Development and growth trends in angiotensin II-induced murine dissecting abdominal aortic aneurysms

Figure 1. Sample 2D and 3D ultrasound images and analysis. 2D B‐mode images shown with example linear measurements placed perpendicular to the vessel walls in healthy (A) and aneurysmal (B) mice. 2D M‐mode images show cursor placement and example systolic and diastolic measurement tracings in a healthy (C) and diseased (D) aorta. The dotted line cursor for PW images runs parallel to the aorta, and a sample velocity waveform tracing is shown in yellow (E). A 2D long‐axis B‐mode image of an aneurysm is shown on the left (F) with a volumetric representation of this aneurysm shown on the right (G). The blue surface is the overall AAA volume, the white surface is the true lumen volume, the red represents the open‐false lumen near the renal arteries, and the yellow is a closed lumen from a more cranial lesion. The hypoechoic region on the B‐mode image confirms the open‐false lumen volume depicted in (G; dotted red line). This volumetric acquisition further demonstrates the benefits of 3D imaging for complex aortic dissections with multiple lesions.

Figure 2. Sample histology taken at day 28 post‐AngII infusion matches well with US imaging data. 2D short‐axis US images near the proximal end of AAAs are shown with corresponding histology (from left to right: VVG, MTC and H&E stains at low 4× and high 10× magnification) in 5 different animals (A–E). VVG stain shows evidence of elastin degradation and aortic dissection (green arrowheads in B, C, and D) and MTC stain highlights adventitial collagen remodeling (blue staining in A–E). An open‐false lumen (OFL) is confirmed by histology for one lesion (D) and atherosclerotic plaque for three lesions (C, D, and E). All lesions (except E) show evidence by US and histology of leftward expansion. A small rightward expanding lesion immediately below the diaphragm (E) is identified proximal to a larger leftward‐expanding lesion in the same animal (see Figure 1G). Blue line = outer vessel wall; White line = inner vessel wall; Yellow line = outline of open‐false lumen; TL = true lumen; Red arrowhead = celiac artery. Scale bars: 1 mm (US), 500 μm (low magnification histology), and 200 μm (high magnification histology).

Figure 3. Baseline suprarenal aortic US measurements compared between groups. Both nondiseased (black triangle) and diseased (gray circles) are plotted from individual animals with means for each group shown as black circles (A–F). No significant differences were observed between groups except for body mass and initial mean diameter (*P ≤ 0.01). As expected, body mass was correlated with age for both the nondiseased and diseased groups (G). Example nondiseased (H) and aneurysmal (I) suprarenal aortas are shown at day 28 after sacrifice. Dashed cyan lines highlight the suprarenal abdominal aorta where AngII‐induced aneurysms form immediately proximal to the renal arteries. The aneurysmal aorta shows a dissection bulging to the animal's left. Scale bars: 5 mm. These plots represent the negligible differences between individual metrics for diseased and nondiseased animals, prompting the need for a statistical model to better delineate between the groups.

Figure 4. Aneurysm growth from volumetric analysis of the total AAA, open‐false lumen, and intramural thrombus. Volume data are shown for days 0 through 28 of total dissecting AAA volume/length (A) and also for thrombus volume (B) and open‐false lumen volume (C) within the aneurysmal false lumen. Each line represents an individual mouse. Day 0 represents measurements at baseline, prior to AngII infusion. Black lines show means ± SD. Initial (or diagnosis day) open‐false lumen and thrombus average volumes show a positive trend when plotted with respect to final dissecting AAA volume (D). These plots show how the composition of an aneurysm may influence the growth of a dissecting AAA over time.

Figure 5. Effects of branching arteries on dissecting AAA growth. Volume segmentations of mice with no branching arteries (A), only the celiac artery branching into the expanded region of the vessel (B), and both the celiac artery and SMA branching into the expanded region (C). The celiac (yellow), SMA (pink), and right renal artery (cyan) are shown in all images. Open‐false lumens are segmented in red for (A) and (B), whereas the false lumen of (B) was entirely open and thus not segmented. Transverse US images show the celiac (top) and SMA (bottom) intersecting the aneurysm. Scale bars: 2 mm. When separated by branching arteries, total, open‐false lumen, and thrombus volume/length, animals with only the celiac artery have the largest volumes and greatest growth (D). The overall change in volume/length from day of diagnosis to day 28 was larger for the celiac only dissecting AAAs compared to the celiac and SMA group (*P ≤ 0.05). These plots represent the need to include branching vessel information when studying factors that influence dissecting AAA growth over time.

Figure 6. Disease probability plots for baseline metrics. Baseline mass (A), aortic diameter (B), age (C), strain (D), and total volume/length (E) for diseased and nondiseased animals are plotted with the corresponding disease probability. These data show that one metric is not sufficient to predict future disease status.

Figure 7. Disease probability plots for baseline metrics. Baseline mass (A), aortic diameter (B), strain (C), and volume/length (D) for diseased and nondiseased animals are plotted with the corresponding disease probability. These data show that one metric is not sufficient to predict future disease status.

Figure 8. Numbers of mice predicted correctly in the diseased group, nondiseased group, and overall for the logistic regression model (A) and quadratic discriminant analysis model (B). These data highlight the ability of the prediction model to determine an animal's risk of developing a dissecting AAA or aortic rupture based only on baseline measurements.

Figure 9. Multiple linear regression models for predicting final aneurysm size. A full 3D model including open‐false lumen and thrombus volumes (A) is comparable to a 2D model including strain (B), and a 2D model excluding strain (C) with actual volume/length at day 28 plotted against predicted values. These highly correlated regressions show that measurements taken at diagnosis day can accurately predict day 28 aneurysm volume/length.