION
The potential impact of applying this methodological TEVAR simulation model with Valiant Captivia to patient-specific cases could be of significant value during preprocedural planning given the ability to predict both perioperative adverse events (such as kinking for the illustrated case) and short-term postoperative technical results. As an additional tool, it may serve physicians in choosing the optimal proximal and/or distal sealing zones in specific cases with challenging aortic anatomy. In fact, as depicted in Fig 3, a simulation was performed to evaluate the stent graft apposition and kinking with a more distal landing zone. In this scenario, we noted that the distal kinking disappeared and that the third nitinol stent ring bulged into the PAU. However, this configuration might not be optimal, not only because of proximal landing zone reduction, but also because of the increased distance between the aorta and stent graft.
The time to obtain virtual results is compatible with the time needed to plan elective TEVAR; this procedure can be performed in 1 day. Furthermore, other patientspecific components such as the presence of intraluminal thrombus (as for this case) or calcifications, can be included in the simulation, to evaluate their impact on the stent graft deployment. Virtual deployment of different thoracic aortic stent grafts of different manufacturers, if verified and validated, 4 may also help to find the most suitable device with optimal sealing for different patient-specific aortic anatomies that vary according to geometrical characteristics (eg, diameter, length, angulation, and tortuosity).
Similar to the present case, another illustrative case by Derycke et al 5 has previously demonstrated the potential of a FE custom-made double branch Relay (Terumo Aortic, Sunrise, FL) TEVAR simulation for an aortic arch aneurysm, to reliably find stent graft collapse that led to postoperative complications. This TEVAR simulation found the deformation of the three nitinol stent rings at the same location as seen on postoperative CTA. Also in our case, the simulation was performed after the clinical procedure to verify if the numerical model was able to predict the perioperative stent graft kinking during the intervention. In our patient-specific case, the event was managed promptly by using a balloon without clinical and technical consequences.

CONCLUSIONS
This study further highlights the potential and reliability of TEVAR simulations to be adopted in and facilitate preprocedural planning in the future. For example, they could investigate optimal proximal landing and stent graft apposition or the ideal stent graft model in demanding aortic anatomies. One of the challenges before a wider implementation of such tools in daily clinical cardiovascular practice, remains the need to further