This article provides an overview of CT and MR imaging of the aortic root, highlighting normal anatomy, pathologic conditions, imaging techniques, measurement thresholds, surgical procedures, postoperative complications, and imaging pitfalls.
Abstract
Three-dimensional datasets acquired using computed tomography and magnetic resonance imaging are ideally suited for characterization of the aortic root. These modalities offer different advantages and limitations, which must be weighed according to the clinical context. This article provides an overview of current aortic root imaging, highlighting normal anatomy, pathologic conditions, imaging techniques, measurement thresholds, relevant surgical procedures, postoperative complications and potential imaging pitfalls. Patients with a range of clinical conditions are predisposed to aortic root disease, including Marfan syndrome, bicuspid aortic valve, vascular Ehlers-Danlos syndrome, and Loeys-Dietz syndrome. Various surgical techniques may be used to repair the aortic root, including placement of a composite valve graft, such as the Bentall and Cabrol procedures; placement of an aortic root graft with preservation of the native valve, such as the Yacoub and David techniques; and implantation of a biologic graft, such as a homograft, autograft, or xenograft. Potential imaging pitfalls in the postoperative period include mimickers of pathologic processes such as felt pledgets, graft folds, and nonabsorbable hemostatic agents. Postoperative complications that may be encountered include pseudoaneurysms, infection, and dehiscence. Radiologists should be familiar with normal aortic root anatomy, surgical procedures, and postoperative complications, to accurately interpret pre- and postoperative imaging performed for evaluation of the aortic root.
Online supplemental material is available for this article.
©RSNA, 2015
SA-CME LEARNING OBJECTIVES
After completing this journal-based SA-CME activity, participants will be able to:
■ List common pathologic conditions affecting the aortic root.
■ Describe various surgical techniques for repair of the aortic root.
■ Identify normal and abnormal imaging findings before and after aortic root surgery.
Introduction
Detailed evaluation of the entire spectrum of aortic root disease is possible with three-dimensional (3D) datasets provided by computed tomographic (CT) angiography and magnetic resonance (MR) imaging. The objective of this article is to review the role of CT angiography and MR imaging in the diagnosis, surveillance, and pre- and postoperative evaluation of aortic root disease. Patients with several underlying conditions, including connective tissue disorders such as Marfan syndrome, bicuspid aortic valve, and Loeys-Dietz syndrome, are predisposed to aneurysmal enlargement of the aortic root with potentially catastrophic outcomes if left untreated. A variety of surgical procedures are available to repair the aortic root, including placement of a composite valve graft, as in the Cabrol and Bentall procedures; placement of an aortic root graft with preservation of the native valve, as in the Yacoub and David valve-sparing techniques; and implantation of a biologic graft, as in the Ross procedure. In this article, we review the anatomy of the aortic root, pathologic conditions, relevant imaging techniques, measurement thresholds, and surgical approaches. We also discuss normal and abnormal pre- and postoperative imaging appearances including potential pitfalls in interpretation.
Anatomy of the Aortic Root
The aortic root is a complex structure that separates the left ventricle from the systemic circulation. It is located in close proximity to other cardiac structures, mainly the pulmonary valve anteriorly, the mitral valve to the left and posteriorly, and the tricuspid valve to the right and posteriorly (1). The aortic root consists of several distinct components, including the aortic annulus, the sinuses of Valsalva (or aortic sinuses), and the sinotubular junction (Fig 1) (1). The internal structure of the aortic root consists of the aortic valve leaflets, the commissures, and the interleaflet triangles.
Figure 1a.
Anatomy of the aortic root. (a, b) Illustrations of the aortic root in virtual longitudinal open section (a) and oblique coronal orientation (b). (c) Oblique coronal volume-rendered CT angiogram of the aortic root with some key anatomic structures labeled.
Figure 1b.
Anatomy of the aortic root. (a, b) Illustrations of the aortic root in virtual longitudinal open section (a) and oblique coronal orientation (b). (c) Oblique coronal volume-rendered CT angiogram of the aortic root with some key anatomic structures labeled.
Figure 1c.
Anatomy of the aortic root. (a, b) Illustrations of the aortic root in virtual longitudinal open section (a) and oblique coronal orientation (b). (c) Oblique coronal volume-rendered CT angiogram of the aortic root with some key anatomic structures labeled.
Aortic Annulus
In surgical terms, the aortic annulus (in Latin, anulus [a diminutive of anus], meaning a small ring) refers to a virtual ring surrounding the ventriculoaortic junction just below the lowest insertion points of the aortic valve. In this sense, the surgical annulus refers to a description of tissues suitable for anastomosing a circular graft or prosthesis to, rather than a true anatomic structure. The true anatomic annulus is a thick fibrous coronet, or crown-shaped structure, that follows the valve insertions, and provides support to the aortic valve complex (2,3). It outlines the true boundary between tissues exposed to ventricular versus aortic pressure changes over the cardiac cycle. The elliptical base of the crown-shaped anatomic aortic annulus corresponds closely to the surgical aortic annulus mentioned previously, which is also referred to as the virtual basal ring in the setting of transcatheter aortic valve replacement (4,5).
Aortic Valve Leaflets and Leaflet Attachments
The leaflets that form the aortic valve constitute the hemodynamic junction and physical boundary between the left ventricle and the aorta, preventing regurgitation of blood from the aorta into the left ventricle during diastole, and allowing the appropriate egress of blood from the left ventricle to the aorta during systole (6). The aortic valve leaflets or cusps are semilunar in appearance, and consist of three parts: the free margin, the belly, and the base. When the aortic valve is closed, the free margins of adjacent cusps touch each other. The coaptation zone refers to the area of the free margins that come in contact between neighboring cusps. At the center of each free margin is a thickened fibrous bulge (these are the nodules of Arantius) that provides the central coaptation area (6). Normally the aortic valve has three leaflets (making it a tricuspid or trileaflet valve), but it may have other, abnormal configurations, including four leaflets (a quadricuspid valve), two leaflets (bicuspid valve), or a single, typically deformed leaflet (unicuspid valve).
Commissures
The commissures refer to the junctions of adjacent aortic valve cusps at the site of their attachment to the aortic wall at the level of the sinotubular junction. In a trileaflet aortic valve, the three commissures are normally evenly spaced around the aortic circumference, and allow opening of the aortic valve all the way to the aortic insertion. The commissures can be partially or completely fused, because of either a congenital abnormality such as a bicuspid aortic valve or a degenerative process such as senile aortic stenosis, resulting in restricted valve motion and incomplete opening during systole.
Interleaflet Triangle
Below the commissures of the valve insertions, and above the virtual basal ring, lie the three interleaflet triangles, which are extensions of the left ventricular outflow tract. Each triangle extends to the level of the sinotubular junction superiorly. Inferiorly, the interleaflet triangle between the left and noncoronary cusps is contiguous with the anterior leaflet of the mitral valve.
Sinuses of Valsalva
The outward bugles of the aortic root associated with each of the three aortic valve cusps are collectively referred to as the sinuses of Valsalva (or aortic sinuses) after the anatomist Antonio Valsalva. The three sinuses are named with reference to the coronary artery ostia: the right, left, and noncoronary sinuses. The right and left coronary sinuses give rise to the right and left coronary arteries, respectively.
Sinotubular Junction
The sinotubular junction is the landmark between the aortic root and the tubular portion of the ascending aorta, and normally forms a waist in the aortic contour between the sinus of Valsalva and the remainder of the ascending aorta. The sinotubular junction is the highest level at which the aortic valve cusps and commissures are attached to the aortic wall.
Imaging of the Aortic Root
Echocardiography
Echocardiography is widely used for the detection and follow-up of aortic root disease, due to its wide availability, relatively low cost, safety, and ability to assess hemodynamic parameters of the aortic valve. Two-dimensional (2D) transthoracic echocardiographic measurements of the aortic root and ascending aorta are usually obtained from the parasternal long-axis view, which is roughly equivalent to the three-chamber view in cardiac CT or MR imaging. In this view, measurements of the sinuses of Valsalva usually correspond to measurements between the right and noncoronary sinuses. Given that 2D echocardiography provides a measurement on the basis of a single plane, this technique may not yield the largest aortic diameter measurement, particularly in the setting of asymmetric root enlargement (7). Evaluation of the extent of an aortic root aneurysm that extends to the mid– or distal ascending aorta is limited, and echocardiography is not well-suited for assessment of the remainder of the thoracic aorta. Other echocardiographic limitations include patient-specific factors (such as body habitus) and operator dependence. Three-dimensional echocardiography and transesophageal echocardiography can overcome some of these limitations, allowing for measurements in multiple planes. Transesophageal echocardiography is typically used for problem solving and for intraoperative imaging.
CT Imaging
Modern CT scanners acquire high-resolution 3D datasets of the thoracic aorta. Synchronization of the data acquisition or the reconstruction with a patient’s electrocardiogram (ECG) allows suppression of pulsation artifacts transmitted from the beating heart. ECG synchronization is critical for detailed assessment of aortic root anatomy and disease (8). ECG gating has been shown to improve the accuracy and reproducibility of aortic measurements, and allows for viewing of images in a particular phase of the cardiac cycle (9,10) (Movie 1).
Movie 1.
Axial oblique volume-rendered CT angiography slab cine loop through a normal trileaflet aortic valve demonstrates unrestricted valve opening.
Several different cardiac synchronization methods are available on modern CT scanners. Prospective ECG triggering synchronizes the radiographic exposure to the cardiac cycle. Images are only acquired during a specified portion of the cardiac cycle, starting at a predetermined delay from the R wave. Retrospective ECG gating acquires redundant helical CT data, and the desired cardiac phase is selected retrospectively from the raw data, which provides the flexibility to reconstruct images at different cardiac phases. However, this is accomplished at the cost of a higher radiation dose (11). Radiation dose may be reduced with use of ECG modulation of the tube current with concentration of the highest dose at a specified cardiac phase (12). Finally, some scanners allow fast, high-pitch, prospectively triggered helical scans. Using this technique, the proper start time of table movement is timed on the basis of the ECG rhythm, so that images of the aortic root are acquired in diastole when the heart is relatively stationary. Only a single phase of the cardiac cycle is acquired with this technique, and therefore cine visualization of the aortic valve and root is not an option.
The technical details of CT angiographic acquisition and ECG synchronization are beyond the scope of this article. For imaging of the aortic root it is necessary, however, to understand how the advantages and disadvantages of each technique can be tailored to the specific clinical situation. For the purpose of surveillance of a known aortic root aneurysm, any technique can be used, as the objective is to suppress motion artifact to obtain reliable measurements. In this setting, a low radiation dose option can be selected, such as prospective sequential triggering without padding, retrospective gating with tube-current modulation optimized for diastolic-phase datasets only, or a prospectively triggered high-pitch helical acquisition. For the initial evaluation of an aortic root aneurysm and for immediate preoperative CT angiography of the aortic root, both systolic and diastolic phases of the cardiac cycle are required to assess valve function and geometry. This can be achieved with a prospectively triggered sequential scan with an exposure window that includes both systole and diastole, or with retrospective gating with ECG-based tube-current modulation maximized from end systole through diastole (for example, from 40% to 70% of the R-R interval). Retrospective ECG gating allows for the best cinematic reconstruction over the entire cardiac cycle, which is often helpful in the evaluation of complex root and valve disease. Similarly, postoperative, predischarge imaging requires high-quality imaging over the cardiac cycle, because complications such as a small leak or pseudoaneurysm may only be visualized in a particular phase of the cardiac cycle.
MR Imaging
MR imaging techniques that are used to image the aortic root include balanced steady-state free precession (SSFP) sequences, black-blood sequences, phase-contrast imaging, and MR angiography (Movie 2, Fig 2). Most cardiovascular sequences are acquired using a breath-hold technique to minimize any artifacts from respiratory motion. Short scan times are achieved through the use of a segmented k-space technique, with a fixed number of k-space readout lines or “views per segment” acquired per cardiac cycle (13). The higher this number is, the shorter the total scan time and the lower the temporal resolution.
Figure 2a.

Preoperative multisequence MR imaging in a 30-year-old man with Marfan syndrome and annuloaortic ectasia. (a) Oblique sagittal maximum intensity projection reformation of a contrast-enhanced MR angiographic image. (b) Oblique coronal T1-weighted black-blood double inversion-recovery image. (c, d) Oblique coronal 2D phase-contrast magnitude (c) and velocity (d) images in systole.
Movie 2.
Axial oblique cine SSFP MR imaging through a normal trileaflet aortic valve.
Figure 2b.

Preoperative multisequence MR imaging in a 30-year-old man with Marfan syndrome and annuloaortic ectasia. (a) Oblique sagittal maximum intensity projection reformation of a contrast-enhanced MR angiographic image. (b) Oblique coronal T1-weighted black-blood double inversion-recovery image. (c, d) Oblique coronal 2D phase-contrast magnitude (c) and velocity (d) images in systole.
Figure 2c.

Preoperative multisequence MR imaging in a 30-year-old man with Marfan syndrome and annuloaortic ectasia. (a) Oblique sagittal maximum intensity projection reformation of a contrast-enhanced MR angiographic image. (b) Oblique coronal T1-weighted black-blood double inversion-recovery image. (c, d) Oblique coronal 2D phase-contrast magnitude (c) and velocity (d) images in systole.
Figure 2d.

Preoperative multisequence MR imaging in a 30-year-old man with Marfan syndrome and annuloaortic ectasia. (a) Oblique sagittal maximum intensity projection reformation of a contrast-enhanced MR angiographic image. (b) Oblique coronal T1-weighted black-blood double inversion-recovery image. (c, d) Oblique coronal 2D phase-contrast magnitude (c) and velocity (d) images in systole.
Two-dimensional, bright-blood, cine imaging utilizing balanced SSFP sequences provides high signal-to-noise ratio and good contrast between the blood pool and vessel wall (14). Contrast is determined on the basis of the ratio of T2 to T1 rather than on the basis of inflow effects as in spoiled gradient-echo methods. SSFP-cine images may be obtained transversely to the aortic root to assess valvular motion or in short-axis orientation through the entire left ventricle to quantify left ventricular volumes and ejection fraction.
Black-blood double inversion-recovery imaging can be combined with ECG gating and respiratory gating, and is particularly useful for assessing the aortic wall and for detecting slow flow or static blood within the lumen. Black-blood sequences typically produce images at a single cardiac phase and do not provide temporal information regarding cardiac motion or blood flow.
Phase-contrast sequences generate quantitative velocity information, and produce two types of images: magnitude images demonstrating anatomy, and one or more phase images encoding the components of the velocity vectors of motion such as blood flow. Phase-contrast sequences are ECG gated and result in cine clips of both magnitude and phase images. Phase-contrast data acquisition can be 2D, with orientation of the image plane perpendicular to the region of interest using a breath-hold technique, or can be 3D, acquired in free breathing. The latter approach has been termed four-dimensional (4D) flow imaging (15). By acquiring flow data in a 3D space, 4D flow imaging eliminates the need to perform multiple 2D–phase-contrast scans, and the time-consuming task of positioning these imaging planes precisely while the patient is in the scanner. With either 2D–phase-contrast or 4D flow acquisitions, velocity vectors within a vessel lumen can be summed to quantify flow. Phase-contrast imaging at the aortic valve may be used to quantify the regurgitant fraction in the setting of aortic insufficiency, and the transvalvular pressure gradient in the setting of aortic stenosis.
Contrast-enhanced MR angiography utilizes the T1-shortening effect of gadolinium-based contrast agents (GBCAs) and can produce high-resolution 3D images of the entire aorta in a single breath-hold. ECG gating is particularly important for imaging of the aortic root to improve vessel wall definition and to reduce phase-ghosting artifact from cardiac motion, at the cost of a longer imaging time. Signal enhancement of the vessel lumen depends on the intra-arterial concentration of the contrast agent. Timing of contrast agent injection must be optimized such that collection of the central lines of k-space occurs during the plateau phase of arterial enhancement (16). Several methods may be used to determine the optimal delay between the start of contrast agent injection and the start of image acquisition, including injection of a small test bolus and the use of automatic triggering (17,18). If first-pass contrast enhancement is not required, MR angiography may be performed at the contrast-equilibrium phase using a blood pool agent. Time-resolved MR angiography, another MR imaging technique, allows for dynamic evaluation of the aorta independently of bolus timing (19).
Noncontrast MR angiography may be performed using a 3D, ECG-gated, respiratory-navigated, balanced SSFP sequence. No exogenous contrast agent is required, as hyperintense signal within the vessel lumen arises as a result of the T2 and T1 properties of blood. To maintain cardiac synchronization, data are recorded in the 3D k-space at a predetermined cardiac phase. This sequence is usually acquired during free breathing, given that filling of k-space at the required matrix size usually takes several minutes. To minimize respiratory motion artifact, respiratory gating is used so that data are only acquired when the diaphragm moves to a predetermined position. The position of the diaphragm is determined with a short navigator sequence interspersed between SSFP sequences.
CT and MR imaging each have unique advantages and limitations with respect to imaging of the aortic root, and the choice between the two modalities depends on availability, the clinical indication, safety considerations, and the condition of the patient. Unlike CT angiography, imaging of the aortic root with MR imaging offers techniques that do not require contrast agent administration. Owing to its lack of ionizing radiation and associated risks, MR imaging is particularly useful in younger patients and those who require multiple sequential follow-up studies. However, image spatial resolution is generally lower in MR imaging than with CT angiography, and MR imaging studies typically require a significantly longer time to perform. The MR imaging environment is not suited for critically ill patients who require close monitoring. Finally, conventional implanted pacemakers and defibrillators are considered relative contraindications to MR imaging.
Image Analysis
CT and MR imaging 3D datasets should be reviewed on a workstation with multiplanar reformatting and measurement capabilities. Images should be manipulated such that reported aortic diameters are measured orthogonally to the aortic lumen, as measurements that are off-axis may significantly overestimate the true aortic diameter (20). Postprocessing techniques that are commonly used include multiplanar reformation, maximum intensity projection, curved planar reformation, and 3D volume rendering.
Different aortic measurement techniques have been reported, including leading edge–to–leading edge (which is frequently the standard used in transthoracic echocardiography), whereas others report inner lumen–only measurements (7). Although there is no clear consensus on the best measurement approach in the literature to date, the measurement technique used at a given institution should be standardized to facilitate meaningful comparison at follow-up studies. It is also important that surgeons and cardiologists caring for these patients, as well as those involved in imaging, are aware of the methodological differences between modalities. Although echocardiography is often considered the practical standard of reference for aortic root measurements, echocardiographic measurements of the aortic root obtained from a parasternal window are often smaller than those obtained by 3D imaging using CT angiography or MR imaging, with potentially important implications with respect to treatment decisions and surgical planning.
Aortic measurements that are typically reported include minimum and maximum diameters of the aortic annulus, sinotubular junction, midascending aorta, distal ascending aorta, aortic arch, and middescending aorta (7,21). For the sinuses of Valsalva, both sinus-to-sinus and sinus-to-commissure measurements should be provided (Fig 3). If the maximum thoracic aortic dimension is not captured in one of these measurements, then this should also be specified and reported.
Figure 3a.

Double-oblique CT angiogram at the level of the sinuses of Valsalva demonstrating sinus-to-sinus (a) and sinus-to-commissure (b) measurement techniques.
Figure 3b.

Double-oblique CT angiogram at the level of the sinuses of Valsalva demonstrating sinus-to-sinus (a) and sinus-to-commissure (b) measurement techniques.
Pathologic Conditions
Aneurysms of the ascending aorta result most often from cystic medial degeneration (22). Aortic root enlargement is associated with many underlying conditions including connective tissue disorders such as Marfan syndrome, bicuspid aortic valve, Loeys-Dietz syndrome, hypertension, and atherosclerosis. Less common underlying causes include infections such as syphilis, vasculitides, trauma, and familial thoracic aneurysm disease.
Marfan Syndrome
Marfan syndrome is an autosomal dominant, inherited connective-tissue disorder that is mainly caused by mutations in the gene (FBN1) that encodes fibrillin-1 (23). Approximately 25% of cases result from a sporadic new mutation in FBN1. Estimates of prevalence are approximately 1 case per 3500–5000 population.
Fibrillin-1 is a major constituent of extracellular microfibrils; clinical manifestations of FBN1 mutations are thought to be due to structural abnormalities of fibrillin-1 and resulting dysregulation of transforming growth factor β (TGFβ) signaling. Aortic root dilatation is the leading cause of morbidity and mortality in patients with Marfan syndrome. The altered structure of the aortic wall predisposes patients to aortic dissection and rupture. The classic phenotypic appearance of the aorta in patients with Marfan syndrome is annuloaortic ectasia, with dilatation of the aortic annulus and sinuses of Valsalva, and effacement of the sinotubular junction, resulting in a tulip-shaped configuration of the aortic root (Fig 4) (7,24). As the annulus expands, poor aortic cusp coaptation and distortion of the cusps may eventually result in aortic insufficiency, which occurs in 15%–44% of patients (25). Mitral valve prolapse is the most prevalent valvular abnormality in patients with Marfan syndrome, affecting 35%–100% of patients (26).
Figure 4.

Aortic root aneurysm in a 33-year-old woman with Marfan syndrome. Oblique sagittal volume-rendered slab CT angiogram demonstrates annuloaortic ectasia with aneurysmal dilatation of the aortic root and proximal ascending aorta, resulting in the classic tulip configuration, with normal caliber of the distal ascending aorta, aortic arch, and descending aorta.
Cardiovascular imagers should be aware of other phenotypic features that might give clues toward a diagnosis of underlying Marfan syndrome, including scoliosis, pectus deformities, and dural ectasia. Importantly, these features may all contribute to the systemic score of the revised Ghent nosology, which is used to establish a diagnosis of Marfan syndrome (7,25,27).
Bicuspid Aortic Valve
Bicuspid aortic valve is the most common congenital cardiac abnormality, with an estimated prevalence of 1%–2% (7). The inheritance pattern is autosomal dominant with incomplete penetrance. Bicuspid aortic valve shares common histopathologic findings with Marfan syndrome, including cystic medial degeneration, increased levels of matrix metalloproteinases, and decreased levels of fibrillin-1 in the aortic wall (22,28). The Sievers classification of bicuspid aortic valve is based on the number of raphes and the positions of cusps and raphes (29). Fusion of the right and left coronary cusps is the most common type.
Bicuspid aortic valve is associated with aortic root, and ascending and transverse arch aortic dilatation, even in the absence of hemodynamically significant valve dysfunction (Fig 5) (30). Aortic root dilatation is an important and potentially lethal complication occurring in nearly one-half of patients. Given the frequency of this complication, patients with bicuspid aortic valve routinely undergo imaging studies to monitor the size of their aorta. Patients are commonly followed using echocardiography if only the aortic root and proximal ascending aorta are involved. However, CT or MR imaging may be useful in patients with poor acoustic windows or in those requiring more detailed evaluation of the aortic arch. Bicuspid aortic valve is also associated with aortic coarctation (31).
Figure 5a.

Aortic root aneurysm in a 40-year-old man with bicuspid aortic valve. (a) Oblique sagittal volume-rendered slab CT angiogram demonstrates an aortic root and ascending aortic aneurysm tapering into the proximal aortic arch. (b) Axial oblique volume-rendered CT angiogram in systole demonstrates a bicuspid aortic valve with no raphe (Sievers type 0), and unrestricted valve opening (in a “fish-mouth” configuration).
Figure 5b.

Aortic root aneurysm in a 40-year-old man with bicuspid aortic valve. (a) Oblique sagittal volume-rendered slab CT angiogram demonstrates an aortic root and ascending aortic aneurysm tapering into the proximal aortic arch. (b) Axial oblique volume-rendered CT angiogram in systole demonstrates a bicuspid aortic valve with no raphe (Sievers type 0), and unrestricted valve opening (in a “fish-mouth” configuration).
Loeys-Dietz Syndrome
Loeys-Dietz syndrome is a less prevalent inherited disorder, which results in aneurysmal aortic root dilatation similar to Marfan syndrome. This syndrome is also inherited in an autosomal dominant pattern, and results from mutations in the genes encoding TGFβ receptors 1 and 2, (TGFBR1 and TGFBR2, respectively) (32). The arterial disease encountered in patients with Loeys-Dietz syndrome is typically severe and widespread, and can involve all aortic segments and major branching arteries, necessitating cardiovascular imaging beyond the aortic root. Importantly, complications, including aortic dissections, occur earlier and at smaller aortic diameters than with patients with Marfan syndrome, warranting close imaging monitoring and potentially earlier surgical intervention (7,32).
Ehlers-Danlos Syndrome, Vascular Type
The vascular type (also known as type IV) of Ehlers-Danlos syndrome is a rare autosomal dominant condition resulting from mutations in the gene (COL3A1) that encodes type III procollagen. Excessive tissue fragility predisposes patients to vascular complications, including aortic aneurysms, dissections, and rupture (33). Approximately 60% of patients experience a vascular complication by age 40 years (34). Procedure-related morbidity among patients requiring surgery is particularly high, due to both vascular complications such as hemorrhage, and graft-related complications such as anastomotic aneurysms and thrombosis (34).
Turner Syndrome
Turner syndrome is a chromosomal abnormality (45,X) associated with dilatation of the ascending aorta in over 40% of patients (35). Additional cardiovascular manifestations include bicuspid aortic valve, juxtaductal coarctation or pseudocoarctation of the aortic arch, elongation of the arch, and partial anomalous pulmonary venous return (36,37). Other phenotypic features include short stature, low-set ears, and a webbed neck (pterygium colli). Patients with Turner syndrome may experience aortic dissections at aortic diameters smaller than the average population (37). Patients with Turner syndrome with both bicuspid aortic valve and aortic dilatation may be at particularly high risk, given that both are important risk factors for aortic dissection (22). Therefore, routine monitoring of aortic size is recommended for the detection of aortic dilatation and potential complications.
Familial Thoracic Aortic Aneurysm Disease
Cystic medial degeneration may also be seen in patients with aortic root aneurysms who do not have established connective-tissue disorders. Patients with a notable family history of thoracic aortic aneurysms and dissections may have familial thoracic aortic aneurysm syndrome, typically presenting with it at a significantly younger age than those with sporadic aneurysms (38). Several mutations have been identified, with most pedigrees suggesting an autosomal dominant mode of inheritance with variability in both expression and penetrance (22).
Other Causes
Atherosclerosis is a known cause of aortic aneurysms, but most commonly results in aneurysms of the descending thoracic aorta, and less commonly the arch or ascending aorta. Atherosclerotic aneurysms are typically diffuse or fusiform. Even if the ascending aorta is involved, the aortic root is typically spared or only minimally affected. Other conditions that may result in aortic root dilatation include vasculitides (such as Takayasu arteritis and giant cell arteritis), infection (such as syphilis), trauma, and chronic aortic dissections.
Aortic Root Dimensions
Abnormal dilatation of the thoracic aorta has been defined as a diameter greater than 50% above the high end of the normal range (39). The normal range of aortic diameters depends on age, sex, and body size (43,44), as well as on the anatomic segment. The average aortic diameter at the sinuses of Valsalva is 3.0 cm ± 0.5 (40). Thoracic aortic measurements greater than 4 cm are generally considered to be consistent with an aneurysm.
Aneurysm size is the primary indicator for surgical aortic root repair of asymptomatic aneurysms. In most instances, surgery is indicated for asymptomatic patients without another underlying cardiovascular condition or disease, with aneurysms measuring greater than or equal to 5.5 cm (41,42). The risk of aortic dissection or rupture at diameters above 5.5 cm is generally perceived to exceed the risk of operation, warranting intervention (41). In patients with evidence of rapid aortic growth (>0.5–1 cm per year), elective surgical intervention should also be considered even if absolute size criteria have not been met (43). Patients who are symptomatic or demonstrate aortic valve insufficiency may also warrant earlier intervention. Conversely, among patients with increased surgical risk, the threshold may be raised to greater than or equal to 6 cm before surgery is recommended, such as in older individuals with atherosclerotic aneurysms.
Among patients with an underlying condition who are at increased risk of aortic dissection or rupture, or among those with a concomitant need for other cardiac surgery, ascending aortic repair may be recommended at smaller diameters (28,41). When patients with bicuspid aortic valve require aortic valve replacement surgery, prophylactic replacement of the ascending aorta and hemiarch may be recommended if the aortic diameter is greater than or equal to 4 cm, given that such patients would otherwise remain at high risk for subsequent aortic dissection (22). In patients with Marfan syndrome, the surgical repair threshold may be 4.5–5 cm depending on the length of aortic dilatation, demonstrated aortic growth of greater than 5 mm per year, the presence of aortic regurgitation, and a family history of aortic dissection at a diameter of less than 5 cm, given that these factors predict a poorer prognosis (44). Aortic dissection may occur at even smaller diameters in patients with Loeys-Dietz syndrome, therefore surgical repair is often recommended at an aortic diameter of greater than 4–4.5 cm, or in the setting of aortic expansion of greater than 5 mm per year (44,45). No strict size criteria are established for surgical intervention in patients with vascular Ehlers-Danlos syndrome, as they have notable surgical morbidity risk and therefore surgical repair is usually reserved for life-threatening cases (34).
Surgical Approaches
Composite Valve Graft
In patients with aneurysms or dissections involving the aortic root and ascending aorta, the traditional surgical aortic root replacement is the so-called composite valve graft. In the Bentall procedure, which was originally described in 1968, the native aortic root and ascending aorta are replaced with a composite valve graft that contains an aortic valve prosthesis. The coronary arteries are reimplanted into the graft (46). A modified technique is now most commonly performed, in which a coronary button of the aorta is taken along with the coronary artery, facilitating anastomosis to the graft (Fig 6) (47,48).
Figure 6a.

Modified Bentall composite valve graft. (a) Illustration shows a synthetic valved conduit (composite valve graft) containing a mechanical aortic valve used to replace the aortic root and ascending aorta. In this procedure, a button of the aorta encircling each coronary ostium is removed with the coronary artery and anastomosed to the ascending aortic graft. (b) Normal postoperative oblique coronal CT angiogram of a 49-year-old man after composite valve graft repair of the aortic root using the modified Bentall procedure. The left coronary button anastomosis (arrow) and high-attenuating mechanical prosthetic valve are identifiable.
Figure 6b.
Modified Bentall composite valve graft. (a) Illustration shows a synthetic valved conduit (composite valve graft) containing a mechanical aortic valve used to replace the aortic root and ascending aorta. In this procedure, a button of the aorta encircling each coronary ostium is removed with the coronary artery and anastomosed to the ascending aortic graft. (b) Normal postoperative oblique coronal CT angiogram of a 49-year-old man after composite valve graft repair of the aortic root using the modified Bentall procedure. The left coronary button anastomosis (arrow) and high-attenuating mechanical prosthetic valve are identifiable.
In the Cabrol procedure, which was developed in 1981 as an alternative to the Bentall procedure, the coronary ostia are anastomosed to a prosthetic conduit in an end-to-end anastomosis (49). This conduit is anastomosed to the ascending aortic graft in a side-to-side anastomosis, typically posterior to the graft (Fig 7) (50). This technique is useful in patients with severe atherosclerosis or postoperative changes (in the setting of a re-do procedure) of the aortic root that preclude obtaining good-quality aortic buttons, in those patients with severe proximal coronary artery disease, or for patients in whom the coronary arteries cannot be mobilized sufficiently to achieve a tension-free anastomosis directly to the aortic graft (50,51). The imaging appearance of the retroaortic conduit may be mistaken for an intimal flap related to aortic dissection (51,52). Complications following composite valve graft surgery include aortic dissection, anastomotic leak, pseudoaneurysm formation at the coronary artery anastomosis, and coronary graft thrombosis (51,53).
Figure 7a.

Cabrol composite valve graft. (a) Illustration shows the Cabrol procedure, in which the coronary ostia are anastomosed to a prosthetic conduit (end-to-end anastomosis). This conduit is then anastomosed to the ascending aortic graft (side-to-side anastomosis). (b, c) Oblique coronal (b) and oblique axial (c) CT angiograms of a 31-year-old man with Loeys-Dietz syndrome demonstrate the normal postoperative appearance of the Cabrol procedure. The retroaortic synthetic conduit graft (long arrow) and high-attenuating mechanical aortic valve are identifiable. The synthetic conduit graft is anastomosed to the left (short arrow) and right coronary arteries in an end-to-end fashion.
Figure 7b.
Cabrol composite valve graft. (a) Illustration shows the Cabrol procedure, in which the coronary ostia are anastomosed to a prosthetic conduit (end-to-end anastomosis). This conduit is then anastomosed to the ascending aortic graft (side-to-side anastomosis). (b, c) Oblique coronal (b) and oblique axial (c) CT angiograms of a 31-year-old man with Loeys-Dietz syndrome demonstrate the normal postoperative appearance of the Cabrol procedure. The retroaortic synthetic conduit graft (long arrow) and high-attenuating mechanical aortic valve are identifiable. The synthetic conduit graft is anastomosed to the left (short arrow) and right coronary arteries in an end-to-end fashion.
Figure 7c.
Cabrol composite valve graft. (a) Illustration shows the Cabrol procedure, in which the coronary ostia are anastomosed to a prosthetic conduit (end-to-end anastomosis). This conduit is then anastomosed to the ascending aortic graft (side-to-side anastomosis). (b, c) Oblique coronal (b) and oblique axial (c) CT angiograms of a 31-year-old man with Loeys-Dietz syndrome demonstrate the normal postoperative appearance of the Cabrol procedure. The retroaortic synthetic conduit graft (long arrow) and high-attenuating mechanical aortic valve are identifiable. The synthetic conduit graft is anastomosed to the left (short arrow) and right coronary arteries in an end-to-end fashion.
Valve-sparing Surgery
To avoid complications associated with mechanical aortic valve prostheses and the requirement for long-term anticoagulation, several techniques have since been developed that allow surgical repair of the aortic root with preservation of the patient’s native aortic valve (54). Two main categories of valve-sparing aortic root repair (V-SARR) include the remodeling technique proposed by Yacoub and the reimplantation technique proposed by David (55–57).
In the remodeling technique (Yacoub), the native aorta is resected down to the level of the valve insertions within the sinuses of Valsalva, leaving the native aortic valve intact. The sinuses are then reconstructed with a scalloped polyethylene terephthalate (Dacron) graft secured to the remnants of aortic tissue above the valve insertions (Fig 8) (55). In the reimplantation technique (David), the native aorta is also resected down to the valve insertions. However, the proximal hemostatic anastomosis of the polyethylene terephthalate graft is to the aortic annulus, below the level of the native aortic valve. A second proximal suture line along the edge of the resected tissue above the valve insertions is then used to resuspend the native aortic valve within the graft (Fig 9a) (51). The coronary arteries are reimplanted onto the aortic graft in both techniques. A modification of the reimplantation technique has been proposed, in which two separate grafts are used to individualize the dimensions of the aortic root and ascending aortic reconstruction (David-V, Stanford modification) (Figs 9b, 9c, and 10) (58,59). This modified technique provides the surgeon with the flexibility to individualize the dimensions and 3D geometry of the reconstruction, and creates neosinuses, mimicking the sinuses of Valsalva (60).
Figure 8a.

Yacoub V-SARR. (a) Illustration shows the Yacoub remodeling procedure, in which the native aorta is removed to the level of the valve insertions, leaving the native aortic valve intact, and a synthetic graft is attached above the native valve. The sinuses are reconstructed and secured above the attachment of the valve leaflets, and the coronary arteries are anastomosed to the ascending aortic graft. (b) Oblique coronal CT angiogram of a 32-year-old woman with Loeys-Dietz syndrome demonstrates the normal postoperative appearance of the Yacoub V-SARR procedure. The reimplanted left coronary artery (red arrow) and preserved native aortic valve (black arrow) are identified.
Figure 9a.

David V-SARR. (a) Illustration shows the David reimplantation procedure, in which the native aorta is removed to the valve insertions and a synthetic graft is attached below the native valve and sewn to the annulus. The native aortic valve is resuspended within the graft using a second suture line. (b) Illustration shows the David-V Stanford modification procedure, in which two separate grafts are used to individualize the dimensions and 3D geometry of the root reconstruction. (c) Intraoperative photograph depicting the David-V Stanford modification procedure showing two separate ascending aortic graft components and a reimplanted coronary artery button (arrow).
Figure 9b.

David V-SARR. (a) Illustration shows the David reimplantation procedure, in which the native aorta is removed to the valve insertions and a synthetic graft is attached below the native valve and sewn to the annulus. The native aortic valve is resuspended within the graft using a second suture line. (b) Illustration shows the David-V Stanford modification procedure, in which two separate grafts are used to individualize the dimensions and 3D geometry of the root reconstruction. (c) Intraoperative photograph depicting the David-V Stanford modification procedure showing two separate ascending aortic graft components and a reimplanted coronary artery button (arrow).
Figure 9c.
David V-SARR. (a) Illustration shows the David reimplantation procedure, in which the native aorta is removed to the valve insertions and a synthetic graft is attached below the native valve and sewn to the annulus. The native aortic valve is resuspended within the graft using a second suture line. (b) Illustration shows the David-V Stanford modification procedure, in which two separate grafts are used to individualize the dimensions and 3D geometry of the root reconstruction. (c) Intraoperative photograph depicting the David-V Stanford modification procedure showing two separate ascending aortic graft components and a reimplanted coronary artery button (arrow).
Figure 10a.

Normal postoperative appearance following David-V Stanford modification V-SARR procedure. (a) Oblique sagittal noncontrast CT image of a 34-year-old man with Marfan syndrome, after David-V Stanford modification V-SARR, demonstrating slightly hyperattenuating graft material that is easily visualized at noncontrast imaging. (b) Oblique sagittal contrast-enhanced CT angiogram demonstrates that the site of the graft component anastomosis can be inferred on contrast-enhanced CT angiograms by changes in contour (arrow).
Figure 8b.
Yacoub V-SARR. (a) Illustration shows the Yacoub remodeling procedure, in which the native aorta is removed to the level of the valve insertions, leaving the native aortic valve intact, and a synthetic graft is attached above the native valve. The sinuses are reconstructed and secured above the attachment of the valve leaflets, and the coronary arteries are anastomosed to the ascending aortic graft. (b) Oblique coronal CT angiogram of a 32-year-old woman with Loeys-Dietz syndrome demonstrates the normal postoperative appearance of the Yacoub V-SARR procedure. The reimplanted left coronary artery (red arrow) and preserved native aortic valve (black arrow) are identified.
Figure 10b.

Normal postoperative appearance following David-V Stanford modification V-SARR procedure. (a) Oblique sagittal noncontrast CT image of a 34-year-old man with Marfan syndrome, after David-V Stanford modification V-SARR, demonstrating slightly hyperattenuating graft material that is easily visualized at noncontrast imaging. (b) Oblique sagittal contrast-enhanced CT angiogram demonstrates that the site of the graft component anastomosis can be inferred on contrast-enhanced CT angiograms by changes in contour (arrow).
The most important long-term complication following V-SARR is aortic insufficiency (57). This complication occurs more commonly following the remodeling (Yacoub) technique, as the unsecured aortic annulus may continue to dilate below the graft, eventually resulting in prolapse and insufficiency of the aortic valve, potentially necessitating another surgical procedure. A major advantage of the reimplantation technique is that the graft is secured proximally to the annulus—similar to an annuloplasty, effectively preventing further dilation of the aortic annulus under the graft. The reimplantation technique is considered the preferred approach for valve-sparing surgery, particularly in patients with Marfan syndrome, given the exceptionally long durability of the graft (57,61).
Bioprosthetic Grafts
Several aortic root replacement techniques have been described using biologic grafts rather than synthetic material, including use of pulmonary autografts, human homografts, and porcine xenografts.
In the Ross procedure, the native aortic valve and aortic root are replaced with the patient’s own pulmonary valve and proximal pulmonary artery, with reimplantation of the coronary arteries onto the autograft. The pulmonary artery is replaced with a synthetic or bioprosthetic graft (Figs 11, 12) (62). The largest advantage of the Ross procedure is the potential for growth of the autograft in children requiring aortic root replacement. However, increasing aortic regurgitation and progressive dilation of the pulmonary autograft are the main limitations of this technique (Movie 3) (62,63). Periodic imaging follow-up is recommended postoperatively to assess for potential complications (62).
Figure 11.
Ross procedure. Illustration shows the Ross procedure, in which the native aortic valve and root are replaced with the patient’s own pulmonary valve and proximal pulmonary artery (autograft) with reimplantation of the coronary arteries. The proximal pulmonary artery is replaced with a synthetic or bioprosthetic valved conduit.
Figure 12a.

Postoperative imaging following Ross procedure in a 26-year-old man with a history of bicuspid aortic valve. (a) Coronal oblique CT angiogram, after Ross procedure 9 years earlier, demonstrates notable dilatation of the pulmonary autograft, which measured more than 7 cm at the sinuses of Valsalva. There was severe neoaortic regurgitation (not shown). (b) Oblique sagittal CT angiogram demonstrates calcification of the cadaveric homograft used to reconstruct the pulmonary artery.
Movie 3.
Four-dimensional flow imaging of a 32-year-old man after Ross procedure for an aortic root aneurysm. There is mild aortic insufficiency with stenosis and flow acceleration across the pulmonary homograft.
Figure 12b.

Postoperative imaging following Ross procedure in a 26-year-old man with a history of bicuspid aortic valve. (a) Coronal oblique CT angiogram, after Ross procedure 9 years earlier, demonstrates notable dilatation of the pulmonary autograft, which measured more than 7 cm at the sinuses of Valsalva. There was severe neoaortic regurgitation (not shown). (b) Oblique sagittal CT angiogram demonstrates calcification of the cadaveric homograft used to reconstruct the pulmonary artery.
Alternative bioprosthetic approaches include replacement of the aortic root with a human aortic homograft from cadaveric or human heart transplant recipient sources. Although aortic homografts have lower transvalvular gradients than do synthetic prostheses, they have limited durability and can undergo late degeneration (64). Use of xenograft bioprosthetic grafts, such as those using treated porcine roots, is an alternative approach with reported late survival rates similar to those after homograft root replacement (Figs 13, 14) (65).
Figure 13a.

Normal postoperative appearance of a 52-year-old woman with giant cell arteritis following replacement of the aortic root with a porcine xenograft. Polyethylene terephthalate grafts were used to replace the ascending aorta and the aortic arch, with an elephant trunk in the proximal descending thoracic aorta. (a) Oblique coronal CT angiogram demonstrates the coronary artery anastomoses (arrows) to the xenograft. (b) Axial CT angiogram demonstrates a hyperattenuating polytetrafluoroethylene felt ring, which was used to reinforce the proximal xenograft anastomosis.
Figure 14a.

Intra- and postoperative appearance following replacement of the aortic root with a porcine root xenograft in a 36-year-old man with congenital bicuspid aortic valve, single right coronary artery, and aortic root aneurysm. (a) Intraoperative photograph demonstrates the porcine root xenograft, which has been sutured in place, with reimplantation of the right coronary artery. Of note, the left coronary artery was reconstructed with a segment of autologous external iliac artery. (b) Postoperative oblique coronal volume-rendered CT angiogram demonstrates the postoperative appearance of the xenograft and reimplanted right coronary artery.
Figure 13b.

Normal postoperative appearance of a 52-year-old woman with giant cell arteritis following replacement of the aortic root with a porcine xenograft. Polyethylene terephthalate grafts were used to replace the ascending aorta and the aortic arch, with an elephant trunk in the proximal descending thoracic aorta. (a) Oblique coronal CT angiogram demonstrates the coronary artery anastomoses (arrows) to the xenograft. (b) Axial CT angiogram demonstrates a hyperattenuating polytetrafluoroethylene felt ring, which was used to reinforce the proximal xenograft anastomosis.
Figure 14b.

Intra- and postoperative appearance following replacement of the aortic root with a porcine root xenograft in a 36-year-old man with congenital bicuspid aortic valve, single right coronary artery, and aortic root aneurysm. (a) Intraoperative photograph demonstrates the porcine root xenograft, which has been sutured in place, with reimplantation of the right coronary artery. Of note, the left coronary artery was reconstructed with a segment of autologous external iliac artery. (b) Postoperative oblique coronal volume-rendered CT angiogram demonstrates the postoperative appearance of the xenograft and reimplanted right coronary artery.
Postoperative Complications and Pitfalls
Normal postoperative appearances following aortic root surgery may mimic those of pathologic conditions. Therefore, cardiovascular imagers must be familiar with common surgical procedures, their normal imaging appearances, and their complications, to interpret postoperative imaging studies accurately. Use of ECG gating is particularly important in the evaluation of patients who have undergone aortic root repair to minimize motion artifact, which could potentially obscure important findings (51). Evaluation of the postoperative aorta is frequently performed with CT angiography, particularly in the immediate postoperative period before hospital discharge. Pertinent clinical information, including surgical notes, should be reviewed, especially when evaluating the first baseline study after surgery. Knowledge of the patient’s preoperative anatomy, surgical procedure, and details of the coronary artery reimplantation technique are crucial in the proper evaluation and interpretation of postoperative imaging.
Postoperative Complications
The presence of contrast material external to the graft or native aortic lumen in the immediate postoperative period is most likely due to dehiscence, which can occur at any anastomosis or site of cannulation (66). Dehiscence may be related to surgical technique, often due to friable native aortic tissue such as in connective tissue diseases and vasculitides. A high index of suspicion should be maintained for infection, particularly in the setting of preoperative infection and in newly detected leaks at previously intact anastomoses.
A pseudoaneurysm is defined as a blood-filled space beyond the expected contours of an artery due to a partial or complete breach of the arterial wall, with a persistent communication to the bloodstream. Pathologically, a pseudoaneurysm may be contained by some, but not all, layers of the original arterial wall (for example, a traumatic aortic pseudoaneurysm may be contained by an intact adventitial layer), or it may be contained by surrounding tissues alone. In the surgical context, the term pseudoaneurym is applied to any extravascular or extragraft blood-perfused space typically arising from a site of anastomosis, cannulation, or arteriotomy, arising from an artery, graft, or the heart (for example, the infravalvular region of the left ventricular outflow tract) (Fig 15, Movie 4). Other complications that may be encountered following aortic root surgery include paravalvular leaks, valve insufficiency, and infection (Fig 16, Movie 5).
Figure 15.

Subannular pseudoaneurysm and paravalvular leak in a 19-year-old man with Marfan syndrome with a history of V-SARR. Oblique coronal volume-rendered CT angiogram demonstrates a small subannular pseudoaneurysm (white arrow) and a small paravalvular leak (red arrow).
Figure 16.

Crescent-shaped pseudoaneurysm in a patient after Ross procedure. Axial CT angiogram through the aortic root shows a small amount of extraluminal contrast material posterior to the aortic root only during systole (arrows), in keeping with a small subvalvular pseudoaneurysm.
Movie 4.
Coronal cine retrospectively gated volume-rendered CT angiographic images in a 19-year-old man with Marfan syndrome, after V-SARR, complicated by a small subannular pseudoaneurysm on the right and small paravalvular leak on the left.
Movie 5.
Axial cine retrospectively gated CT angiographic images through the aortic root demonstrate a small posterior subvalvular pseudoaneurysm in a patient after Ross procedure. The pseudoaneurysm only filled during systole, consistent with a subvalvular connection to the left ventricle.
Normal postoperative imaging findings seen following all types of open aortic root repair include perigraft fluid, soft-tissue stranding, and mediastinal air (66). Mediastinal air is often associated with surgical drain removal several days after the operation, and may persist for several weeks postoperatively. Findings that should raise suspicion for infection include a fluid collection with associated rim contrast enhancement or intrinsic air, as well as new or increasing perigraft air (66). Mild rim enhancement of fluid collections in the early postoperative period can be a normal finding, however, and often cannot be distinguished from low-grade infection in the acute and subacute postoperative period. Further investigation with biopsy, aspiration, re-do sternotomy, or a nuclear medicine white blood cell scan may be required for definitive diagnosis if a postoperative infection is suspected (Fig 17).
Figure 17a.

Postoperative infection in a 51-year-old man, after aortic valve replacement and arch repair, who presented with anterior chest pain and leukocytosis. (a) Axial CT angiogram demonstrates a rim-enhancing mediastinal fluid collection (arrow). (b) Technetium 99m white blood cell scan demonstrates abnormal uptake corresponding to the fluid collection (arrow) identified at CT, in keeping with postoperative infection, in addition to prosthetic valve infection.
Figure 17b.

Postoperative infection in a 51-year-old man, after aortic valve replacement and arch repair, who presented with anterior chest pain and leukocytosis. (a) Axial CT angiogram demonstrates a rim-enhancing mediastinal fluid collection (arrow). (b) Technetium 99m white blood cell scan demonstrates abnormal uptake corresponding to the fluid collection (arrow) identified at CT, in keeping with postoperative infection, in addition to prosthetic valve infection.
Postoperative Imaging Pitfalls
Performing nonenhanced imaging is particularly important in the postoperative setting, as surgical material may be most conspicuous at this phase. Synthetic grafts are usually composed of polyethylene terephthalate and are slightly hyperattenuating on noncontrast CT images relative to the native aortic wall, and are typically visualized as a thin, curvilinear hyperattenuating structure (Fig 18) (66). Graft material may be obscured on contrast-enhanced CT angiograms, as the attenuation of intraluminal contrast material exceeds that of the thin graft. However, the site of graft material may be inferred by changes in contour or visualization of hyperattenuating felt rings (typically made of polytetrafluorethylene [Teflon]) used to reinforce the site of anastomoses (Figs 10, 13) (66,67). Surgical grafts are usually uniform in diameter. However, the graft may occasionally fold on itself due to angulation or redundancy, or at the junction of two grafts, potentially mimicking a dissection flap (Fig 19). Historically—before the introduction of preclotted (fibrin-coated) graft material—the native aortic aneurysm sac was wrapped around the synthetic graft during surgery, which may appear unusual at postoperative imaging given the presence of an additional potential space between the graft and surrounding mediastinal tissue (Fig 20).
Figure 18a.

Normal postoperative appearance of graft material at CT. (a) Axial unenhanced CT image shows a synthetic polyethylene terephthalate aortic graft (arrow), which is usually conspicuous as a thin structure, slightly hyperattenuating relative to the aortic lumen on noncontrast CT images. (b) Axial contrast-enhanced CT angiogram demonstrates that graft material (arrow) is obscured by higher-attenuation intraluminal contrast material.
Figure 19a.

Postoperative appearance of folded graft material on CT images potentially mimicking aortic dissection. (a, b) Oblique coronal noncontrast CT image (a) and contrast-enhanced CT angiogram (b) demonstrate synthetic graft material (arrows) that is redundant and has folded in on itself, potentially mimicking the appearance of aortic dissection on the CT angiogram. However, hyperattenuating graft material can be confidently identified on the noncontrast image.
Figure 20.

Postoperative appearance of a small contained leak at the left coronary artery anastomosis in a 48-year-old man with Marfan syndrome, following Bentall composite valve graft procedure. Oblique coronal CT angiogram demonstrates a small contained leak (red arrow) at the left coronary artery anastomosis. The native aortic aneurysm sac (white arrows), which was wrapped around the graft during this procedure, is also identified.
Figure 18b.

Normal postoperative appearance of graft material at CT. (a) Axial unenhanced CT image shows a synthetic polyethylene terephthalate aortic graft (arrow), which is usually conspicuous as a thin structure, slightly hyperattenuating relative to the aortic lumen on noncontrast CT images. (b) Axial contrast-enhanced CT angiogram demonstrates that graft material (arrow) is obscured by higher-attenuation intraluminal contrast material.
Figure 19b.

Postoperative appearance of folded graft material on CT images potentially mimicking aortic dissection. (a, b) Oblique coronal noncontrast CT image (a) and contrast-enhanced CT angiogram (b) demonstrate synthetic graft material (arrows) that is redundant and has folded in on itself, potentially mimicking the appearance of aortic dissection on the CT angiogram. However, hyperattenuating graft material can be confidently identified on the noncontrast image.
Hyperattenuating felt pledgets may be mistaken for a pseudoaneurysm on contrast-enhanced CT angiograms, warranting careful evaluation of nonenhanced images to confirm the presence of the high-attenuation (polytetrafluoroethylene) surgical material (Fig 21). In comparison, pseudoaneurysms are usually iso- or hypoattenuating relative to surrounding tissue, and isoattenuating with the blood pool, on noncontrast CT images. Another potential mimic of a pseudoaneurysm relates to the coronary button technique, in which case a focal bulge of native tissue at the anastomotic site may be misinterpreted as a pathologic condition (66).
Figure 21a.

Postoperative appearance of a polytetrafluoroethylene felt pledget at CT potentially mimicking a pseudoaneurysm in a 93-year-old man after aortic valve replacement. (a) Axial CT angiogram demonstrates hyperattenuating felt material (arrow) at an arterial puncture site, which could potentially be mistaken for a small pseudoaneurysm. (b) Axial noncontrast CT image demonstrates that this represented a hyperattenuating felt pledget (arrow).
Figure 21b.

Postoperative appearance of a polytetrafluoroethylene felt pledget at CT potentially mimicking a pseudoaneurysm in a 93-year-old man after aortic valve replacement. (a) Axial CT angiogram demonstrates hyperattenuating felt material (arrow) at an arterial puncture site, which could potentially be mistaken for a small pseudoaneurysm. (b) Axial noncontrast CT image demonstrates that this represented a hyperattenuating felt pledget (arrow).
Bioabsorbable hemostatic agents such as gelatin (Gelfoam; Pfizer, New York, NY) or oxidized regenerated cellulose (Surgicel; Johnson & Johnson, New Brunswick, NJ) are used when intraoperative bleeding cannot be controlled by conventional means. This typically occurs in the setting of a re-operation. These agents are usually absorbed within 7–14 days, but if used extensively, may remain unabsorbed for a longer period of time. Before complete absorption, they may appear as an ill-defined, gas-filled heterogeneous mass, sometimes with rim enhancement, potentially mimicking the appearance of an abscess, hematoma, or retained foreign body (Fig 22) (68,69). Features that have been suggested to be more characteristic of absorbable hemostatic agents versus abscess include a linear arrangement of tightly packed gas bubbles that maintain their position on subsequent images and lack of an air-fluid level or an enhancing wall (68). BioGlue (Cryolife; Kennesaw, GA), a two-component adhesive incorporating glutaraldehyde and bovine serum albumin is used during cardiovascular surgery to improve hemostasis perioperatively and to strengthen and reinforce vascular anastomoses. BioGlue cannot be seen on CT images, but its use indicates a technically difficult surgical procedure and thus higher risk of anastomotic complications. BioGlue has also been reported to result in tissue necrosis, late anastomotic leaks, and pseudoaneurysm formation (70).
Figure 22.

Postoperative appearance of absorbable hemostatic sponge (arrow) at axial CT angiography, potentially mimicking infection in a 37-year-old man with Marfan syndrome, after aortic valve replacement and aortic root, ascending aorta, and hemiarch repair. Sponge is adjacent to the aortic root, containing multiple locules of air.
Conclusion
A variety of surgical techniques are used to repair the aortic root. Because normal postoperative appearances may mimic those of pathologic conditions, and because several postoperative complications exist, cardiovascular imagers must be familiar with these procedures and their complications. This article provides an overview of anatomy, imaging techniques, pathologic conditions, measurement thresholds, surgical approaches, and postoperative imaging of the aortic root.
Presented as an education exhibit at the 2014 RSNA Annual Meeting.
For this journal-based SA-CME activity, the authors F.P.C. and D.C.M. have provided disclosures (see “Disclosures of Conflicts of Interest”); all other authors, the editor, and the reviewers have disclosed no relevant relationships.
Funding: The work was supported by the National Registry of Genetically Triggered Thoracic Aortic Aneurysms and Cardiovascular Conditions (GenTAC) [D.C.M.].
Disclosures of Conflicts of Interest.—: F.P.C. Activities related to the present article: disclosed no relevant relationships. Activities not related to the present article: nonfinancial support from Siemens Healthcare. Other activities: disclosed no relevant relationships. D.C.M. Activities related to the present article: disclosed no relevant relationships. Activities not related to the present article: grants from Abbott Vascular, Edwards Lifesciences, and Medtronic; fees from Abbott Vascular, Medtronic, and RTI International; and nonfinancial support from Edwards Lifesciences. Other activities: disclosed no relevant relationships.
Abbreviations:
- 4D
- four-dimensional
- SSFP
- steady-state free precession
- TGFβ
- transforming growth factor β
- 3D
- three-dimensional
- 2D
- two-dimensional
- V-SARR
- valve-sparing aortic root repair
References
- 1.Anderson RH. Clinical anatomy of the aortic root. Heart 2000;84(6):670–673. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Sievers HH, Hemmer W, Beyersdorf F, et al. The everyday used nomenclature of the aortic root components: the tower of Babel? Eur J Cardiothorac Surg 2012;41(3):478–482. [DOI] [PubMed] [Google Scholar]
- 3.Anderson RH, Lal M, Ho SY. Anatomy of the aortic root with particular emphasis on options for its surgical enlargement. J Heart Valve Dis 1996;5(suppl 3):S249–S257. [PubMed] [Google Scholar]
- 4.Leipsic J, Gurvitch R, Labounty TM, et al. Multidetector computed tomography in transcatheter aortic valve implantation. JACC Cardiovasc Imaging 2011;4(4):416–429. [DOI] [PubMed] [Google Scholar]
- 5.Piazza N, de Jaegere P, Schultz C, Becker AE, Serruys PW, Anderson RH. Anatomy of the aortic valvar complex and its implications for transcatheter implantation of the aortic valve. Circ Cardiovasc Interv 2008;1(1):74–81. [DOI] [PubMed] [Google Scholar]
- 6.Charitos EI, Sievers HH. Anatomy of the aortic root: implications for valve-sparing surgery. Ann Cardiothorac Surg 2013;2(1):53–56. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Freeman LA, Young PM, Foley TA, Williamson EE, Bruce CJ, Greason KL. CT and MRI assessment of the aortic root and ascending aorta. AJR Am J Roentgenol 2013;200(6): W581–W592. [DOI] [PubMed] [Google Scholar]
- 8.Litmanovich D, Bankier AA, Cantin L, Raptopoulos V, Boiselle PM. CT and MRI in diseases of the aorta. AJR Am J Roentgenol 2009;193(4):928–940. [DOI] [PubMed] [Google Scholar]
- 9.Roos JE, Willmann JK, Weishaupt D, Lachat M, Marincek B, Hilfiker PR. Thoracic aorta: motion artifact reduction with retrospective and prospective electrocardiography-assisted multi-detector row CT. Radiology 2002;222(1):271–277. [DOI] [PubMed] [Google Scholar]
- 10.Wu W, Budovec J, Foley WD. Prospective and retrospective ECG gating for thoracic CT angiography: a comparative study. AJR Am J Roentgenol 2009;193(4):955–963. [DOI] [PubMed] [Google Scholar]
- 11.Blanke P, Bulla S, Baumann T, et al. Thoracic aorta: prospective electrocardiographically triggered CT angiography with dual-source CT—feasibility, image quality, and dose reduction. Radiology 2010;255(1):207–217. [DOI] [PubMed] [Google Scholar]
- 12.Paul JF, Abada HT. Strategies for reduction of radiation dose in cardiac multislice CT. Eur Radiol 2007;17(8): 2028–2037. [DOI] [PubMed] [Google Scholar]
- 13.Boxerman JL, Mosher TJ, McVeigh ER, Atalar E, Lima JA, Bluemke DA. Advanced MR imaging techniques for evaluation of the heart and great vessels. RadioGraphics 1998;18(3):543–564. [DOI] [PubMed] [Google Scholar]
- 14.Nayak KS, Hargreaves BA, Hu BS, Nishimura DG, Pauly JM, Meyer CH. Spiral balanced steady-state free precession cardiac imaging. Magn Reson Med 2005;53(6): 1468–1473. [DOI] [PubMed] [Google Scholar]
- 15.Vasanawala SS, Hanneman K, Alley MT, Hsiao A. Congenital heart disease assessment with 4D flow MRI. J Magn Reson Imaging 2015 Feb 24. [Epub ahead of print] [DOI] [PubMed] [Google Scholar]
- 16.Gruber PJ, Epstein JA. Development gone awry: congenital heart disease. Circ Res 2004;94(3):273–283. [DOI] [PubMed] [Google Scholar]
- 17.Prince MR, Grist TM, Debatin JF. 3D contrast MR angiography. 2nd ed. Berlin, Germany: Springer-Verlag, 2003; 11–16. [Google Scholar]
- 18.Foo TK, Saranathan M, Prince MR, Chenevert TL. Automated detection of bolus arrival and initiation of data acquisition in fast, three-dimensional, gadolinium-enhanced MR angiography. Radiology 1997;203(1):275–280. [DOI] [PubMed] [Google Scholar]
- 19.Lohan DG, Krishnam M, Tomasian A, Saleh R, Finn JP. Time-resolved MR angiography of the thorax. Magn Reson Imaging Clin N Am 2008;16(2):235–248, viii. [DOI] [PubMed] [Google Scholar]
- 20.Rubin GD, Dake MD, Semba CP. Current status of three-dimensional spiral CT scanning for imaging the vasculature. Radiol Clin North Am 1995;33(1):51–70. [PubMed] [Google Scholar]
- 21.Burman ED, Keegan J, Kilner PJ. Aortic root measurement by cardiovascular magnetic resonance: specification of planes and lines of measurement and corresponding normal values. Circ Cardiovasc Imaging 2008;1(2):104–113. [DOI] [PubMed] [Google Scholar]
- 22.Isselbacher EM. Thoracic and abdominal aortic aneurysms. Circulation 2005;111(6):816–828. [DOI] [PubMed] [Google Scholar]
- 23.Cañadas V, Vilacosta I, Bruna I, Fuster V. Marfan syndrome. I. Pathophysiology and diagnosis. Nat Rev Cardiol 2010;7(5):256–265. [DOI] [PubMed] [Google Scholar]
- 24.Beroukhim RS, Roosevelt G, Yetman AT. Comparison of the pattern of aortic dilation in children with the Marfan’s syndrome versus children with a bicuspid aortic valve. Am J Cardiol 2006;98(8):1094–1095. [DOI] [PubMed] [Google Scholar]
- 25.Milewicz DM, Dietz HC, Miller DC. Treatment of aortic disease in patients with Marfan syndrome. Circulation 2005;111(11):e150–e157. [DOI] [PubMed] [Google Scholar]
- 26.van Karnebeek CD, Naeff MS, Mulder BJ, Hennekam RC, Offringa M. Natural history of cardiovascular manifestations in Marfan syndrome. Arch Dis Child 2001;84(2): 129–137. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Loeys BL, Dietz HC, Braverman AC, et al. The revised Ghent nosology for the Marfan syndrome. J Med Genet 2010;47(7):476–485. [DOI] [PubMed] [Google Scholar]
- 28.Tadros TM, Klein MD, Shapira OM. Ascending aortic dilatation associated with bicuspid aortic valve: pathophysiology, molecular biology, and clinical implications. Circulation 2009;119(6):880–890. [DOI] [PubMed] [Google Scholar]
- 29.Sievers HH, Schmidtke C. A classification system for the bicuspid aortic valve from 304 surgical specimens. J Thorac Cardiovasc Surg 2007;133(5):1226–1233. [DOI] [PubMed] [Google Scholar]
- 30.Nkomo VT, Enriquez-Sarano M, Ammash NM, et al. Bicuspid aortic valve associated with aortic dilatation: a community-based study. Arterioscler Thromb Vasc Biol 2003;23(2):351–356. [DOI] [PubMed] [Google Scholar]
- 31.Siu SC, Silversides CK. Bicuspid aortic valve disease. J Am Coll Cardiol 2010;55(25):2789–2800. [DOI] [PubMed] [Google Scholar]
- 32.Van Hemelrijk C, Renard M, Loeys B. The Loeys-Dietz syndrome: an update for the clinician. Curr Opin Cardiol 2010;25(6):546–551. [DOI] [PubMed] [Google Scholar]
- 33.Zilocchi M, Macedo TA, Oderich GS, Vrtiska TJ, Biondetti PR, Stanson AW. Vascular Ehlers-Danlos syndrome: imaging findings. AJR Am J Roentgenol 2007;189(3):712–719. [DOI] [PubMed] [Google Scholar]
- 34.Oderich GS, Panneton JM, Bower TC, et al. The spectrum, management and clinical outcome of Ehlers-Danlos syndrome type IV: a 30-year experience. J Vasc Surg 2005;42(1): 98–106. [DOI] [PubMed] [Google Scholar]
- 35.Elsheikh M, Casadei B, Conway GS, Wass JA. Hypertension is a major risk factor for aortic root dilatation in women with Turner’s syndrome. Clin Endocrinol (Oxf) 2001;54(1):69–73. [DOI] [PubMed] [Google Scholar]
- 36.Hjerrild BE, Mortensen KH, Sørensen KE, et al. Thoracic aortopathy in Turner syndrome and the influence of bicuspid aortic valves and blood pressure: a CMR study. J Cardiovasc Magn Reson 2010;12:12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Kim HK, Gottliebson W, Hor K, et al. Cardiovascular anomalies in Turner syndrome: spectrum, prevalence, and cardiac MRI findings in a pediatric and young adult population. AJR Am J Roentgenol 2011;196(2):454–460. [DOI] [PubMed] [Google Scholar]
- 38.Coady MA, Davies RR, Roberts M, et al. Familial patterns of thoracic aortic aneurysms. Arch Surg 1999;134(4): 361–367. [DOI] [PubMed] [Google Scholar]
- 39.Erbel R, Eggebrecht H. Aortic dimensions and the risk of dissection. Heart 2006;92(1):137–142. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Hager A, Kaemmerer H, Rapp-Bernhardt U, et al. Diameters of the thoracic aorta throughout life as measured with helical computed tomography. J Thorac Cardiovasc Surg 2002;123(6):1060–1066. [DOI] [PubMed] [Google Scholar]
- 41.Elefteriades JA. Indications for aortic replacement. J Thorac Cardiovasc Surg 2010;140(6 suppl):S5–S9; discussion S45–S51. [DOI] [PubMed] [Google Scholar]
- 42.Elefteriades JA. Natural history of thoracic aortic aneurysms: indications for surgery, and surgical versus nonsurgical risks. Ann Thorac Surg 2002;74(5):S1877–S1880; discussion S1892–S1898. [DOI] [PubMed] [Google Scholar]
- 43.Coady MA, Rizzo JA, Hammond GL, et al. What is the appropriate size criterion for resection of thoracic aortic aneurysms? J Thorac Cardiovasc Surg 1997;113(3):476–491; discussion 489–491. [DOI] [PubMed] [Google Scholar]
- 44.Hiratzka LF, Bakris GL, Beckman JA, et al. 2010 ACCF/AHA/AATS/ACR/ASA/SCA/SCAI/SIR/STS/SVM guidelines for the diagnosis and management of patients with thoracic aortic disease: executive summary. A report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines, American Association for Thoracic Surgery, American College of Radiology, American Stroke Association, Society of Cardiovascular Anesthesiologists, Society for Cardiovascular Angiography and Interventions, Society of Interventional Radiology, Society of Thoracic Surgeons, and Society for Vascular Medicine. Catheter Cardiovasc Interv 2010;76(2):E43–E86. [DOI] [PubMed] [Google Scholar]
- 45.Williams JA, Loeys BL, Nwakanma LU, et al. Early surgical experience with Loeys-Dietz: a new syndrome of aggressive thoracic aortic aneurysm disease. Ann Thorac Surg 2007;83(2):S757–S763; discussion S785–S790. [DOI] [PubMed] [Google Scholar]
- 46.Bentall H, De Bono A. A technique for complete replacement of the ascending aorta. Thorax 1968;23(4):338–339. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47.Cherry C, DeBord S, Hickey C. The modified Bentall procedure for aortic root replacement. AORN J 2006;84(1):52–55, 58–70; quiz 71–74. [DOI] [PubMed] [Google Scholar]
- 48.Kouchoukos NT, Marshall WGJ, Jr, Wedige-Stecher TA. Eleven-year experience with composite graft replacement of the ascending aorta and aortic valve. J Thorac Cardiovasc Surg 1986;92(4):691–705. [PubMed] [Google Scholar]
- 49.Cabrol C, Pavie A, Gandjbakhch I, et al. Complete replacement of the ascending aorta with reimplantation of the coronary arteries: new surgical approach. J Thorac Cardiovasc Surg 1981;81(2):309–315. [PubMed] [Google Scholar]
- 50.Gelsomino S, Frassani R, Da Col P, et al. A long-term experience with the Cabrol root replacement technique for the management of ascending aortic aneurysms and dissections. Ann Thorac Surg 2003;75(1):126–131. [DOI] [PubMed] [Google Scholar]
- 51.Prescott-Focht JA, Martinez-Jimenez S, Hurwitz LM, et al. Ascending thoracic aorta: postoperative imaging evaluation. RadioGraphics 2013;33(1):73–85. [DOI] [PubMed] [Google Scholar]
- 52.Kruser TJ, Osaki S, Kohmoto T, Chopra PS. Computed tomography finding mimicking aortic dissection after Cabrol procedure. Asian Cardiovasc Thorac Ann 2009;17(1): 108–109. [DOI] [PubMed] [Google Scholar]
- 53.Milano AD, Pratali S, Mecozzi G, et al. Fate of coronary ostial anastomoses after the modified Bentall procedure. Ann Thorac Surg 2003;75(6):1797–1801; discussion 1802. [DOI] [PubMed] [Google Scholar]
- 54.David TE, Ivanov J, Armstrong S, Feindel CM, Webb GD. Aortic valve-sparing operations in patients with aneurysms of the aortic root or ascending aorta. Ann Thorac Surg 2002;74(5):S1758–S1761; discussion S1792–S1799. [DOI] [PubMed] [Google Scholar]
- 55.Sarsam MA, Yacoub M. Remodeling of the aortic valve anulus. J Thorac Cardiovasc Surg 1993;105(3):435–438. [PubMed] [Google Scholar]
- 56.David TE, Feindel CM. An aortic valve-sparing operation for patients with aortic incompetence and aneurysm of the ascending aorta. J Thorac Cardiovasc Surg 1992;103(4):617–621; discussion 622. [PubMed] [Google Scholar]
- 57.Tian D, Rahnavardi M, Yan TD. Aortic valve sparing operations in aortic root aneurysms: remodeling or reimplantation? Ann Cardiothorac Surg 2013;2(1):44–52. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 58.Miller DC. Valve-sparing aortic root replacement in patients with the Marfan syndrome. J Thorac Cardiovasc Surg 2003;125(4):773–778. [DOI] [PubMed] [Google Scholar]
- 59.Demers P, Miller DC. Simple modification of “T. David-V” valve-sparing aortic root replacement to create graft pseudosinuses. Ann Thorac Surg 2004;78(4):1479–1481. [DOI] [PubMed] [Google Scholar]
- 60.Fleischmann D, Liang DH, Mitchell RS, Miller DC. Pre- and postoperative imaging of the aortic root for valve-sparing aortic root repair (V-SARR). Semin Thorac Cardiovasc Surg 2008;20(4):365–373. [DOI] [PubMed] [Google Scholar]
- 61.Cameron DE, Alejo DE, Patel ND, et al. Aortic root replacement in 372 Marfan patients: evolution of operative repair over 30 years. Ann Thorac Surg 2009;87(5):1344–1349; discussion 1349–1350. [DOI] [PubMed] [Google Scholar]
- 62.Kouchoukos NT, Masetti P, Nickerson NJ, Castner CF, Shannon WD, Dávila-Román VG. The Ross procedure: long-term clinical and echocardiographic follow-up. Ann Thorac Surg 2004;78(3):773–781; discussion 773–781. [DOI] [PubMed] [Google Scholar]
- 63.Frigiola A, Ranucci M, Carlucci C, Giamberti A, Abella R, Di Donato M. The Ross procedure in adults: long-term follow-up and echocardiographic changes leading to pulmonary autograft reoperation. Ann Thorac Surg 2008;86(2): 482–489. [DOI] [PubMed] [Google Scholar]
- 64.Lim E, Ali A, Theodorou P, et al. Longitudinal study of the profile and predictors of left ventricular mass regression after stentless aortic valve replacement. Ann Thorac Surg 2008;85(6):2026–2029. [DOI] [PubMed] [Google Scholar]
- 65.El-Hamamsy I, Clark L, Stevens LM, et al. Late outcomes following freestyle versus homograft aortic root replacement: results from a prospective randomized trial. J Am Coll Cardiol 2010;55(4):368–376. [DOI] [PubMed] [Google Scholar]
- 66.Hoang JK, Martinez S, Hurwitz LM. MDCT angiography after open thoracic aortic surgery: pearls and pitfalls. AJR Am J Roentgenol 2009;192(1):W20–W27. [DOI] [PubMed] [Google Scholar]
- 67.Quint LE, Francis IR, Williams DM, Monaghan HM, Deeb GM. Synthetic interposition grafts of the thoracic aorta: postoperative appearance on serial CT studies. Radiology 1999;211(2):317–324. [DOI] [PubMed] [Google Scholar]
- 68.Sandrasegaran K, Lall C, Rajesh A, Maglinte DT. Distinguishing gelatin bioabsorbable sponge and postoperative abdominal abscess on CT. AJR Am J Roentgenol 2005;184(2):475–480. [DOI] [PubMed] [Google Scholar]
- 69.Gayer G, Petrovitch I, Jeffrey RB. Foreign objects encountered in the abdominal cavity at CT. RadioGraphics 2011;31(2):409–428. [DOI] [PubMed] [Google Scholar]
- 70.Luk A, David TE, Butany J. Complications of Bioglue postsurgery for aortic dissections and aortic valve replacement. J Clin Pathol 2012;65(11):1008–1012. [DOI] [PubMed] [Google Scholar]









