Expanding the Radiologist’s Arsenal against Abdominal Aortic Aneurysms, a Versatile Adversary

See also the article by Meyrignac et al in this issue.

Dimitrios Mitsouras, PhD, is associate professor of radiology and biomedical imaging at UCSF. He focuses on end-to-end novel cardiovascular imaging technique development fusing the physics and mathematics of signal encoding with high-performance computing and data-driven analyses. He has contributed to advances in high-efficiency 3D fast spin-echo MRI, multicontrast vessel wall MRI, and coronary CT angiography. He has developed numerous courses for the RSNA annual meeting and is recipient of a 2016 RSNA Magna Cum Laude education award.

Joseph R. Leach, MD, PhD, is assistant professor of radiology and biomedical Imaging at UCSF. His work aims to uncover imaging features and biomechanical conditions related to the progression of vascular disease. As an RSNA and NIH T32 research fellow, he has focused on the development of methodologies by which to study the biomechanics of abdominal aortic aneurysms on a patient-specific basis, incorporating advanced MRI techniques.

Abdominal aortic aneurysm (AAA), a dilatation of the abdominal aorta to a diameter larger than 3 cm, poses unique challenges and opportunities in vascular radiology. It is the only vessel-specific pathologic abnormality for which imaging-based screening is recommended by the U.S. Preventive Services Task Force (by one US examination in men age 65–75 years who have ever smoked, and selectively in those who have not). For patients with AAA who do not yet meet criteria for endovascular or surgical repair, imaging surveillance by US is recommended at varying intervals depending on AAA size and often performed on an annual basis. When repair thresholds are met, contrast agent−enhanced CT is the reference standard for aneurysm evaluation and procedure planning. Postrepair imaging at 1 and 12 months (initially with contrast-enhanced CT) is subsequently used to assess for endoleaks, aneurysm sac enlargement, and postrepair complications. Finally, because AAA repair is a durable therapy for most patients, imaging follow-up continues for a number of years. Thus, radiologists are involved at every stage of AAA management from screening to posttreatment, life-long surveillance.

AAA is diagnosed at an annual rate of four to seven per 1000 people and the cause of more than 7000 deaths per year in the United States. Smokers and older white men are at highest risk of developing an AAA, which affects up to 4%−8% of men older than 65 years. Nationwide screening programs in this population confer a mortality benefit but can lead to increased rates of elective repairs, not all of which may be necessary (1). Conversely, nonselective screening has not been proven to be effective in women and nonsmokers who have significantly lower (two- to sixfold) AAA prevalence. These groups nonetheless account for a significant proportion of AAA-related ruptures and deaths: Women accounted for 26% of Medicare beneficiaries with a diagnosis of ruptured AAA in 2006 and 41% of deaths related to AAA in the United States in 2005.

The AAA management strategy aims to avoid aneurysm rupture, which has mortality rates in excess of 80% overall and 50% for those who undergo repair. Maximal aortic diameter is the strongest known predictor of rupture. It is generally agreed that AAAs smaller than 4.0 cm in diameter are at low risk of rupture, whereas those 5.5 cm or larger should be repaired. Whereas earlier repair confers no mean survival benefit, a number of aneurysms will rapidly (and often unexpectedly) enlarge and become high risk. Some smaller aneurysms will rupture despite not meeting repair criteria. A recent study (2) reported that 41% of ruptures occurred in AAAs smaller than 5.4 cm diameter at last surveillance. Because reports suggesting that rates of failure to order a follow-up examination are not insignificant, improving surveillance follow-up rate via radiology-driven care coordination programs is one important avenue to further improve outcomes.

The role of the radiologist throughout AAA management is critical, with well-established guidelines that use our sophisticated scanners primarily as rulers (3,4). Maximal aneurysm diameter remains the primary data point on which AAA risk estimates and treatment decisions are based. The rate of change of maximal aneurysm diameter is a common ancillary metric that can prompt intervention (expansion > 1 cm per year) but is not much more sophisticated. This simplicity would be acceptable, if not preferable, were it not for the significant number of AAA-related deaths. The factors that promote AAA expansion or precipitate aneurysm rupture are complex and likely intersecting, and incompletely understood. Major advances in medicine over the past decades have been used to understand and predict the progression of the disease, from genetics to circulating biomarkers, and advanced imaging and imaging-based computational analyses (5).

Advanced imaging has targeted increased biologic activity, a characteristic of the disease that accompanies tissue degradation and loss of aneurysm wall strength. Early unsuccessful attempts have been supplanted by research directly assessing the inflammatory status of AAAs by using fluorine 18 sodium fluoride PET (5) and advanced MRI techniques by using T2* changes before versus after administration of ultrasmall superparamagnetic iron oxide contrast media (6). Evidence from larger, well-designed, albeit insufficiently powered, prospective studies shows promise in predicting AAA progression and the composite of repair and rupture events. However, further study is needed to understand the associations of those markers to AAA progression versus rupture, and whether indiscriminate application, for example to larger AAAs that are already considered higher risk for rupture, masks the significance of any findings (6).

Considering the mechanical aspect of the disease, rupture occurs when the stress experienced by the aneurysm wall because of blood pressure surpasses its strength. Wall stress is affected by both AAA geometry and wall tissue composition, and accordingly, also their evolution. To a first approximation, the law of Laplace states that a vessel with a larger radius (eg, progressing AAAs) must support a larger stress to withstand a given blood pressure. Complex computational modeling studies on the basis of finite element analyses by using the AAA anatomy at CT or MRI have established that wall stress depends on the shape of each AAA. Imaging-derived peak wall stress on the basis of such analyses has been successfully associated to rupture in smaller, typically retrospective case-control studies. However, this association is not unequivocal (7). Furthermore, the effect of peak wall stress on progression, its relationship with focal mural inflammation, and its value beyond maximal diameter remain uncertain.

Finally, factors affecting wall stress and strength are not independent. For example, wall stress is generally reduced by the presence and extent of intraluminal thrombus (ILT), which is seen in most clinically relevant AAAs. However, ILT has been independently associated with AAA progression (8). ILT likely distributes and reduces wall stress; however, histologic studies confirm it is biologically active and can promote an inflammatory environment that may reduce wall strength. Further complicating the dynamic balance of wall strength and stress, hemodynamics in the aneurysm sac, such as recirculating flow, may modulate ILT formation (eg, by increasing blood residence time near the endoluminal surface, thereby promoting platelet aggregation and adhesion). Ultimately, results from the imaging and image-based computational modeling literature regarding aneurysm anatomy, biologic activity, and biomechanics are often conflicting or of uncertain significance because of confounding biases, lack of accounting for factor interdependence, and often small sample sizes.

In this issue of Radiology, Meyrignac et al (9) combined volumetric assessment of AAA with image-derived hemodynamic assessment to predict progression in 81 patients with asymptomatic AAA who underwent contrast-enhanced CT at baseline and 1-year follow-up. Volumetric assessment considered aneurysm morphologic structure (saccular vs fusiform) and lumen, ILT, and total AAA volume. Hemodynamic assessment considered the average shear stress, that is, the frictional force exerted on the endoluminal surface by the flow of blood, and maximum pressure experienced by the wall throughout the cardiac cycle. Some relevant demographic factors (age, sex, hypertension, and anticoagulant use) were also controlled for to produce a logistic model predicting AAA volume growth larger than 10 mL in the first 50 patients (derivation group). The model, which included baseline lumen volume and shear stress normalized to that in the more proximal, nonaneurysmal aorta, was then tested in the remaining 31 patients (validation group). Model performance compared well against baseline maximum aneurysm diameter in the derivation cohort (area under the receiver operating characteristic curve [AUC], 0.78 vs 0.52, respectively). Although it compared less favorably in the validation cohort (AUC, 0.79 vs 0.71, respectively), it maintained overall diagnostic accuracy and good balance of sensitivity and specificity.

There are two aspects to this study that merit important discussion. First, it highlights the promise of combined assessment of factors believed to influence progression (eg, inflammatory pathways and wall mechanical stress) that others have begun to explore, or, in this study, morphologic structure and hemodynamics. Such combined interrogation will not only improve risk stratification, but also likely advance our understanding of the disease process. For example, in this study shear stress was more markedly decreased in the aneurysm sac for progressing AAA (21-mL growth) compared with stable AAA (3-mL growth), with most (80%) of that progression because of ILT growth. Elucidating such influences might, for example, contribute to our understanding of the failure of platelet inhibitors to modulate AAA growth (10).

Second, the study highlights that because of the rarity of rupture (<0.4% per year in men under surveillance), assessment of new risk stratification markers (and therapies [10]) are often limited to studying their association with progression during short intervals. Validating predictive model accuracy over a fixed time interval is in line with the increasing appreciation of both discontinuous growth of AAA and the temporal evolution of the many factors that we believe affect progression. However, volume measurements may be a more sensitive measure of AAA progression to assess such models over the 6–12-month period relevant to discontinuous growth. Standardized volume AAA measurements have generally been found to be equally reproducible as diameter (3). Moreover, on the basis of studies in both the pre- and postrepair surveillance periods, volume progression may not be reflected by maximum diameter progression with respect to measurement reproducibility in 27%–42% of AAA.

Given the capabilities of our scanners and postprocessing software, reducing our picture of an aneurysm by two dimensions (diameter only) is unnecessary. Diameter assessment alone may mask important features, particularly as we begin to address the spatial and temporal colocalization of multiple anatomic, biomechanical, and biologic aneurysm features.