PSI - Issue 49

6

Mongrain et al. / Structural Integrity Procedia 00 (2023) 000–000

S. McLennan et al. / Procedia Structural Integrity 49 (2023) 51–58

56

3.2 Modelling Assumptions To date there is limited consensus on how to define aortic calcification material in simulation studies. A number of prior simulation efforts have based calcification material constants on experimental data produced by Loree et al. (1994) however, as noted by Maier et al. (2010), this study gave values very close (and potentially less than given the error bounds) to the initial stiffness of AAA tissue as given by the parameters of Raghavan and Vorp (2000), which suggests the calcification is less stiff than the wall which seems intuitively wrong. As such, future work should characterize further the calcification material properties and the effect of the properties on the overall behaviour of aortic tissue. Several limitations exist with our simulation. Firstly, we neglected the viscoelastic behaviour of the aorto-iliac structure. While this may have an insignificant impact on relatively short analyses such as ours, longer timeframe analyses, such as EVAR post-operative complication prediction, would likely require aortic viscoelastic behaviour characterization. In addition, we have assumed a homogenous perivascular tissue medium, however, depending on the composition of perivascular tissues and the location along the vascular structure, the mechanical support brought by surrounding organs, para-spinal muscles and bowels varies along the aorto-iliac segment. to the computational cost of modelling this complexity, we neglected it in our simulations. Similarly to the present study, Li et al. (2008) treated the calcification component of their model as separate to the wall, with the mechanical properties of the wall characterized as healthy tissue. Meanwhile, Speelman et al. (2007) used modified material parameters for wall elements immediately surrounding calcification regions. From the existing studies it is not yet clear if further complexity in the healthy-calcified wall boundary modelling would play a significant role in aortic PWS, particularly for EVAR simulation, however, further study of this is important. Regarding our CT scan segmentation methodology, Buijs et al. (2018) noted that calcification scoring on CT angiography tends to overestimate volume and mass due to individual image voxels containing multiple tissue types but only being able to represent one. They also noted that accuracy is further reduced by interference of intravascular contrast. As such, our study was limited by our inability to include calcification volumes smaller than 1.5 mm 3 due to meshing restrictions. Further, as previously discussed, our selected lower threshold of 500 HU will have resulted in smaller calcifications going undetected during the segmentation process, likely leading to underestimated total calcification volume. In addition, it has been established that the higher the density of a calcification piece, the higher the X-ray absorption, Giannini et al. (2019). This suggests that our simulations neglected the soft calcification and only included the stiffer hydroxyapatite calcification by using the 500 HU lower threshold limit. Finally, while our simulations had a relatively short run time of approximately 1 hour 30 minutes per patient model, from a clinical sense this may be considered too time consuming. It should be noted that we were using a computer with relatively low processing power during our study. If the simulations were ran on a clinical grade computer, the run time would likelysignificantly reduce. When comparing the increase in PWS of the entire aorto-iliac structure between the No-Ca and With-Ca simulations, a positive correlation ( R = 0 . 794 , p < 0 . 001) was found between the RCP and the PWS increase (Fig. 4). This finding supports the notion that severe calcification presence can limit the feasibility of EVAR as higher PWS in our simulations suggests a higher likelihood of rupture due to tool deployment. 4 Conclusion In this paper, we developed a numerical simulation tool for assessing the impact of calcification presence during EVAR. FEA was used to study the PWS across the entire aorto-iliac structure during endovascular tool insertion for 12 patients. In addition, the PWS stress at different regions of the aorto-iliac structure was investigated for each patient. We quantified the impact of calcification presence by running simulations with and without consideration of calcification presence for each patient. Our numerical investigation shows that FEA can be used to guide clinicians on the feasibility of EVAR on patient-to- patient basis. In addition, our results suggest the iliac arteries experiencethe highest PWS of the defined aorto-iliac regions during