|
 
 |
| ORIGINAL ARTICLE |
|
| Year : 2016 | Volume
: 2
| Issue : 1 | Page : 22-29 |
|
Computer reconstruction of the cardiac skeleton and its application in locating heart valve planes
Ying Li1, Wei Chen2, Yonglin Chen2, Kaijun Liu1, Liwen Tan1, Shaoxiang Zhang1
1 Institute of Digital Medicine, Biomedical Engineering College, Third Military Medical University, Chongqing 400038, China
2 Department of Dermatology Radiology, Southwest Hospital, Third Military Medical University, Chongqing 400038, China
| Date of Web Publication | 11-May-2016 |
Correspondence Address:
Liwen Tan
Institute of Digital Medicine, Third Military Medical University, Chongqing 400038
China
Shaoxiang Zhang
Institute of Digital Medicine,Third Military Medical University, Chongqing 400038
China
 Source of Support: None, Conflict of Interest: None  | Check |
DOI: 10.4103/2226-8561.182298
Objective: To fully understand the original spatial position and three-dimensional (3D) anatomical morphology of cardiac skeleton (CS), and to quickly locate its position in patient-specific computed tomography angiography (CTA) images. Materials and Methods: First, we segmented and reconstructed 3D models of CS and its attached valves with Amira software, defined valve planes based on Chinese visible human 5 (CVH5), and then computed its geometric transformation matrix and applied them in locating the valve planes in patient-specific CTA images. Results: We reconstructed a 3D CS model based on CVH5 images which keep the original spatial position and its normal anatomical appearance. The 3D structures include aortic valve annulus (AVA), mitral valve annulus, tricuspid valve annulus, pulmonary valve annulus, and its attached valves. With the relative geometric transformation matrix, we quickly located the patient-specific valve planes that are vertical to each valve in CTA images. Conclusions: CVH5 dataset can be used in reconstructing the 3D model of CS, which is difficult for clinical images, such as CT, magnetic resonance imaging, and traditional anatomical method to achieve. Our method of 3D reconstruction presents more anatomical details than clinical images and keeps the original shape and position. We can define each valve plane on the CVH5 model and show its corresponding plane in patient-specific CTA images, which can be observed on each valve plane at the same time based on the consistent reference. Keywords: Chinese visible human, computed tomography angiography, locating patient.specific heart valve planes, model of cardiac skeleton, three-dimensional
How to cite this article:
Li Y, Chen W, Chen Y, Liu K, Tan L, Zhang S. Computer reconstruction of the cardiac skeleton and its application in locating heart valve planes. Digit Med 2016;2:22-9 |
How to cite this URL:
Li Y, Chen W, Chen Y, Liu K, Tan L, Zhang S. Computer reconstruction of the cardiac skeleton and its application in locating heart valve planes. Digit Med [serial online] 2016 [cited 2023 Mar 29];2:22-9. Available from: http://www.digitmedicine.com/text.asp?2016/2/1/22/182298 |
| Introduction | |  |
The cardiac skeleton (CS), also known as the fibrous skeleton of the heart, is composed of high-density connective tissue that forms and anchors the valves and influences the forces exerted through them. The cardiac valve leaflet and the myocardium are attached to the CS, and the atria are separated and partitioned from the ventricles.[1] The detailed knowledge of three-dimensional (3D) morphology and dynamics has a significant impact on the diagnosis, as well as the understanding of the mechanisms underlying heart movement and the blood hydrodynamics between chambers. Both animal and human studies using different techniques including anatomy or imaging modalities have already investigated the morphology and dynamics of the mitral valve (MV) and tricuspid valve (TV), whereas the CS has been less studied. The main limitation of current imaging technologies is that the resolution of images is not fine enough to identify its anatomical details.
The Chinese visible human (CVH) project was initiated at the Third Military Medical University, China, in October 2002, as a means of presenting abundant anatomical information based on high-resolution, cross-sectional slice images gained from cadavers (Zhang et al. 2006). The advantages of this dataset are that it cannot only help observing CS clearly (with its natural color) but also maintain the original spatial relationship between anatomical structures.
The purposes of this study were to reconstruct a 3D computer model of CS identified by high-resolution slices of CVH and to quickly locate the patient-specific cardiac planes. First, we utilized the space relationship of fibrous annulus to design valve planes, and then we registered a 3D CVH model of aortic root to a patient-specific computed tomography angiography (CTA) 3D model and applied the same geometric transformation matrix to quickly locate patient-specific cardiac planes.
| Materials and Methods | | |
Imaging acquisition and preprocessing
Acquisition of CVH images
In our study, images of the heart region generated from a 22-year-old, young, healthy female from CVH5 were selected. The image resolution of a section of 780 cryo-sectional slices of the heart region was 940 *870 pixels per slice. The pixel resolution was 0.12 mm *0.12 mm and the slice interval was 0.2 mm.
Preprocessing of CVH images
CS and its adjacent structures were labeled based on CVH dataset, including aortic valve (AV), MV, TV, pulmonary valve (PV), right fibrous annular, left fibrous annular, left coronary artery (LCA), right coronary artery (RCA), aortic root (AR), vena cava, pulmonary artery, pulmonary valve, left atrium, right atrium, left myocardium, left ventricle (LV), right ventricle, AR, LCA, RCA, etc. These anatomical structures were labeled with different color values manually by experienced anatomists and radiologists using the Amira software (). Each color value we defined represented a unique anatomical structure. Then, we selected Compute → SurfaceGen on labeled to generate 3D surface models of CS.
Acquisition of computed tomography angiography images
Data acquisition parameters for CT angiography were as follows: Collimation of 0.6 mm, rotation time of 330 ms, tube voltage of 120 kV, and tube current of 400 m. A contrast-enhanced volume data set was acquired with retrospective electrocardiogram (ECG) gating to allow reconstructions during all phases of the cardiac cycle. Transaxial images were reconstructed with a section thickness of 0.75 mm, the increment of 0.4 mm, and a medium-soft convolution kernel (B26f). The position of the reconstruction window in the cardiac cycle was individually selected to minimize artifacts. The resolution of image was 512 *512 pixels per slice, and the pixel resolution was 0.12 mm *0.12 mm. Original sequence slices were used for generating cardiac plane on specific anatomical landmark.
Preprocessing of computed tomography angiography images
Coronary artery tree was segmented by threshold and region growing method in three steps: (1) T1 is a threshold used for presegmenting the region of high CT HU values, which is the maximal value on the histogram of whole CT HU values, (2) the connections between AR and LV were cutoff, and (3) region growing method was used on the aorta part. The model was used for the registration on the coronary artery tree model of CVH5.
Locating patient-specific cardiac planes in computed tomography angiography images based on the cardiac skeleton model of CVH 5
To quickly locate the patient-specific cardiac valve planes in CTA images, we defined the cardiac planes on a CVH5 model, gained geometric transformation matrix by registration on the 3D model of AR from CVH5 to CTA, and then applied the same geometric transformation matrix to all CVH 5 anatomical structures and the cardiac planes defined on CVH5, based on the CS relative spatial relationship between AR and CS. Finally, we gained patient-specific valve planes in CTA images, which were perpendicular to each valve of CS. The flowchart is shown in [Figure 1]. | Figure 1: The flowchart of locating patient-specific cardiac planes based on the cardiac skeleton three-dimensional model of CVH 5
Click here to view |
Point coordinates of gravity center
All local gravity center coordinates of valves in CVH5 follow this parameter equation: AX + bY + cZ + d = 0, (X, Y, Z/xi, yi, zi, i = 1, 2, 3, 4), in which i = 1, 2, 3, and 4 represents the X, Y, and Z axes of region on AV, MV, TV, and PV, respectively. The maximal intensity projection of the valve ring of CS is shown in [Figure 2], when z = 0, y = 0, and x = 0, respectively. Four local center axes of the region were calculated for each image, which was substituted into equation system of solved value of a, b, c, and d, and each region axis of x, y, and z in 3D space.
Axis definition of valves
Short-axis (SAX) images, which correspond to the plane through the central point and have the shortest distance between two points at the intersection of the region; long-axis (LAX) images, which correspond to the plane through the central point and have the longest distance between two points that are at intersection of the region. The gravity center and axes of valves are shown in [Figure 3]. | Figure 3: The gravity center and axes of valves. (a,b,c,d) show the gravity center and axes of aortic valve annulus (AVA), mitral valve annulus (MVA), tricuspid valve annulus(TVA), and pulmonary valve annulus (PVA), respectively
Click here to view |
Plane definition of valves
Valve planes, which correspond to the plane through the central point and any two points in each annular region.
Cardiac plane definition based on the cardiac skeleton
In this study, we defined the cardiac planes corresponding to four valves as an example to show how to use CS to guide the location on a specific plane. There are ventriculoaortic junction (VAJ), sinotubular junction (STJ), SAX and LAX of MV, SAX and LAX of TV, and SAX and LAX of PV.
Registration
We took the gravity center, STJ, VAJ of AR as the control point set which determines the geometric position of images, and then derived a geometric transformation matrix T for registering CVH to patient-specific CTA coronary tree model from those points.
Geometric transformation
We gained geometric transformation matrix G by the following formula:
f' (x', y', z') = f (x, y, x) *G(1)
in which f (x, y, z)and f' (x', y', z') are vectors consisting of the points of AV on a 3D model of CVH5 and CTA, respectively, with x, y, z and x', y', z' representing the coordinates on the axis of x, y, z before and after geometric transformation, respectively. G is a geometric transformation matrix: G = (S, T, R) in which S represents the scale factor matrix and
, T is a translation matrix and
and dx, dy, dz represent the distance of translation on the axis of X, Y, and Z, respectively. R is a the vector of rotation, and R = (Rx(θ), Ry(θ), Rz(θ)); Rx(θ), Rz(θ), Ry(θ) reflect the angles, rotated by the axes of X, Y, and Z, respectively.
Then, coordinates were calculated after geometric transformation based on the following formula: F' (x', y', z') = F (x, y, z)* G, f (x, y, z) ε F (x, y, z) F (x, y, z) consists of all 3D coordinates of CVH5 models and planes that we defined in section 2.2, and F' (x', y', z') are these coordinates transformed to CTA image space.
| Results | | |
Computer reconstruction of fibrous skeleton on CVH 5 heart
CS is surrounded by the valves of the heart, fuses with one another, and merges with the interventricular septum. The spatial relationship between CS and left and right ventricles, left and right atriums is important. The cardiac geometric shape is divided by CS into four chambers (4ch). In other words, once we gain position of CS, the interface of inflow-outflow can be located.
The irregular morphology of the cardiac base was characterized by the following aspects. (1) The centrally located fibrous skeleton represents the anatomical element with which the other cardiac structures share fibrous continuity [Figure 4]a; (2) the four valve orifices are not on the same plane, which is more obvious in 3D space than in 2D space, and their inflow-outflow axes are not paralleled [Figure 4]b. In an apical-to-basal direction, TV is the most apical or the most inferior of the four valves, followed by the mitral, aortic, and pulmonary valves. MV is the most posterior while PV is the most anterior [Figure 4]b, [Figure 4]c. (3) Each of the four valves takes complex anatomical shape. AV [Figure 5]a takes a 3-pronged crown shape, MV [Figure 5]b and TV [Figure 5]c take a semicircular shape, and the shape of PV [Figure 5]d is similar to that of AV, only with thinner leaflets due to lower pressure and it has no arteries arising from its sinuses. The annulus of each valve attached to CS is the corresponding annulus. There are aortic valve annulus (AVA), MV annulus (MVA), TV annulus (TVA), and pulmonary valve annulus (PVA), respectively. CS consists of four annuluses that hold the shape and spatial position of valves. (4) The cardiac geometric shape is divided by CS into 4ch. Once the position of CS is gained, the interfaces of inflow-outflow will be located. | Figure 4: The spatial position of cardiac skeleton in heart and its three-dimensional shape. a is the front view of whole heart; b is the back view of the heart (the spatial relationship of CS and four chambers); c is the top view of CS; d is the annulus 3D shape of fibrous skeleton
Click here to view |
 | Figure 5: The 3D models of valves. a,b,c,d are AV,MV,TV,and PV respectively
Click here to view |
Registering the aortic root of CVH and patient-specific computed tomography angiography model
The AR is a complex functional unit supporting the leaflets of AV and giving origin to the coronary arteries. The AR morphology determines the direction of the flows that are essential for the AV function and coronary perfusion. The aortic annulus is an important part of the supporting frames of the heart. As all cardiac surgery should refer to the AR, the AR chosen as the topography reference is very important for 3D model registration. Once the apex position of three-coronet shaped VAJ was fixed according to the relative geometric relationship of CS, each valve plane was generated.
It should be noted that we do not need to match all points of aortic annulus of CVH on a patient-specific CTA model as the size of AR may have individual difference but have similar topography, namely geometrical invariability. We just need to register the gravity center of the CVH model to that of the individual CTA model, and register the three apices of 3-pronged crown shape on AR; then, the orientation of AR can be determined.
The difference between standard planes, cardiac planes, and valve planes
Based on the shape of the heart [Figure 6]a, we can see that the right and the left sides of atria and ventricles are not always perpendicular to each other. Therefore, this plane imaging method might lead to artificial error in quantitatively evaluating morphological parameters on orthogonal planes. The standard planes are perpendicular to original image spaces [Figure 6]b. The cardiac planes are perpendicular to SAX and LAX of heart orientation [Figure 6]c. Valve planes differ from cardiac planes in that they are not perpendicular to the plane of SAX and LAX of the heart but perpendicular to each valve plane of the fibrous skeleton [Figure 6]d. | Figure 6: The shape of heart and the difference between standard, cardiac, and valve planes. Standard body planes (Figure 6b), three orthogonal cuts in axial transverse, coronal, and sagittal orientations. Heart planes (Figure 6c) Short-axis (SAX) images, Horizontal long-axis (four-chamber, 4ch), Vertical long-axis (two-chamber, 2ch). Valve planes (Figure 6d) generated by our method, which are perpendicular to each valve
Click here to view |
The image of patient-specific valve planes
In a 3D view, as shown in the first column of [Figure 7], the spatial relationship of CS with coronary tree can be observed directly so that the orientation vertical to valves is directly confirmed. Once we got the orientation of the valve plane, we can generate the section views in CVH5 images [shown in the second column of [Figure 7] and in patient-specific CTA images [shown in the third column of [Figure 7], respectively, according to its relative geometric transformation mentioned in 2.2. Namely, we used a 3D model of CS to build the connection between the high-resolution CVH5 images and CTA images. There are some anatomical details that we can easily identify in CVH5 images and models, which are difficult to identify in CTA images, especially some anatomy structures of complex and irregular shape, such as PV. Using our models and its valve planes, we can quickly locate its position and view the section views in patient-specific CTA images to observe its morphological features. The advantage of our method is that the section views of CVH5 and CTA are synchronized, and the position and orientation are consistent, which will help to deepen our understanding of the anatomical details of patient-specific CTA images. | Figure 7: Valve planes in a three-dimensional view, in CHV 5 images, and in patient-specific computed tomography angiography images. The first column are these planes that we defined in a 3D view, the second and third columns are its section views in CVH5 images and the results of valve planes in patient-specific CTA images respectively; A1)-A2) are the planes vertical to PVA and AVA; A3)-A4) are the planes vertical to short-axis and long-axis of TVA respectively; A5)-A6) are the planes vertical to short-axis and long-axis of MVA respectively
Click here to view |
| Discussion | | |
Cardiac skeleton morphological features and the feasibility of its application in locating cardiac planes
A complete understanding of CS morphology and motion has the potential to enhance pathophysiological knowledge and in turn improve diagnosis and surgical treatment. The application of 3D computer reconstruction in the virtual simulation can bridge morphology and function. The simulation is very difficult because of the complex shape of CS and the position of the heart. Conventionally, dissection is the most direct way to gain the natural shape of CS. Due to the deep position of the heart, CS would be exposed after destroying its surround structures and original spatial relationship. A study [2] hold that the aortic, pulmonary, mitral, and TVs are positioned on a plane, the so-called 'base' of the heart. However, our results show that the four valves have their own orientation and are not completely positioned on a plane. The complex and irregular shape of CS is shown in [Figure 4]d. It is a challenging job in clinical imaging to reconstruct its 3D shape, which depends on doctor's experience and their familiarity with its anatomy. Normal or abnormal morphology of 2D clinical images, their changes during cardiac cycle and 3D adjacent relationship about CS should be understood because they play important roles in diagnosing and evaluating the cardiac function. Some studies quantitatively studied the morphology of MV [3],[4],[5] and other studies quantitatively researched the TV morphology [6],[7] and its effects on functions.[8] In current clinic imaging, it is very difficult to identify the CS directly, and few studies have analyzed CS in clinical images. Maffessanti et al. located MVA and TVA by manually drawing annular points by inferring from its anatomical relationship between valves on each plane, and then connecting a ring to quantitatively evaluation their morphology and dynamics. This is the first study to our knowledge to quantify the same subject in terms of both MVA and TVA morphology and motion in magnetic resonance imaging. How about the morphological regulation and motion between four valves? CTA may be a potential noninvasive imaging technique to monitor the motion between four valves. Cross-sectional imaging data of CTA are acquired, which contained high-quality anatomical information about the valves apparatus.[9],[10],[11],[12],[13],[14],[15] Retrospective synchronization of the ECG-gating allows for the reconstruction of data sets at various systolic or diastolic phases throughout the cardiac cycle.[16] Images also can be generated to show the open or closed valves in respective part of the cycle.[17] In addition, CTA can show the whole coronary artery tree with well-temporal spatial resolution, so it has the potential to refine both the assessment of valve dysfunction and the resultant hemodynamic effects on surrounding cardiac structures.[18] The main limitations of this study are listed as follows: (1) It is time-consuming to locate the ideal cardiac planes, especially for TV and PV. (2) Because of the complex and irregular shape of valves during motion, if each valve is just independently measured without a consistent reference, whole errors may occur. CS is naturally connected to four valves, but it is also hard to identify CTA imaging. In our study, we reconstructed CS 3D shape in original position of heart by CVH, including MVA, TVA, AVA, and PVA and applied the spatial relationship to quickly locate the patient-specific four valve planes, which has the potential to monitor morphological changes of the four valves at the same time and on a consistent reference. Besides, we also built the connection between CVH slice images and patient-specific CTA slice images, which helps the understanding of patient-specific valve planes on the same reference of CVH.
For the framework function of CS, which is close to most anatomical structure within the heart, CVH5 anatomical model can identify anatomical details more clearly than clinical images. Therefore, the locating method can also be used in designing cardiac interior anatomical landmark plane and show its relative position in patient-specific CTA images. Combining the cine image technology, which is able to follow opening-closing synchronous movement of four valves at the same time, CTA also has the potential to monitor whole coronary artery tree and our method of locating valve planes may have further implications for understanding the hemodynamic relationship between cardiac chambers and coronary arteries and its impacts on atherosclerotic plaques.
Valve plane imaging
In general, a cross-sectional imaging study of the heart requires reconstruction of two groups of imaging planes including body and heart planes. Standard body planes help us to perceive the location of heart, which includes three orthogonal cuts in axial (transverse), coronal, and sagittal orientations. Standard planes are not suitable for the imaging of the interior anatomy structures of the heart, such as LV, because its location can lead to varying obliquity which may cause significant artifacts [Figure 5]b.[16]
Heart planes [Figure 5]c can help us to perceive the interior anatomy of heart. Cardiac views such as SAX and 4ch perpendicular to the interatrial septum in different phases are required for the complete anatomic study of the fossa ovalis including the PFO shunt.[19],[20] Although heart planes are three orthogonal planes in relation to the heart axes,[21],[22] the body planes transect the heart obliquely, while the heart planes transect the body obliquely. At the crux of the heart, there is an area where TV is attached to the septum closer to the ventricular apex than the MV.[16] Namely, the left atrioventricular grooves and the right atrioventricular grooves are not perpendicular. It is difficult to fix the relative axis in 3D space.
The main reason that we define valve planes based on CS is that the interior anatomy of the heart is so complex that any oblique section may reflect artificial errors in geometric information, which leads to misunderstanding of cardiac morphology and functions. Based on our previous knowledge, valve motion has three main directions: (1) The sphincteric mechanism designates the systolic-to-diastolic variation in annular circumference,[22] and annular enlargement happens preferentially along the shortest diameter direction,[5] (2) the excursion or annular descent is in the apical-to-basal direction, and (3) the rotation indicates the torque movement of the atrioventricular junction. Valve planes shown in [Figure 5]c are defined by our method and are perpendicular or paralleled to these directions of valve motion. The aortic leaflet of MV and the mitral aortic curtain change their shape during the phases of the cardiac cycle, allowing the passage of blood into the LV. On the other hand, the blood ejects from the LV. The direction of valve planes is vertical to the valve direction of motion. Thus, the use of both SAX and LAX cardiac images enables the most accurate localization of the ventricles at the basal and apical levels.[23]
| Conclusions | | |
In this study, we reconstruct a computer 3D model based on CVH5 heart datasets and develop a new approach for patient-specific valve plane localization. Using CVH5 heart datasets, we can localize the original spatial position of CS, which is at the deep center of heart and takes a complex and irregular shape that are difficult to comprehend. As CS is adjacent to most cardiac interior anatomical structure, CVH5 3D anatomical model can identify anatomical details more clearly than clinical images. Therefore, its location method can also be used in designing cardiac interior anatomical landmark planes and showing its relative position in patient-specific CTA images. In addition, we confirm the orientation and geometric relationship of the four valve annulus and apply it to patient-specific valve planes. These methods for generating 3D anatomy models and planes may have further implications for valve plane localization during cardiac cycle with the aim to monitor hemodynamic changes in morphology. In addition, cine loops [24] can be created to demonstrate regional valve and coronary arteries motion, thus having the potential to analyze the dynamic morphology and function of the four valves at 1 time and synchronously analyze the hemodynamic effects on the morphological change of four champers and their surrounding cardiac structures such as coronary arteries.
Financial support and sponsorship
This work was supported by the National Science. Foundation of China (no.61190122)(http://www.nsfc.gov.cn/) and the Social Livelihood Science and Technology Innovation Special Project of CSTC(cstc2015shmszx120002).
Conflicts of interest
There are no conflicts of interest.
| References | | |
| 1. | Zhu XD. The Cardiac Skeleton and the Aortic Root. Surgical Atlas of Cardiac Anatomy. Netherlands, Springer 2015:101-18.  |
| 2. | Misfeld M, Sievers HH. Heart valve macro- and microstructure. Philos Trans R Soc Lond B Biol Sci 2007;362:1421-36. |
| 3. | Mihaila S, Muraru D, Piasentini E, Miglioranza MH, Peluso D, Cucchini U, et al. Quantitative analysis of mitral annular geometry and function in healthy volunteers using transthoracic three-dimensional echocardiography. J Am Soc Echocardiogr 2014;27:846-57. |
| 4. | Gordic S, Nguyen-Kim TD, Manka R, Sündermann S, Frauenfelder T, Maisano F, et al. Sizing the mitral annulus in healthy subjects and patients with mitral regurgitation: 2D versus 3D measurements from cardiac CT. Int J Cardiovasc Imaging 2014;30:389-98. |
| 5. | Maffessanti F, Gripari P, Pontone G, Andreini D, Bertella E, Mushtaq S, et al. Three-dimensional dynamic assessment of tricuspid and mitral annuli using cardiovascular magnetic resonance. Eur Heart J Cardiovasc Imaging 2013;14:986-95. |
| 6. | Veronesi F, Addetia K, Lamberti C, Lang RM, Mor-Avi V. 3D evaluation of tricuspid annulus morphology in patients with pulmonary hypertension.in Computing in Cardiology 2013;2013. p. 241-4. |
| 7. | Owais K, Taylor CE, Jiang L, Khabbaz KR, Montealegre-Gallegos M, Matyal R, et al. Tricuspid annulus: A three-dimensional deconstruction and reconstruction. Ann Thorac Surg 2014;98:1536-42. |
| 8. | Miglioranza MH, Mihaila S, Muraru D, Cucchini U, Iliceto S, Badano LP. Dynamic changes in tricuspid annular diameter measurement in relation to the echocardiographic view and timing during the cardiac cycle. J Am Soc Echocardiogr 2015;28:226-35. |
| 9. | Alkadhi H, Desbiolles L, Husmann L, Plass A, Leschka S, Scheffel H, et al. Aortic regurgitation: Assessment with 64-section CT. Radiology 2007;245:111-21. |
| 10. | Alkadhi H, Wildermuth S, Plass A, Bettex D, Baumert B, Leschka S, et al. Aortic stenosis: Comparative evaluation of 16-detector row CT and echocardiography. Radiology 2006;240:47-55. |
| 11. | Pflederer T, Achenbach S. Aortic valve stenosis: CT contributions to diagnosis and therapy. J Cardiovasc Comput Tomogr 2010;4:355-64. |
| 12. | Cuéllar H, Roque A, Pineda V, Rodríguez J. Cardiac computed tomography for valve disease. Radiologia 2013;55:24-36. |
| 13. | Bouvier E, Logeart D, Sablayrolles JL, Feignoux J, Scheublé C, Touche T, et al. Diagnosis of aortic valvular stenosis by multislice cardiac computed tomography. Eur Heart J 2006;27:3033-8. |
| 14. | Alkadhi H, Wildermuth S, Bettex DA, Plass A, Baumert B, Leschka S, et al. Mitral regurgitation: Quantification with 16-detector row CT – initial experience. Radiology 2006;238:454-63. |
| 15. | Ketelsen D, Fishman EK, Claussen CD, Vogel-Claussen J. Computed tomography evaluation of cardiac valves: A review. Radiol Clin North Am 2010;48:783-97. |
| 16. | Saremi F. Revisiting Cardiac Anatomy: A Computed-Tomography-Based Atlas and Reference. Oxford, UK: John Wiley & Sons; 2011. |
| 17. | Baumert B, Plass A, Bettex D, Alkadhi H, Desbiolles L, Wildermuth S, et al. Dynamic cine mode imaging of the normal aortic valve using 16-channel multidetector row computed tomography. Invest Radiol 2005;40:637-47. |
| 18. | Chen JJ, Jeudy J, Thorn EM, White CS. Computed tomography assessment of valvular morphology, function, and disease. J Cardiovasc Comput Tomogr 2009;3 1 Suppl: S47-56. |
| 19. | Saremi F, Channual S, Raney A, Gurudevan SV, Narula J, Fowler S, et al. Imaging of patent foramen ovale with 64-section multidetector CT. Radiology 2008;249:483-92. |
| 20. | Saremi F, Attai SF, Narula J. 64 multidetector CT in patent foramen ovale. Heart 2007;93:505. |
| 21. | McAlpine WA. Heart and Coronary Arteries: An Anatomical Atlas for Clinical Diagnosis, Radiological Investigation, and Surgical Treatment. New York: Springer-Verlag Berlin; 1975. |
| 22. | Kirklin JW, Blackstone EH, Jonas RA, Kouchoukos NT. Anatomy, dimensions, and terminology. In: Barrat-Boyes BG, ed. Cardiac Surgery, 3 rd ed. Edinburgh, Scotland: Churchill Livingstone; 2004. p. 50. |
| 23. | Perperidis D, Spatio-temporal registration and modelling of the heart using cardiovascular MR imaging. Imperial College London, 2006;84:177-185. |
| 24. | Coulthard T, Wu J. Comparison of ultrasound and compact MRI for the assessment of cardiac function in mice. Circ Res 2014;115:A94. |
[Figure 1], [Figure 2], [Figure 3], [Figure 4], [Figure 5], [Figure 6], [Figure 7]
|
Search Pubmed for