In this issue of JASE , in a well-written and thoughtful article, Park et al . report on a study in which they retrospectively compared transthoracic echocardiographic and computed tomographic angiographic measurements of the aortic root and ascending aorta in 53 patients with bicuspid aortic valves. These authors evaluated three measurement methods (inner edge to inner edge [I-I], leading edge to leading edge [L-L], and outer edge to outer edge [O-O]), as well as three timing points of these measurements (end-diastole, mid-diastole, and end-systole; Figure 1 ). This report calls attention to the ongoing controversy and lack of uniformity among both institutions and guidelines as to the “preferred” method(s) for measuring aortic dimensions, leading to the comparison of “apples to oranges” rather than “apples to apples.”

Diagram illustrating the three methods for measuring the diameter of the aorta. The measurements below are approximate representative correlating measurements.

Accurate measurements of the aorta are necessary for early detection of enlargement and, perhaps more important, have significant prognostic and therapeutic implications in patients with aortic disease. A dilated aorta requires careful, close imaging follow-up because aortic diameter is the strongest predictor of catastrophic events such as dissection and rupture. Aortic size is also used to determine the appropriate timing for prophylactic surgery. Therefore, accurate and standardized measurement techniques of the diameter of the thoracic aorta among the various imaging modalities are desirable and important.

A bit of history behind one of the aforementioned guidelines provides further insight into these issues and also into the process of guideline creation. The following is a quotation from an introductory letter from Drs. Evangelista and Goldstein to the writing committee before beginning work on the document: “Currently, measurement techniques and locations for measurements vary among each of the imaging modalities [referring to echocardiography, computed tomography [CT], and magnetic resonance imaging (MRI)]. We believe that it would be a great contribution if we could arrive at some consensus on how and where to make standard measurements, even if this means a deviation from some traditions [referring to I-I, L-L, and O-O]. We doubt that we will get 100% consensus on this issue. Nevertheless, we feel that it would be of value if we could accomplish this.” It was our intent to improve comparison among the major imaging modalities (echocardiography, CT, and MRI) by unifying the measurement technique. Surprisingly, there was total initial agreement on using the I-I method for all three modalities, thus accomplishing the goal of comparing apples to apples. The American Society of Echocardiography/European Association of Cardiovascular Imaging document in progress continued with this recommendation for more than a year. However, a few months before publication, a group of the writing committee raised major concern based partly on new unpublished normative data but largely on concern for patient safety. Measurements of the aorta using the I-I convention are consistently approximately 2 mm smaller than measurements using the L-L technique. Most of the prognostic data and therapeutic recommendations were established using criteria based on the L-L method. Because the I-I method yields smaller dimensions than the L-L method, there was concern that patients would be followed beyond the standard recommendations for surgery, resulting in unnecessary risk for dissection and/or rupture—a potential “lemon” or bitter outcome. “Last-minute” telephone meetings led to a reversal of the initial plan to unify measurements by all modalities to the I-I method. After much consideration and deliberation, the writing group decided to continue to recommend that echocardiographic measurements be made in the traditional fashion from leading edge to leading edge. This issue also provides insight into the process of guideline creation. Decisions in writing groups are made on the basis of majority opinion (consensus), which may or may not be unanimous. In other words, practice recommendations are often based on consensus of experts, despite frequently not reaching 100% agreement.

Park et al . have raised and addressed a number of points and important issues that deserve emphasis, and we add a few comments.

• 1.
  

  The authors distinguish between the aortic root (usually defined as the portion of the aorta between the aortic annulus and the sinotubular junction) and the tubular portion of the aorta (which they refer to as the ascending aorta). Although not the technique of choice for the complete assessment of the thoracic aorta, transthoracic echocardiography (TTE) is both practical and useful for imaging the aortic root for both diagnosis and follow-up. Imaging of the aorta is a routine part of the standard transthoracic echocardiographic examination, and in the majority of patients the aortic root is well imaged in the left parasternal long-axis view. However, imaging the ascending aorta beyond the sinotubular junction presents challenges in many patients. When the predominant area of dilatation is above the sinotubular junction, as it is in most patients with bicuspid aortic valves, TTE is not an ideal screening tool.

  
  
• 2.
  

  The authors imply that computed tomographic angiography (and MRI) remain the “gold standard” for quantitative imaging of the aorta. We agree with this premise. Electrocardiographically gated multidetector CT provides high-resolution images that allow precise three-dimensional measurements and guarantee double-oblique, perpendicular alignment to the long axis of the aorta. This minimizes the likelihood that oblique measurements will be made, which falsely increase dimensions. Similarly, it avoids long-axis views that are foreshortened, therefore rendering falsely decreased dimensions. Nevertheless, in clinical practice, TTE can be used for serial imaging follow-up of the dilated aortic root or ascending aorta when agreement between the dimensions measured by TTE and CT/MRI has been documented. The advantages of a three-dimensional (3D), double-oblique methodology (avoiding oblique and foreshortened measurements) are evident, but these advantages may not be modality-related but rather the result of two-dimensional versus 3D techniques.

  
  
• 3.
  

  We want to emphasize the recommendation that the maximum diameter be measured perpendicular to the centerline (long axis) of the aorta. Pitfalls include tortuous, kinked, and curved portions of the aorta. It should also be mentioned that the aorta is not always round, especially when aneurysmal or tortuous.

  
  
• 4.
  

  The authors highlight the importance of asymmetry of the sinuses of Valsalva in patients with bicuspid aortic valves. They determined and recommended that the largest transthoracic echocardiographic diameters at the sinus of Valsalva level be obtained using the mid-diastolic short-axis L-L method (the resulting dimensions are about 2 mm larger than by the conventional I-I method). Although there is some merit to this method, further prospective evaluation will be required to establish its value. Several drawbacks for TTE include reduced lateral resolution and difficulty avoiding oblique images.

  
  
• 5.
  

  Measurements should be made at standard locations: the aortic annulus, sinuses of Valsalva, sinotubular junction, and the ascending aorta at a specified level (e.g., 2 or 3 cm above the sinotubular junction or at the level of the right pulmonary artery).

  
  
• 6.
  

  We would like to emphasize the importance of electrocardiographically gated CT for measuring the aortic root and ascending aorta. Electrocardiographic gating is used to freeze the motion of the aorta, and this will significantly minimize artifacts caused by cardiac motion, particularly at the level of the sinuses of Valsalva and ascending aorta.

  
  
• 7.
  

  Aortic root and ascending aortic diameters vary depending on the trigger time (such as end-systole vs end-diastole). Although arguments for either can be made, we agree with Park et al . that diastole may be preferable for the aortic root and ascending aorta. Measuring in diastole improves reproducibility, as blood pressure is more stable in end-diastole. In addition, the least cardiac motion occurs during diastole, which can be important for CT. On the other hand, early pediatric investigations measured aortic dimensions at maximum expansion (i.e., at the time of peak aortic wall stress) because of the thought that this measurement would be more predictive of the risk for dissection.

  
  
• 8.
  

  The authors state that “regardless of the measurement method used, it is critical to physically report the measurement methodology used…to allow further interval comparison with the same measurement method.” We fully agree and emphasize that it is critically important to report the measurement method used so that meaningful interval comparisons can be made. We would add that whenever possible, interval measurements be made with direct comparison to prior images to truly compare apples to apples.

In summary, in their excellent study, Park et al . raise and discuss a number of issues relating to measurement techniques for the aortic root and ascending aorta. The recently published report from the GenTAC registry suggests that all of the professional imaging societies should work together to develop and adopt a unified method for measuring the thoracic aorta for consistency and improved communication. Some would state that because the transthoracic echocardiographic L-L technique correlated better with computed tomographic I-I, these two different approaches should continue to be used. In reality, TTE underestimates the dimension, likely because of the inability to measure the largest dimension in the z plane (third dimension, foreshortening). Rather than simply using different methods (I-I vs L-L, i.e., apples vs oranges) it is reasonable to aim to unify our techniques so that measuring the structure results in similar dimensions regardless of the imaging modality used. We believe that this may be possible with newer imaging modalities that are 3D based, such as x -plane, biplane, or real-time 3D imaging. Further investigation is needed to validate this concept. We anticipate that, provided that this concept proves to be true, the I-I convention will be adopted for all 3D imaging modalities (echocardiography, CT, and MRI) in the near future.