Peak wall stress and peak wall rupture index in ruptured and asymptomatic intact abdominal aortic aneurysms

Dear Editor Recently results of biomechanical analysis of AAA in 75 patients have been published in BJS 1. In this Correspondence we wish to bring to the readers’ attention certain technical aspects of the biomechanical modelling methodology that may shed new light on the results of this paper. Our parametric studies indicated an approximately linear relationship between measured/assumed wall thickness and computed peak wall stress (PWI, doi.org/10.1007/978-3-319-75589-2_5). This relationship implies that half-a-pixel (within a CT slice) size inaccuracy (∼0.5 mm) in AAA wall thickness estimation will immediately result in +/- 25 per cent inaccuracy in the PWI calculation. To visualise this uncertainty, each point in Fig. 2a1 reproduced below (Fig. 1), should be accompanied by a vertical error bar. Peak wall stress (PWS) versus AAA diameter 1 with error bars representing +/- 25 per cent uncertainty due to unknown patient specific AAA wall thickness, inserted for the lowest, highest and typical PWS Peak wall stress (PWS) versus AAA diameter 1 with error bars representing +/- 25 per cent uncertainty due to unknown patient specific AAA wall thickness, inserted for the lowest, highest and typical PWS Moreover, there are (at least) two other sources of uncertainty in patient-specific PWRI computation: i) inaccuracy of patient-specific AAA wall strength estimation based on population-based statistics (the magnitude of this uncertainty is unknown but potentially very large due to e.g., localised weakening of tissue); and ii) very significant impact of residual stress on computed PWS (doi10.5277/ABB-01143-2018-03). Authors of 1 used finite element meshes of about 7000 elements. This is orders of magnitude lower than required to obtain a reliable numerical solution for PWS. Others use ca. 200,000 elements, we use ca. 500 000. Our image data, segmentations and finite element meshes are freely available online https://doi.org/10.1016/j.jss.2020.01.025. We believe that clear understanding of biomechanical modelling and computational methods used in AAA analysis for large cohorts is necessary to correctly evaluate the conclusions of such studies. Disclosure. The authors declare no conflicts of interest.