Exploring the Biological and Mechanical Properties of Abdominal Aortic Aneurysms Using USPIO MRI and Peak Tissue Stress: A Combined Clinical and Finite Element Study

This is a sub-study of the MRI in AAA to predict Rupture or Surgery (MA³RS) study (http://​www.​isrctn.​com/​ISRCTN76413758): a large multi-centre prospective observational cohort study aiming to determine the added value of USPIO-enhanced MRI in predicting AAA rupture or surgery in 342 patients with AAA under routine clinical surveillance. The MA³RS study protocol has been described in full elsewhere [16], and the study is in the follow-up stages, due to report in 2017. In brief, the MA³RS study cohort underwent USPIO MRI and CTA between November 2012 and December 2014 at the University of Edinburgh, Scotland. Patients were eligible for inclusion if they had an AAA ≥ 40 mm as measured on an ultrasound scan, if they were over 40 years of age, and with no contraindications to USPIO MRI or CTA. Patients were excluded if the suspected aetiology of the aneurysm was inflammatory. An outline of the patient selection criteria is given in Fig. 1a.

For the present sub-study, we selected and analysed scans of consecutive patients recruited into the MA³RS study using the process outlined in Fig. 1b. During this phase, 14 of the available consecutive scans were unable to be processed due to factors such as poor contrast on the CT scan (n = 6), segmentation difficulties (n = 4), and a lack of corresponding MRI-USPIO data (n = 4). These patients were therefore excluded from the final analysis. As per the selection algorithm (Fig. 1b), if a patient became ineligible, reconstruction and analysis moved on to the next consecutive patient until the predetermined target sample size of 50 eligible patients was reached. During the selection and reconstruction phase of the study, we were blinded to all demographic data, clinical outcomes, and USPIO groupings for each patient.

The USPIO MRI and image analysis techniques used in the study have previously been described [16]. Briefly, participants underwent T2 and T2*-weighted MRI scanning at 3 T (Siemens Magnetom Verio, Erlangen, Germany) both before and 24–36 h after administration of a weight-adjusted dose of USPIO. Both datasets (pre- and post-USPIO) were then registered and analysed, using bespoke software. USPIO causes a rapid decline in T2*, which is the constant for the decay of MRI signal intensity over time, and as such, changes in T2* can be used to assess USPIO accumulation in the AAA [8]. Colour maps representing the percentage change in T2* (%∆T2*) were generated using custom written scripts (MATLAB, The MathWorks Inc., Natick, MA). In order to minimise the effect of artefact, a threshold of ≥ 71% ∆T2* was used to identify areas of signal change attributable to true USPIO uptake, based on previous reproducibility data from our group [16].

Colour maps were classified according to predefined criteria [8], into USPIO-negative (no mural USPIO uptake) or USPIO-positive (area(s) of significant mural USPIO enhancement indicative of macrophage-driven inflammation). Significant mural USPIO enhancement was defined as 10 or more contiguous USPIO-positive voxels adjacent to the aneurysm wall. Periluminal USPIO uptake is not considered to represent true inflammation and likely represents passive trapping of USPIOs in periluminal thrombus [17] although mural and periluminal uptake can be difficult to distinguish from posterior wall uptake due to their close proximity.

A standard high-resolution contrast-enhanced CTA was performed with a slice thickness of 1.0 mm and pixel size of 0.625 mm (Aquilion One, Toshiba Medical Systems Ltd., UK). These images were used to create patient-specific models of each aneurysm. Segmentation and 3D reconstruction were performed using commercial software (VASCOPS GmbH, Sweden). Early studies of AAA often assumed a uniform wall thickness [18]; however, variable wall thickness has been shown to highly influence the predicted results [12, 19‐22]. Therefore, this package employs a specialist algorithm to calculate a more physiological aneurysm wall thickness distribution, which varies between 1.5 and 1.13 mm at the thrombus-free and covered sites, respectively [19]. Finite element (FE) meshes were then created from the 3D aneurysm geometry using the A4 clinical research software (VASCOPS). After suitable refinement, each AAA volume mesh typically consisted of > 160,000 (C3D8H) elements. These meshes were then exported to Abaqus 6.10-1 (Dassault Systemes, Simulia, Providence, RI, USA) for analysis. Both the aortic wall and thrombus regions were modelled as hyperelastic, homogeneous, incompressible, and isotropic materials, using well-established constitutive models [23, 24] with material constants based on population data. Loading representative of peak systolic blood pressure was applied as an outward-facing uniformly distributed pressure load acting on the luminal surface of the aneurysm. To remove any variability due to loading and to allow for comparison across patient cases, a peak systolic blood pressure of 120 mmHg (0.016 MPa) was chosen, as in many previous studies [25, 26]. In the present study, the effect of wall shear stress due to blood flow was not considered due to its negligible magnitude. Residual stresses in the aortic wall itself and the interaction of the aorta with the surrounding structures of the body (e.g. organs and spine) were also not considered. However, displacements at the distal and proximal most regions of each aneurysm were restrained, in all degrees of freedom, to model attachment of the AAA to the rest of the aorta.

For efficiency, a custom script was developed in Python (Python Software Foundation, Python Language Reference, version 2.7, available at http://​www.​python.​org) to automate the definition of the model parameters and batch process all 50 patients. All simulations were computed on a Dell Precision T7600 workstation with 16 cores and 64 GB of RAM. Contour plots of von Mises stress were output for all aneurysms, and their locations were manually aligned to USPIO uptake colour maps. Maximum AAA diameter, as measured orthogonal to the AAA centreline, was also extracted from each CT reconstruction.

In order to compare the spatial relationship between elevated stress and areas of inflammation represented by USPIO uptake, the two-dimensional (2D) contour maps of von Mises stress were manually co-aligned with the USPIO colour maps. The MRI slice with the largest area of USPIO uptake (i.e. most diseased segment) was chosen for analysis, ensuring that the corresponding cross-sectional slice was analysed from the 2D contour map (Fig. 2a). Regions of elevated stress and areas of mural USPIO enhancement were examined for visual overlap on the chosen slice.

Global comparisons between peak wall stress (PWS) predicted for each patient (from FE) and maximum and peak USPIO uptake (%∆T2*) per patient were also investigated, using values derived from the entire aneurysm. Maximum AAA diameter was also included in the analysis, as this is the most widely used predictor of aneurysm rupture. To test if correlations varied with classification of USPIO uptake (i.e. USPIO-negative or USPIO-positive), comparisons by group were also investigated.

To examine the trends with respect to the focal mural inflammation observed in USPIO-positive aneurysms more closely, the correlation of diameter and whole-vessel PWS with mural USPIO uptake values (mean and peak USPIO values identified on the most diseased segment) was investigated. Non-focal areas of USPIO uptake (i.e. those which did not meet the definition of mural USPIO enhancement) were removed prior to this analysis.

Linear dependence between variables was investigated using the Pearson’s correlation coefficient. Unpaired two-sample t tests were used to assess differences in mean values between AAA groups. Statistical significance was determined using a two-tailed p < 0.05. All statistical analyses were performed using Minitab® 17 (Minitab Ltd., Coventry CV 2TE, UK).

The maximum areas of PWS were most commonly (n = 28; 56%) found in the posterior wall and regions of high curvature such as the shoulder region. When analysing the most diseased segment, 40 aneurysms (80%) demonstrated some degree of visual overlap between regions of elevated stress and USPIO enhancement anywhere (e.g. adjacent to lumen and/or wall) in the vessel, while in the remaining 10 cases, no visual overlap was observed (e.g. Fig. 2b). Of the 40 aneurysms which did exhibit overlap, only 19 (38%) demonstrated spatial co-location of increased stress and USPIO enhancement adjacent to the aneurysm wall, while the remaining aneurysms exhibited overlap of elevated stress at areas of peri-luminal USPIO uptake where classification is challenging (e.g. Fig. 2c). Overall, only eight aneurysms (16%) demonstrated co-location of elevated stress with an area meeting the definition of mural USPIO enhancement (e.g. Fig. 2d).

Details of inter- and intra-observer variability for stress and USPIO comparisons as well as additional contour plot comparisons for the full 50 patients can be found in the supplementary text (ESM 1).

Maximum diameter was not associated with PWS (r = 0.13, p = 0.36) or maximum USPIO uptake (r = 0.05, p = 0.74). There was no correlation between PWS and maximum USPIO enhancement over the entire aneurysm (r = 0.17, p = 0.23), as shown in Fig. 3.

When comparing between groups, diameter was not correlated with PWS in either group (USPIO-negative r = 0.17, p = 0.39; USPIO-positive r = 0.13, p = 0.61), nor was diameter and peak USPIO uptake (USPIO-negative r = 0.04, p = 0.85; USPIO-positive r = −0.15, p = 0.52), as shown in Fig. 4. There was no difference in correlation between PWS and mural USPIO enhancement between groups (r = 0.23 vs. r = 0.09; p = 0.22).

There were no associations between PWS and mean or peak USPIO uptake within the individual areas of mural USPIO enhancement on the most diseased segment (r = 0.37, p = 0.10; r = 0.32, p = 0.16), as shown in Fig. 5a and b respectively.

Similarly, mean USPIO uptake within the identified regions of mural USPIO enhancement did not correlate with diameter (r = −0.05, p = 0.61), as shown in Fig. 6a. However, a significant weak inverse correlation was observed between peak USPIO uptake and diameter (r = −0.45, p = 0.04), as shown in Fig. 6b.