Author + information
- Received April 30, 2012
- Revision received September 28, 2012
- Accepted October 4, 2012
- Published online May 1, 2013.
- Sasi Sithamparanathan, MBBS⁎,
- Simon P.G. Padley, MBBS†,
- Michael B. Rubens, MBBS†,
- Michael A. Gatzoulis, PhD‡,
- Siew Yen Ho, PhD‡ and
- Edward D. Nicol, MD, MBA†,‡,⁎ ()
- ↵⁎Reprint requests and correspondence:
Dr. Edward Nicol, Royal Brompton Hospital and Harefield NHS Trust, Sydney Street, London, SW3 6NP, United Kingdom
In patients with transposition of the great arteries, the identification of coronary anatomy is fundamental to optimal surgical outcome. A number of classifications describing the coronary vessels' origin and course in transposition of the great arteries have been published. However, all are limited to operative or pathological case series. They are often alphanumeric classifications that do not lend themselves to clinical practice; they do not consider certain important anatomical variations that may increase surgical morbidity and mortality, nor do they fully delineate coronary anatomy or define the relationship to adjacent structures seen with cardiovascular computed tomography. Using cardiovascular computed tomography for illustrative purposes, we propose and validate a universal sequential descriptive classification and an associated alphanumeric classification that may be used for all coronary anomalies with or without associated congenital heart disease.
Coronary computed tomography (CT) angiography allows for a complete visualization of the origin and course of the coronary arteries, in context with the surrounding cardiovascular anatomy. Although of great potential for the analysis of anomalous coronary anatomy, this leads to the requirement of a comprehensive descriptive framework for the coronary arteries and great vessels to allow cardiologists, radiologists, and surgeons alike to accurately describe and classify coronary and cardiovascular anatomy. Existing alphanumeric classifications of coronary anatomy, though useful in the research and taxonomy settings are of limited value within clinical settings where they do not cover all coronary variants and where their use requires a complete understanding of the alphanumeric code by all clinicians. Whereas a sequential descriptive classification exists for atrioventricular and ventriculo-arterial connections, there is currently a lack of a universal, systematic descriptive classification for coronary and great vessel anatomy.
Clinically, coronary anomalies matter. In patients with transposition of the great arteries (TGA), the identification of coronary anatomy is fundamental to optimal surgical outcome (1). At the more benign end of the spectrum this may include 3 separate coronary ostia (Fig. 1), whereas a single coronary ostium (Fig. 2) confers a 3-fold mortality increase, a “malignant” intramural/interarterial coronary artery (Fig. 3) is associated with 6-fold increase in mortality, whereas overall mortality in those patients with abnormal anatomy is nearly double (1). (We use “malignant course” to describe both interarterial [major epicardial vessel running between the right ventricular outflow tract and the aortic root] and intramural [coronary arteries additionally reside within the aortic wall for a variable duration of their interarterial course]. The use of multidetector CT to differentiate between these 2 types of malignant courses is both unproven and challenging; however, both are associated with poorer prognosis.) As a result of the association of coronary anomalies with adverse outcomes in TGA, the classification of coronary anatomy in TGA has been attempted by several investigators (Table 1). However, all previous classifications are incomplete as they are based on relatively small surgical, pathological, or echocardiographic series and, therefore, are limited to these case series and cannot be applied to other anomalies. No previous classification comprehensively covers important anatomical variations that have been demonstrated to, or have the potential to, increase surgical morbidity and mortality. Finally, they cannot be used to describe isolated coronary anomalies that may predispose to sudden cardiac death.
Partial classifications do exist for certain coronary anomalies (such as Lipton's classification of isolated single coronary artery patterns and Angelini's proposed descriptive approach to coronary anatomy at angiography ), but again these are limited, for example, to invasive angiography and only applicable to patients without congenital heart disease (CHD). Additionally, they are not transferable for use with other imaging modalities.
Clinically, Leiden's classification is often used. However, this alphanumeric classification is complicated and may not fully describe coronary anatomy, and the codex needs to be understood by all clinicians to be of practical use. The description of the aortic sinuses within the Leiden classification is also confusing (Fig. 4), and the relationship between the aortic root and pulmonary artery is not routinely described. The Leiden classification cannot be used to describe the anatomy in congenitally corrected transposition of the great arteries (ccTGA) where the “normal” variation of the coronary anatomy also influences both natural history and prognosis nor in isolated coronary anomalies that are increasingly recognized in patients.
We propose both a universal, descriptive, sequential approach for use clinically and an alphanumeric equivalent for the purposes of taxonomy and research where a shorthand description allows for a simple, universal, and standardized codex. Both classifications allow for the complete delineation of great vessel and coronary anatomy, comprehensively encompassing variations with significant morbidity and mortality.
Sequential descriptive and alphanumeric classification of coronary anatomy
For the clinical setting, we developed a descriptive approach to delineate the great vessel and coronary anatomy. However, we also believed there was a value in an alphanumeric system for taxonomy and research, and we recognized the potential to break down the descriptive classification into an alphanumeric classification. Therefore, we also present our descriptive classification as an extended Leiden alphanumeric classification that can be used in all CHD patients and isolated coronary anomalies based on multidetector CT.
We describe the relationship of the aorta (Ao) to the pulmonary artery (PA) and then the coronary anatomy. In both classifications, we allow for additional ostia to be described, describe which arteries supply which ventricles in ccTGA, and, if present, denote an interarterial malignant course. We describe aortic sinuses as their morphological position as related to the patient: anterior or posterior; right or left (Fig. 5). In the alphanumeric classification, we extend and modify the Leiden method to include the methodology described. These simple changes in the classification take into account the important role of the coronary anatomy in the surgical outcome of CHD patients.
The proposed new descriptive terminology is applied sequentially as follows:
1. Use the axial views to describe the anatomical relationship of the aorta to the pulmonary artery (if present): Ao directly anterior to PA; Ao directly posterior to PA; anterior right; anterior left; posterior right; posterior left; side-by-side left (Ao is left of PA); and side-by-side right (Ao is right of PA).
2. Describe the origin of the coronary arteries: describing the aortic sinuses as left or right, anterior or posterior with respect to the patient and stating the number of ostia present off each sinus. If >1 ostium is present on a given sinus, the more anterior is described first, with subsequent, more posterior ostia following in order. Bicuspid valves would be described as either anterior or posterior or left and right depending on orientation (Fig. 5).
3. Describe the anterior descending artery according to cardiac morphology; that is, do not automatically begin with the left anterior descending artery in TGA.
4. Describe clearly the high risk or malignant interarterial/intramural coronary course between Ao and PA.
5. Describe the coronary arterial supply to the left or right ventricle for patients with ccTGA.
The proposed new alphanumeric terminology is applied sequentially as follows:
1. Use the axial view to describe the morphological relationship of Ao to PA (if present): A = Ao directly anterior to PA; P = Ao directly posterior to PA; AR = anterior right; AL = anterior left; PR = posterior right; PL = posterior left; SSL = side-by-side left (Ao is left of PA); and SSR = side-by-side right (Ao is right of PA).
2. Describe the origin of the coronary arteries: describing the aortic sinuses as left or right, anterior or posterior with respect to the patient and stating the number of ostia present off each sinus. If >1 ostium is present on a given sinus, the more anterior is labeled i, with subsequent, more posterior sinuses labeled ii, iii, etc. The use of lowercase avoids confusion with coronary labeling. a = anterior (if only 1 anterior sinus); ar = anterior right; al = anterior left; p = posterior (if only 1 posterior sinus); pr = posterior right; pl = posterior left. Bicuspid valves would be a and p or l and r depending on orientation (Fig. 5). A semicolon is used to separate the different aortic sinuses and if >1 ostium is present in a given sinus.
3. Label the anterior descending artery according to cardiac morphology; that is, do not automatically begin with the left anterior descending artery in TGA.
4. Describe clearly the high risk or malignant interarterial coronary course between the Ao and PA and annotate the alphanumeric classification with aM notation.
5. Denote the coronary arterial supply to the left or right ventricle for patients with ccTGA (i.e., supplying morphological left ventricle or right ventricle).
To demonstrate these new sequential classifications, we illustrate several cases with CT angiography and compare the proposed new classification with the previous Leiden terminology and other classifications where appropriate. We show standard caudal-cranial multiplanar reformatted radiological views alongside the schematic diagrams. In clinical practice, however, the craniocaudal volume rendered images are often the most useful for fully delineating the coronary anatomy, and these are shown are separate color images (Figs. 6 to 9).⇓⇓⇓⇓
Results and interobserver variability
To validate the proposed classification and to test interobserver variability, 2 level-3 Society of Cardiovascular Computed Tomography operators (consultant cardiologist and consultant radiologist) tested the validity of the classification on 50 standard CT coronary angiograms, including those with aberrant anatomy, and 50 cases with CHD (including TGA, ccTGA, and other simple and complex CHD). Agreement was achieved in all but 1 case, a case of single coronary artery where the proximal, mid, and distal left anterior descending consisted of 3 individual vessels arising from different locations.
Applicability to other imaging modalities
We propose a clinically relevant, universal descriptive, sequential, morphological approach consisting of describing the relationship of the Ao to the PA, the number of coronary ostia in each aortic sinus, and the origin and course of each major epicardial coronary artery. The alphanumeric classification proposed is difficult to use in clinical settings because both users and readers need to possess full knowledge and understanding of the coding system. In the fields of taxonomy and research, however, this extended alphanumerical methodology provides the flexibility to describe coronary anatomy more comprehensively, accurately, and reproducibly in patients with and without CHD.
The new classification takes into consideration the increasing knowledge of coronary anatomy following the advent of multidetector CT coronary anatomy and can be used in the reporting of cardiac CT. Importantly, both classifications describe the 2 variants of coronary arteries associated with the highest mortality in TGA patients: malignant interarterial courses; and arteries arising from a single ostium. It also provides a fundamental method of communicating complex coronary anatomy within clinical and academic settings.
The new classification has been designed using CT angiography; however, it is transferrable to other imaging modalities. It is best suited to cross-sectional techniques such as CT and cardiac magnetic resonance; however, it is not always possible to visualize the entire coronary tree using cardiac magnetic resonance. It is possible to use the classification with ICA so long as all branches are visualized. The classification is also theoretically applicable to other imaging modalities, even though techniques such as transesophageal echocardiography and transthoracic echocardiography do not usually visualize the coronary tree sufficiently well for diagnostic purposes.
The new classification will potentially allow a more sophisticated analysis of coronary anomalies and support further research into the effect of specific abnormalities on both pre-surgical and general prognosis in patients with TGA and other coronary anomalies, both in isolation and with associated CHD.
The authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- congenitally corrected transposition of the great arteries
- congenital heart disease
- computed tomography
- pulmonary artery
- transposition of the great arteries
- Received April 30, 2012.
- Revision received September 28, 2012.
- Accepted October 4, 2012.
- American College of Cardiology Foundation