Skip to content

Hemodynamics and Biomechanics of Thoracic Aortic Aneurysm

TL;DR — Dissection and rupture occur when local wall stress exceeds local wall strength; diameter is a surrogate for only the first half of that inequality and a poor one. Aneurysmal ascending aorta is simultaneously stiffer and weaker than control aorta — tensile strength falls 29–34% while maximal stiffness rises 44–72% (Vorp 2003, PMID 12683565) — so a Laplace-based estimate of stress says nothing about the margin to failure. Patient-specific finite-element analysis finds peak wall stress spanning 28–94% of measured failure stress across individuals, with rupture indices only weakly correlated with diameter (ρ ≈ −0.29) (Trabelsi 2015, PMID 25979384), and in bicuspid-valve aneurysms peak stress does not correlate with diameter at all (r ≈ 0.007) (Xuan 2018, PMID 29656820). On the flow side, 4D-flow MRI shows that bicuspid valves generate eccentric systolic jets with peak wall shear stress ~0.9 ± 0.3 N/m² versus 0.4 ± 0.3 N/m² in age- and size-matched controls (Barker 2012, PMID 22730420), and — the pivotal observation — within-patient regions of elevated WSS show measurably greater elastin degradation and ECM dysregulation than adjacent normal-WSS regions of the same aorta (Guzzardi 2015, PMID 26293758). The unsolved problem is that none of these biomechanical metrics has yet been prospectively validated against dissection or rupture endpoints.

1. Wall stress fundamentals and the limits of Laplace

The Laplace relation for a thin-walled cylinder, σ ≈ P·r/h (circumferential stress = pressure × radius / wall thickness), is the intuition behind every diameter threshold in risk stratification. It fails for the diseased ascending aorta in at least five ways:

Laplace assumption Reality in ascending TAA Evidence
Uniform wall thickness Thickness varies by region: 2.21 ± 0.4 mm on the outer curve vs 2.50 ± 0.12 mm on the inner curve Salmasi 2022, PMID 35894942
Cylindrical, axisymmetric geometry Root, sinotubular junction and tubular ascending carry markedly different stresses; peak circumferential stress is highest at the STJ (501 ± 119 kPa) not the widest segment (AscAo 340 ± 57.6 kPa) Gomez 2022, PMID 34718581
Isotropic material Aorta is anisotropic: energy loss differs significantly between circumferential and longitudinal directions (p<0.0001) Chung 2016, PMID 27888778
Stress alone determines failure Failure requires stress > strength, and strength falls independently of geometry Vorp 2003, PMID 12683565
No residual/prestress Patient-specific FEA must explicitly account for prestress geometry before loading to systemic pressure PMID 29656820; PMID 34718581

The material consequence of medial degeneration is a stiffening-and-weakening pair. In ascending aneurysm tissue versus autopsy control aorta, tensile strength was 29% lower longitudinally and 34% lower circumferentially, while maximum tissue stiffness was 72% higher longitudinally and 44% higher circumferentially (PMID 12683565). A stiffer wall transmits more load to its cellular and matrix constituents at the same pressure — and a stiffer, weaker wall shortens the distance between physiologic stress and failure stress without necessarily changing diameter.

Wall stress is not, however, a sufficient explanation for aneurysm formation. Comparative biaxial testing across 10 murine models (wild-type, acute elastase, and eight genetic models spanning ECM proteins, transmembrane receptors, cytoskeletal proteins and intracellular signalling) showed that reduced elastic energy storage or distensibility did not predispose to aneurysm. Aneurysm-prone aortas were characterised by inability of intramural cells to maintain or restore intrinsic circumferential material stiffness, despite normal or lower-than-normal circumferential and axial wall stresses — pointing at dysfunctional mechanosensing/mechanoregulation rather than at overload (Bellini 2017, PMID 28490606).

2. Wall stress versus wall strength

The clinically operative statement is that dissection or rupture occurs when local wall stress exceeds local wall strength (PMID 29656820). Diameter proxies the numerator only. Three lines of evidence show the denominator moves independently:

  • Strength is regionally patterned. Peel (delamination) force averaged 35.5 N/m (SD 22) and dissection energy 88.5 J/m² (SD 69), both lowest at the outer curve in the longitudinal orientation, implying the outer curve is more prone to dissection propagation while possibly being less prone to rupture than the inner curve. Circumferential tensile strength exceeded longitudinal by anatomical region. Collagen abundance correlated with both circumferential and longitudinal strength (p=0.010); SMC count correlated with no mechanical property (p>0.05) (Salmasi 2022, PMID 35894942).
  • Strength has an identifiable microstructural basis. A fibre-bridge failure model validated against peel tests on human aortic strips showed that the number density and failure energy of radially-running collagen fibres control peel strength, and that the lower delamination strength observed in the circumferential direction reflects fewer radially-running collagen fibres in that direction (Pal 2014, PMID 24484644).
  • Stress and strength are only loosely coupled to size. Across five patients whose resected tissue underwent bulge-inflation testing to failure and whose CT-derived geometry was analysed by FEA, peak wall stress ranged from 28% to 94% of that patient's own failure stress; both the rupture risk index and the overpressure index correlated only weakly with maximum diameter (ρ = −0.29 for each) (Trabelsi 2015, PMID 25979384).

A fourth, mechanistic contributor is the mucoid material described in pathophysiology. Pools of glycosaminoglycans are proposed to decrease tensile strength, create stress concentrations, and raise intralamellar swelling pressure, any of which could initiate a local delamination that then propagates as a dissection producing a false lumen or rupture (Humphrey 2012, PMID 23018968).

3. Ex vivo biomechanical testing of aneurysmal tissue

Metric What it measures Representative findings Citation
Uniaxial tensile strength / stiffness Load to failure; maximal tangent modulus ATAA strength ↓29% (long.), ↓34% (circ.); stiffness ↑72% / ↑44% vs control PMID 12683565
Biaxial tensile testing Direction-dependent stress–strain behaviour under physiologic biaxial loading Basis for energy-loss and anisotropy metrics; used across 10 mouse models PMID 25129601; PMID 28490606
Energy loss (hysteresis area of the loading/unloading stress–strain loop) Fraction of cardiac-cycle energy absorbed rather than returned by the wall Aneurysms had higher energy loss than controls (p<0.0001); energy loss correlated with aortic size (p<0.0001, r²=0.60) with a hinge point at 5.5 cm after which it rises rapidly; indexing size to BSA linearised the relationship (r²=0.78); correlated with collagen/elastin ratio (p=0.0002, r²=0.49) Chung 2014, PMID 25129601
Energy-loss anisotropy index Directional dependence of energy loss Energy loss greater circumferentially than longitudinally (p<0.0001). Anisotropy diminished in larger aortas (r=0.15, p=0.01; indexed to BSA r=0.29, p=0.002), in aortas with higher overall energy loss (r=0.44, p<0.0001), with tricuspid valves (p=0.004), and with higher collagen:elastin (r=0.29, p=0.001). Aortas with collagen:elastin > 2 were uniformly isotropic Chung 2016, PMID 27888778
Peel / delamination testing Force and energy to separate medial lamellae — the mechanics of dissection propagation, not rupture Peel force 35.5 N/m; dissection energy 88.5 J/m²; lowest at outer curve longitudinally PMID 35894942
Bulge inflation to failure Patient-specific failure stress, rupture stretch, physiologic tangent modulus Enables per-patient rupture risk index; peak stress 28–94% of failure stress PMID 25979384; PMID 31331875

The energy-loss work is notable for two reasons. First, its hinge point at 5.5 cm mirrors the clinical size hinge points used for surgical thresholds — a rare direct convergence of tissue mechanics with epidemiology (PMID 25129601). Second, loss of anisotropy is a mechanical readout of histology: as the collagen/elastin ratio rises and medial degeneration advances, the aorta becomes mechanically isotropic, so a single global biomechanical measurement carries microstructural information (PMID 27888778).

4. 4D-flow MRI and wall shear stress

Time-resolved 3D phase-contrast MRI with three-directional velocity encoding ("4D flow") maps velocity throughout the thoracic aorta and permits estimation of wall shear stress — the tangential frictional force of blood on endothelium, distinct from the transmural wall stress of §1–2.

Bicuspid valves produce eccentric jets and elevated regional WSS. In 60 subjects, ascending aortic WSS in right–left cusp fusion BAV was significantly elevated versus age-appropriate and age/size-matched tricuspid controls, with peak WSS 0.9 ± 0.3 N/m² versus 0.4 ± 0.3 N/m² in age/aorta-size-controlled subjects (p<0.001), independent of stenosis severity. Co-registering cine valve images with 4D-flow data directly linked cusp fusion pattern to jet pattern: right–left fusion produced right-anterior wall impingement corresponding to the statistically elevated WSS regions, while right–non-coronary fusion produced a different jetting pattern (Barker 2012, PMID 22730420).

Elevated WSS co-localises with ECM degradation in the same patient. Twenty BAV patients undergoing ascending resection had preoperative 4D-flow WSS mapping, and paired within-patient wall samples were taken from elevated-WSS and normal-WSS regions. Elevated-WSS regions showed decreased total elastin (p=0.01), thinner elastic fibres (p=0.00007) and fibres further apart (p=0.001), plus increased TGF-β1 (p=0.04), MMP-1 (p=0.03), MMP-2 (p=0.06), MMP-3 (p=0.02) and TIMP-1 (p=0.04) (Guzzardi 2015, PMID 26293758). This paired design controls for genotype, age, blood pressure and every other patient-level confounder, and is the strongest available evidence that valve-related haemodynamics mediate rather than merely accompany bicuspid aortopathy.

The correlation extends across a larger cohort and is strongest early. In 27 BAV patients with 93 tissue samples, elastic fibres were thinner and WSS higher along the greater curvature (vs anterior wall p=0.003 and p=0.0001; vs lesser curvature both p=0.001). Increased regional WSS was associated with decreased elastic fibre thickness (r=−0.25, p=0.02), and the association strengthened in patients with aortic stenosis (r=−0.36, p=0.002) and in smaller aortas (<4.5 cm: r=−0.39, p=0.03) — i.e. the haemodynamic–histologic link is most detectable at earlier stages of aortopathy. Elastic fibre thinning was in turn associated with circumferential stiffness (r=−0.41, p=0.06) (Bollache 2018, PMID 30060930).

Methodological caveat — age-matching is mandatory. Peak systolic velocity and WSS decline significantly with age in healthy controls (R² = 0.32 and 0.39 respectively, p<0.001). Comparing a young BAV cohort with older controls overestimated the aortic fraction exposed to abnormally elevated velocity/WSS (~25 ± 14% vs ~8 ± 5% when correctly age-matched); comparing an older BAV cohort with younger controls overestimated abnormally decreased velocity/WSS (~9 ± 7% vs 1 ± 1%) (van Ooij 2016, PMID 26477691). Any WSS "heat map" is only as valid as its normative reference.

Tricuspid valve dysfunction also deranges flow. Among 86 patients with ascending diameter ≥40 mm and trileaflet valves plus 25 healthy controls, all patient groups showed markedly elevated vortex and helix flow versus controls; WSS was significantly elevated in combined stenosis+regurgitation and in stenosis, and significantly reduced in patients with neither stenosis nor regurgitation (Suwa 2019, PMID 31169969). Eccentric flow is therefore not a bicuspid-only phenomenon — see bicuspid aortopathy.

Beyond WSS. 4D-flow can also derive intra-cross-sectional pressure differences. In 17 ATAA patients, 17 age/sex-matched and 13 younger healthy subjects, maximum WSS correlated inversely with indexed ascending diameter (r=−0.49, p<0.001) and indexed volume (r=−0.63, p<0.001). Within the 30 healthy subjects, cross-sectional pressure difference was positively associated with vorticity amplitude (r=0.55, p=0.002) and with maximum WSS (r=0.59, p<0.001), and remained so after adjustment for diameter, age and systolic pressure; ATAA patients were then superimposed on those normal-ageing trends to identify individuals with disproportionately high pressure differences (Bouaou 2024, PMID 38403074). Note the sign: WSS falls as the aorta dilates, so a single WSS value cannot be read as "risk" without knowing size and stage.

5. Computational modelling: FEA, CFD and FSI

5.1 Finite element analysis of wall stress

Patient-specific FEA reconstructs 3D geometry from ECG-gated CT angiography, accounts for the prestressed in vivo configuration, and loads the model to systemic pressure with a fibre-embedded material law.

Study Cohort Result
Xuan 2018 (PMID 29656820) BAV-ATAA n=17 vs TAV-ATAA n=19, all >4.5 cm 99th-percentile longitudinal stress 280 vs 242 kPa (p=0.028); at the STJ 405 vs 329 kPa (p=0.023); circumferential 548 vs 462 kPa (p=0.033). Stress did not correlate with diameter in BAV (r=0.007 circ., −0.004 long.) but did in TAV (r=0.677)
Gomez 2022 (PMID 34718581) TAV-associated aneurysms, n=204, diameter ≥4.0 cm Peak longitudinal stress: sinuses 326 ± 61.7, STJ 246 ± 63.4, ascending 195 ± 38.7 kPa. Peak circumferential: sinuses 416 ± 85.1, STJ 501 ± 119, ascending 340 ± 57.6 kPa; ascending differed significantly from both sinuses and STJ (p<0.001)
Trabelsi 2015 (PMID 25979384) n=5 with paired bulge-inflation failure testing Peak stress 28–94% of that patient's failure stress; rupture risk index and overpressure index weakly correlated with diameter (ρ=−0.29)

The BAV result is the most consequential for practice: if wall stress carries dissection risk and wall stress is uncorrelated with BAV aneurysm diameter, then diameter-based thresholds are structurally unable to stratify that population (PMID 29656820).

5.2 Coupled flow modelling

Fluid–structure interaction simulation isolates the valve's contribution from confounded patient variables. Simulating four fully-passive elastic valve models (tricuspid; and left–right, right–non, non–left cusp fusion bicuspid) in a single patient-specific healthy aortic geometry produced relatively uniform tricuspid flow with little secondary or reverse flow and essentially no transvalvular pressure gradient, versus localised forward jets, excess streamwise momentum, elevated secondary and reverse flow and clinically significant stenosis for every bicuspid configuration. Locations of high local flow corresponded to the locations at which dilation is observed clinically, with the location depending on which cusps were fused — supporting chronic local flow exposure as a contributor to localised dilation (Kaiser 2022, PMID 35748961).

Combining in vivo, in vitro and in silico data in the same ten patients — CT and 4D-flow-driven CFD, then bulge-inflation testing of the resected tissue — found a highly significant positive correlation between time-averaged wall shear stress and rupture stretch (r=0.867, p=0.001), i.e. relatively low TAWSS associated with reduced rupture properties of the tissue (Condemi 2019, PMID 31331875). This is a different, and partly opposing, signal to the high-WSS/ECM-degradation association of §4, and is discussed as an open question below.

6. Why diameter is a crude surrogate — and what biomechanics offers

Diameter is retained because it is measurable, reproducible enough, and tied to outcome data. Its mechanistic weaknesses are now well characterised:

  1. It estimates one term (stress) of a two-term failure criterion and ignores strength, which varies by region, direction and microstructure (PMID 35894942; PMID 24484644).
  2. In BAV aneurysms it does not track computed wall stress at all (PMID 29656820).
  3. Peak stress occurs at the sinuses and sinotubular junction, not at the maximal-diameter tubular segment (PMID 34718581).
  4. The margin between operating stress and failure stress varies roughly threefold between patients at similar size (PMID 25979384).
  5. It ignores flow: two aortas of identical size with different valve morphology experience different jet impingement, WSS and regional ECM remodelling (PMID 22730420; PMID 26293758; PMID 35748961).

What biomechanics adds, in rough order of maturity: (i) energy loss as a tissue-level severity marker that tracks histology and shows a 5.5 cm hinge point (PMID 25129601; PMID 27888778); (ii) regional WSS maps as a non-invasive read-out of where ECM degradation is occurring, potentially permitting individualised resection extent (PMID 26293758; PMID 30060930); (iii) patient-specific stress and rupture indices from FEA plus material characterisation (PMID 25979384; PMID 31331875). All three remain research tools; see risk stratification and size thresholds for what is actually used clinically and imaging and surveillance for acquisition considerations.

The mechanobiological interpretation is that the aortic wall normally holds stress and stiffness near set points through negative feedback, and that disease represents replacement of those loops by biomechanical and biochemical positive feedback — stiffening begets altered mechanosensing begets maladaptive remodelling (Humphrey 2021, PMID 34255994; Humphrey 2015, PMID 25858068). On this reading, biomechanics is not merely a better risk calculator; it is the phenotype.

7. Pressure as a mechanical second hit

Blood pressure enters the failure criterion twice — directly, as the P term setting wall stress, and indirectly, as a chronic remodelling stimulus. Thoracic aortic aneurysm and dissection are associated with poorly controlled hypertension as well as with mutations in genes for ECM constituents, membrane receptors, contractile proteins and associated signalling molecules — a grouping that motivated the mechanosensing/mechanoregulation framework rather than a purely load-based one (Humphrey 2015, PMID 25858068).

The cleanest experimental demonstration that pressure acts as a second hit rather than a primary cause comes from contractile-mutant mice. Myh11^R247C/R247C aortas exhibited near-normal biaxial mechanics under normotension and near-normal adaptation to induced hypertension on average — yet induced hypertension produced intramural delaminations or premature death in over 20% of these animals, with localised pools of mucoid material at the delamination sites mirroring the histology of human TAAD. The authors' explicit conclusion was that contractile-protein mutations place the thoracic aorta at increased risk from epigenetic factors, and that the field should look for focal rather than global changes in structure and properties — including glycosaminoglycan/proteoglycan pooling (Bellini 2015, PMID 25433566). A global average biomechanical measurement can therefore be reassuring in an aorta that is locally about to fail.

Pressure also leaves a measurable structural signature. In the regional testing series, advanced patient age and higher externally measured pulse wave velocity were predictors of increased aortic wall thickness (PMID 35894942) — meaning wall thickness, the h term in the Laplace denominator, is itself a function of chronic haemodynamic exposure rather than an independent constant. Integrating these multimodal signals — genetics, histology, regional mechanics, flow — is the stated motivation for data-informed and data-driven computational models of aortic disease progression (Humphrey 2026, PMID 42140666).

Open questions

  • Is high or low wall shear stress the dangerous one? Elevated regional WSS co-localises with elastin degradation and MMP upregulation within the same aorta (PMID 26293758; PMID 30060930), yet lower time-averaged WSS correlated with reduced tissue rupture properties in a combined in vivo/in vitro/in silico cohort (PMID 31331875), and maximum WSS falls as the aorta dilates (PMID 38403074). Whether these describe different disease stages, different mechanisms (initiation vs failure), or a measurement artefact is unresolved.
  • Has any biomechanical metric been prospectively validated against dissection? All current rupture indices are retrospective and derived from resected tissue at elective surgery (PMID 25979384; PMID 35894942); the authors of the largest such series explicitly call for a prospective study including surveillance patients and patients who dissect. No such validation exists.
  • Does WSS mapping change surgical decisions? Guzzardi et al. proposed 4D-flow WSS as a non-invasive biomarker able to individualise resection extent and stated that validation is warranted (PMID 26293758); no trial has tested WSS-guided resection against standard extent.
  • Why does the aorta lose mechanical anisotropy, and is isotropy prognostic? Loss of directional dependency tracks collagen:elastin ratio and medial degeneration, becoming complete above collagen:elastin > 2 (PMID 27888778), but whether isotropy predicts events independently of diameter has not been tested.
  • What determines the outer-curve vulnerability to dissection propagation? Outer-curve tissue is thinner (2.21 vs 2.50 mm) with the lowest longitudinal peel force and dissection energy (PMID 35894942), and radially-running collagen fibre density governs peel strength (PMID 24484644) — but why radial fibre architecture differs regionally, and whether it is modifiable, is unknown.
  • Can mouse biomechanics guide human risk? Across 10 murine models, aneurysm-proneness tracked failure to maintain circumferential material stiffness rather than any measure of reduced mechanical function (PMID 28490606); whether an analogous stiffness-homeostasis metric can be measured non-invasively in humans is untested.

References

  1. Vorp DA, Schiro BJ, Ehrlich MP, et al. Effect of aneurysm on the tensile strength and biomechanical behavior of the ascending thoracic aorta. Ann Thorac Surg. 2003;75:1210-4. PMID 12683565
  2. Salmasi MY, Sasidharan S, Frattolin J, et al. Regional variation in biomechanical properties of ascending thoracic aortic aneurysms. Eur J Cardiothorac Surg. 2022;62:ezac392. PMID 35894942
  3. Gomez A, Wang Z, Xuan Y, et al. Regional wall stress differences on tricuspid aortic valve-associated ascending aortic aneurysms. Interact Cardiovasc Thorac Surg. 2022;34:1115-1123. PMID 34718581
  4. Chung J, Lachapelle K, Cartier R, Mongrain R, Leask RL. Loss of mechanical directional dependency of the ascending aorta with severe medial degeneration. Cardiovasc Pathol. 2016;26:45-50. PMID 27888778
  5. Xuan Y, Wang Z, Liu R, et al. Wall stress on ascending thoracic aortic aneurysms with bicuspid compared with tricuspid aortic valve. J Thorac Cardiovasc Surg. 2018;156:492-500. PMID 29656820
  6. Bellini C, Bersi MR, Caulk AW, et al. Comparison of 10 murine models reveals a distinct biomechanical phenotype in thoracic aortic aneurysms. J R Soc Interface. 2017;14:20161036. PMID 28490606
  7. Pal S, Tsamis A, Pasta S, et al. A mechanistic model on the role of "radially-running" collagen fibers on dissection properties of human ascending thoracic aorta. J Biomech. 2014;47:981-8. PMID 24484644
  8. Trabelsi O, Davis FM, Rodriguez-Matas JF, Duprey A, Avril S. Patient specific stress and rupture analysis of ascending thoracic aneurysms. J Biomech. 2015;48:1836-43. PMID 25979384
  9. Humphrey JD. Possible mechanical roles of glycosaminoglycans in thoracic aortic dissection and associations with dysregulated transforming growth factor-β. J Vasc Res. 2012;50:1-10. PMID 23018968
  10. Chung J, Lachapelle K, Wener E, et al. Energy loss, a novel biomechanical parameter, correlates with aortic aneurysm size and histopathologic findings. J Thorac Cardiovasc Surg. 2014;148:1082-8. PMID 25129601
  11. Condemi F, Campisi S, Viallon M, Croisille P, Avril S. Relationship Between Ascending Thoracic Aortic Aneurysms Hemodynamics and Biomechanical Properties. IEEE Trans Biomed Eng. 2019;67:949-956. PMID 31331875
  12. Barker AJ, Markl M, Bürk J, et al. Bicuspid aortic valve is associated with altered wall shear stress in the ascending aorta. Circ Cardiovasc Imaging. 2012;5:457-66. PMID 22730420
  13. Guzzardi DG, Barker AJ, van Ooij P, et al. Valve-Related Hemodynamics Mediate Human Bicuspid Aortopathy: Insights From Wall Shear Stress Mapping. J Am Coll Cardiol. 2015;66:892-900. PMID 26293758
  14. Bollache E, Guzzardi DG, Sattari S, et al. Aortic valve-mediated wall shear stress is heterogeneous and predicts regional aortic elastic fiber thinning in bicuspid aortic valve-associated aortopathy. J Thorac Cardiovasc Surg. 2018;156:2112-2120. PMID 30060930
  15. van Ooij P, Garcia J, Potters WV, et al. Age-related changes in aortic 3D blood flow velocities and wall shear stress: Implications for the identification of altered hemodynamics in patients with aortic valve disease. J Magn Reson Imaging. 2016;43:1239-49. PMID 26477691
  16. Suwa K, Rahman OA, Bollache E, et al. Effect of Aortic Valve Disease on 3D Hemodynamics in Patients With Aortic Dilation and Trileaflet Aortic Valve Morphology. J Magn Reson Imaging. 2019;51:481-491. PMID 31169969
  17. Bouaou K, Dietenbeck T, Soulat G, et al. Four-dimensional flow cardiovascular magnetic resonance aortic cross-sectional pressure changes and their associations with flow patterns in health and ascending thoracic aortic aneurysm. J Cardiovasc Magn Reson. 2024;26:101030. PMID 38403074
  18. Kaiser AD, Shad R, Schiavone N, Hiesinger W, Marsden AL. Controlled Comparison of Simulated Hemodynamics Across Tricuspid and Bicuspid Aortic Valves. Ann Biomed Eng. 2022;50:1053-1072. PMID 35748961
  19. Humphrey JD, Schwartz MA. Vascular Mechanobiology: Homeostasis, Adaptation, and Disease. Annu Rev Biomed Eng. 2021;23:1-27. PMID 34255994
  20. Humphrey JD, Schwartz MA, Tellides G, Milewicz DM. Role of mechanotransduction in vascular biology: focus on thoracic aortic aneurysms and dissections. Circ Res. 2015;116:1448-61. PMID 25858068
  21. Bellini C, Wang S, Milewicz DM, Humphrey JD. Myh11(R247C/R247C) mutations increase thoracic aorta vulnerability to intramural damage despite a general biomechanical adaptivity. J Biomech. 2015;48:113-21. PMID 25433566
  22. Humphrey JD, Milewicz DM. The aorta in Marfan syndrome: from molecular mechanisms to mechanobiological dysfunction. Cardiovasc Res. 2026;122:1427-1442. PMID 42140666