- Email: [email protected]
The Frequency of Acquisition and Quality of Images Obtained by Transesophageal Echocardiography: A Prospective Study of 100 Consecutive Cases
The Frequency of Acquisition and Quality of Images Obtained by Transesophageal Echocardiography: A Prospective Study of 100 Consecutive Cases Gerald I. Cohen, MD, Malcolm B. Davison, MBBS,a Allan L. Klein, MD, Emesto E. Salcedo, MD, and William J. Stewart, MD, Cleveland, Ohio
Although transesophageal echocardiography is capable of generating detailed images of the heart and aorta, the quality of these images and the frequency of their acquisition has not been previously defined. We performed monoplane transesophageal echocardiograms in lOO consecutive patients to determine the quality and frequency of acquisition of 10 standard, transverse views of the heart and aorta. Each image was graded for quality by ascribing a value from 0 (poor quality or not obtained) to 3 (excellent quality). The effect of age on image acquisition and quality was determined for patients less than or equal to 60 years (n = 54) versus greater than 60 years old (n = 46). For the 10 image planes, the average image quality score and the frequency of image acquisition(%) was (l) 2.23 and 100% for the four-chamber view, (2) 1.96 and 98% for the right ventricular inflow view and coronary sinus, (3) 2.17 and 99% for the five-chamber view, (4) 1.80 and 84% for the aortic valve short axis, (5) 1.54 and 82% for the mitral valve short axis, (6) 1.84 and 91% for the mid-left ventricular short axis, (7) 1.02 and 59% for the left ventricular apical short axis, (8) 1.79 and 93% for the ascending aorta, (9) 1.23 and 77% for the aortic arch, and (10) 1.88 and 92% for the descending thoracic aorta view. Ascending aorta and aortic valve short-axis views were obtained with a significantly higher frequency in patients less than or equal to 60 years old. In addition, image quality was significantly better for views of the left ventricular apical short axis, aortic arch, mitral valve short axis, mid-left ventricular short axis, and the right ventricular inflow view in the younger population. Image quality for most basal transesophageal images is high, higher than that of transgastric images, which are less frequently obtained. Although the frequency of image acquisition is usually not affected by age, image quality is often better in patients who are 60 years of age or younger. (JAM Soc EcHOCARDIOGR 1993;6: 577-82.)
Monoplane transverse transesophageal echocardiography generates a limited number of images because the transducer is confined to the esophagus and
From the Department of Cardiology of The Cleveland Clinic Foundation. 'Current address for M.D. is St. Andrews' War Memorial Hospital, Brisbane, Australia. Data presented at the Second Scientific Session of the American Society of Echocardiography, June 1991. Reprint requests: Gerald I. Cohen, MD, The Cleveland Clinic Foundation, Department of Cardiology F-15, 9500 Euclid Ave., Cleveland, OH 44106. Copyright© 1993 by the American Society ofEchocardiography. 0894-7317/93$1.00 + .10 27/l/48557
stomach and because the ultrasound plane cannot be rotated. 1-3 Although newer biplane and multiplane probes have increased the number of views,4- 7 the advantages of these new developments are difficult to evaluate without first defining the limitations of single-plane imaging. Neither the frequency nor the quality of image acquisition have been previously described for monoplane transesophageal echocardiography. Furthermore, the impact of patients' increased age on transesophageal echocardiography has not been previously examined, although this adversely affects the quality of transthoracic images. 8 Therefore, the purpose of our study was ( l) to determine how often standard transverse transesophageal views can be obtained, (2) to assess the quality 577
Journal of the American Sociery of Echocardiography November-December 1993
578 Cohen et al.
Table 1 The demographics and M-mode measurements of the two groups studied (patients s60 and >60 years old)
n
Age Sex OPD/OR LA RT LVd LVs FS(%)
:s60
>60
54 44±11 27F:27M 26/28 47 ± lO 32 ± 6 56± 11 36 ± 10 36 ± 8
46 68 ± 5 14F:32M 27/19 49 ± 10 33 ± 7 56± l l 39 ± 11 31 ± l l
M-mode measurements of the two groups (reported as mean ± 2 SD) were compared by use of unpaired student t tests; chi-square analysis was used for sex and examination location ratios. Except for age, sex (p = 0.05), and fractional shortening (p = 0.01), the two groups are not significantly different. OPD I OR = the number of studies performed on an ambulatory versus an intraoperative basis. M-mode measurements are of the left atrium (lA), aortic root (RT), and of the left ventricle in diastole (LVd) and systole (L Vs). FS(%) = fractional shortening.
of these images, and (3) to determine the effect of age on image acquisition.
METHODS Patient Population
Monoplane transesophageal echocardiograms were performed prospectively on 100 consecutive patients to demonstrate standard views of the heart and aorta. Fifty-four of the patients were less than or equal to 60 years of age, and 46 patients were more than 60 years old (Table 1). Patients were referred for transesophageal imaging to assess the mitral valve (n = 43), the aortic valve (n = 14), or both valves (n = 9), to exclude endocarditis (n = 13) or an intracardiac source of embolism (n = 7), and to evaluate suspected aortic or pulmonary disease (n = 4), atrial septal defects (n = 3), masses (n = 3), constrictive pericarditis (n = 2), and hypertrophic cardiomyopathy (n = 2). Equipment
Machines used for transesophageal imaging included the Hewlett Packard 1000 (Andover, Mass.) and Acuson 128 (Moutainview, Calif.). Informed consent was obtained before the examination, which was performed in a standard manner that has been previously described. 1-3 During the entire examination, machine settings (depth, gain, and compress) were optimized. A 5 MHz single-plane probe was used. Studies were performed on an ambulatory basis
(n = 53) or before surgery (n = 47) in the surgical suite with the patient under general anesthesia. Before the series was started, we identified, by consensus, 10 standard transverse views of the heart and thoracic aorta that merited evaluation. In addition to routine image acquisition performed to answer clinical questions at hand, the physician attempted to obtain and optimize these views: basal views of (1) the left ventricular inflow or four-chamber view, (2) the right ventricular inflow, which includes an image of the coronary sinus, (3) the left ventricular outflow or five-chamber view, and (4) the short axis of the aortic valve (specific structures were also sought at or near this level, including the atrial septum, left atrial appendage, pulmonary arteries, and coronary arteries); standard transgastric views of the short axis of the (5) mitral valve and the (6) mid -left ventricle and (7) apical left ventricle, and views of the (8) ascending aorta, (9) arch, and (10) descending thoracic aorta. Probe positioning and manipulation to obtain these images was performed in a standard manner that has been previously described. 3•5 On average the intraoperative study took 20 minutes to perform whereas the outpatient study took 30 minutes to perform. The study time varied with the complexity of the cases. Although all standard views were attempted in a timely manner, there was never a clinical event that led to an abbreviated or incomplete study in any patient.
Analysis The videotaped studies were subsequently reviewed by two observers blinded to the patient's clinical data to determine the frequency of image acquisition and the quality of images. Image quality was graded by consensus and scored from 0 to 3. The score assigned was 0 when a meaningful image was not obtained, 1 when the image was comprehensible but border definition was only fair, 2 when most but not all borders had good definition, 3 when all borders had excellent definition. For each view, the grades for image quality were averaged as an overall quality score for the patient population. The obliquity of short-axis images of the aortic and mitral valve and the left ventricle were evaluated as absent for a perfectly round short-axis image, minimal for an imperfect short-axis or oval image, and marked for a very oblique short-axis or very oval image. The effect of age (::S versus >60 years) on the frequency of image acquisition and quality was determined by use of the Fisher's exact test. The
Journal of the American Society of Echocardiography
Cohen et al.
Volume 6 Number 6
579
Table 2 The quality of standard images of the heart and aorta indicated by a grade of 0 (not obtained) to 3 (excellent quality) and the overall quality score. Numbers represent the number of patients or percentages in each category of image quality (N = lOO) Quality Scale Image
Basal views Four-chamber RV inflow Five-chamber Aortic valve short-axis Transgastric MV short-axis LV (mid) short-axis LV (apex) short-axis Aorta Ascending Arch Descending
0
1
2
3
Quality score (all patients)
0 2 16
4 17 6 6
69 64 68 60
27 17 25 18
2.23 1.96 2.17 1.80
18 9 41
21 20 26
50 51 23
20 10
11
1.54 1.84 1.02
7 23 8
23 43 9
54 32 70
16 2 13
1.79 1.23 1.88
l
RV, Right ventricle; MV, mitral valve; LV, left ventricle.
Table 3 Image quality (0 = not obtained to 3 = excellent) and overall score of the left atrial appendage, atrial septum, and left, main, and right pulmonary arteries. Numbers represent the number of patients or percentages in each category of image quality (N = lOO) Quality scale Image
0
1
2
3
Quality score (all patients)
Specific structures Atrial septurn LA appendage LPA MPA RPA
15 55 9 9
l
10 22 20 32 27
61 45 22 47 50
28 18 3 12 14
2.16 1.66 0.73 1.62 1.69
LA, Left atrial; LPA, left pulmonary artery; .MPA, main pulmonary artery; RPA, right pulmonary artery.
obliquity of views of the aortic valve, mitral valve, and mid-left ventricle were compared by chi-square analysis. RESULTS Image Quality Versus Standard View
Table 2 summarizes the frequency of image acquisition and image quality of the l 0 standard image planes. In the majority of patients, most standard basal views and views of the thoracic aorta were obtained and graded l to 3 for image quality; failure to obtain an image (grade 0) occurred infrequendy. Good or excellent transgastric views were less often obtained than the basal transesophageal views (Figure l). Short-axis views of the aortic and mitral valves
were obtained in most patients, though usually oblique to some extent. A perfect short axis of the left ventricle could usually be obtained (Figure 2). In most patients, it was possible to image most of the thoracic aorta. The best images were obtained of the descending thoracic aorta. In contrast, the arch was more difficult to image and was seen adequately in only 77% of patients. Visualization of Specific Structures
Good images of the atrial septum, left atrial appendage, and main and right pulmonary artery were usually acquired (Table 3); however, the left pulmonary artery was usually not seen well. Coronary artery visualization was good but not universal. The left main trunk was seen in 74% of patients whereas its branches, the left anterior descending artery and cir-
580
Journal of the American Society of Echocardiography November-December 1993
Cohen et al.
Image Quality of Standard Views
(n=lOO)
"100 80
60
40 20
.,
¥
,
Basal Views
Figure 1 Bar graph representation of the percentage of patients with good or excellent quality images of standard basal, transgastric, and aortic views. 4C, Four-chamber; RV, right ventricle inflow; SC, five-chamber; AV-SX, aortic valve short-axis; MV, mitral valve; LV, left ventricle; Asc, ascending; Desc, descending.
cumflex artery, were visible in 43% and 67%, respectively. The right coronary artery could be seen in only 51% of patients. Effect of Age on Image Acquisition and Quality
Aging had more of an adverse effect on image quality than on frequency of image acquisition (Table 4). Only the short-axis view of the aortic valve and the view of the ascending aorta were less frequently obtained in the older population. For the other imaging planes, the frequency of acquisition was not significantly different for the two age groups. However, image quality was significantly worse in older patients for the views of the right ventricular inflow, transgastric short-axis view of the mitral valve and left ventricle, and the view of the aortic arch. DISCUSSION
This study documents the frequency with which transverse transesophageal echocardiography generates high-quality images. Compared with transthoracic imaging, it is well known that transesophageal echocardiography has the advantage of being in close
proximity to the heart with less intervening tissue to impede ultrasound penetration. Thus good to excellent definition of edges and endocardium was obtainable in most images. Inadequate visualization of other structures was likely due to their far-field position to the transducer, especially for transgastric views, or due to tracheal interposition, as with the aortic arch. 1-3 In the case of the left pulmonary artery, this structure quickly moves out of the plane of imaging and was often not well seen. Transesophageal echocardiography imaged the posterior heart best. Basal views had a better quality score than transgastric images in which the heart was more in the far field of the ultrasonic beam. Despite their small diameter, coronary arteries could usually be seen, especially the left main and circumflex arteries. The right coronary artery was also often not well seen, at least in part because of its position in the ultrasonic far field. Our success rate was not as high as Zwicky et al., 9 who imaged the proximal right and left coronary arteries in 50 consecutive patients. This discrepancy may be due to differences of technique (including length of time devoted to imaging the coronary arteries), equipment, and definition of adequate coronary visualization. The inability to rotate the transducer or move it beyond the confines of the esophagus and stomach is the most likely cause of the frequent inability to obtain a perfect short-axis view of the aortic and mitral valves and of the left ventricle. This is an important limitation because it may prevent one from localizing pathologic lesions of the leaflet and from diagnosing wall motion abnormalities. Without an adequate short axis, it is more difficult to determine the medial versus lateral relations of mitral pathologic lesions or the site of regurgitation along the coaptation line of the leafiets. 1°For the aortic valve, an adequate short axis is important to determine the number of cusps present. 6 The Effect of Age
This is the first study to show that patient characteristics, such as age, have an impact on transesophageal image quality and acquisition. We speculate that this may be due to a greater incidence of esophageal pathology (such as esophageal dilatation and hiatal hernia) with reduced transducer contact with the mucosa. 11 Dilatation of cardiac chambers and tortuosity of the aorta may have distorted the position of these structures relative to the esophagus and stomach, though M-mode measurements did not show a significant difference between the two age groups. However, it is well known that aging has a
Journal of the American Society ofEchocardiography Volume 6 Number 6
581
Cohen et al.
Figure 2 Bar graph comparison of the obliquity of the aortic and mitral valves (v) and the mid left ventricle (LV) during transverse transesophageal imaging. Views of the left ventricle were significantly less oblique than those of the aortic valve (p < 0.0001) and mitral valve (p < 0.006). Numbers represent the number of patients or percentage in each category of obliquity.
Table 4 The effect of age on image acquisition (%) and quality score for patients < versus > 60 years of age Frequency of acquisition (%) Image plane
:s60years
>60years
Four-chamber RV inflow Five-chamber AoV Short Axis MV Short Axis LV (mid) Short Axis LV (apex) Short Axis Ascending Aorta (Asc) Aortic Arch Descending Aorta (Desc)
lOO lOO lOO
lOO
92 87 95 65 98 80 94
96 98 75 77 87 52 87 74 89
Quality score
p
* *
:s60years
>60years
2.24 2.07 2.20 1.85 1.72 1.92 1.20 1.91 1.20 1.98
2.22 1.83 2.13 1.67 1.33 1.65 0.80 1.65 1.04 1.85
p
* * * * *
*Significant (p :s 0.05) differences between the two age groups for image acquisition frequency (left) and quality (right).
much more profound effect on image quality during transthoracic imaging because of obesity, thickening of the chest wall, and lung disease. 8 Limitations
This study does not establish whether variation in image quality of obliquity during single-plane transesophageal echocardiography has an adverse affect on patient outcome. This question is a major issue because of the greater physician time commitment and expense of transesophageal echocardiography. Our findings represent a reference point for an evolving technology because newer probes may generate better images. Muitiple frequency probes may improve near-field and far-field resolution and multiplane probes may help offset problems with shadowing and image obliquity.H Other factors that af-
feet image quality include machine type, technology, and settings, transducer contact with the esophagus, and operator skill. Though these factors were not specifically examined by our study, utilization of multiple machine types under different clinical settings and the participation of many skilled echocardiographers should make our averaged findings more widely applicable. Single-plane transesophageal echocardiography was able to provide high-quality and detailed views of the heart and aorta in the majority of our patients. REFERENCES l. Gussenhoven EJ, Taams MA, Roelandt JR, et al. Transesophageal two-dimensional echocardiography: its role in solving clinical problems. JAm Coil Cardiol 1986;8:975-9.
582 Cohen et al.
2. Mitcheil MM, Sutherland GR, Gussenhoven EJ, Taams MA, Roelandt JR. Transesophageal echocardiography. J AM Soc ECHOCARDIOGR 1988;1:362-77. 3. Seward JB, Khandheria BK, Oh JK, et al. Transesophageal echocardiography: technique, anatomic correlations, implementation, and clinical applications. Mayo Clin Proc 1988;63:649-80. 4. Omoto R, Kyo S, Matsumura M, et al. Recent technological progress in transesophageal color Doppler flow imaging with special reference to newly developed biplane and pediatric probes. Heidelberg: Springer-Verlag, 1989:21. 5. Seward JB, Khandheria BK, Edwards WD, Oh JK, Freeman WK, Tajik. AJ. Biplanar transesophageal echocardiography: anatomic correlations, image orientation, and clinical applications. Mayo Clin Proc 1990;65:1193-213. 6. Cohen GI, Chan KL. Biplane transesophageal echocardiography: clinical applications of the long-axis plane. J AM Soc ECHOCARDIOGR 1991;4:155-63.
Journal of the American Society of Echocardiography November-December 1993
7. FlachskampfFA, Hoffmann R, Hanrath P. Experience with a transesophageal echo transducer allowing full rotation of the viewing plane: the ornniplane probe [Abstract]. J Am Coil Cardiol1991;17:34A. 8. Bansal RC, Tajik. AJ, Seward JB, Offord KP. Feasibility of detailed two-dimensional echocardiographic examination in adults. Prospective study of 200 patients. Mayo Clin Proc 1980;55:291-308. 9. Zwicky P, Daniel WG, Mugge A, Lichtlen PR. Imaging of coronary arteries by color-coded transesophageal Doppler echocardiography. Am J Cardiol 1988;62:639-40. 10. Stewart WJ, Currie PJ, Salcedo EE, et al. Evaluation of mitral leaflet motion by echocardiography and jet direction by Doppler color flow mapping to determine the mechanism of mitral regurgitation. JAm Coil Cardiol1992;20:l353-6l. ll. Altman DF. The effect of age on gastrointestinal function. Philadelphia: WB Saunders, 1989:162.
Although transesophageal echocardiography is capable of generating detailed images of the heart and aorta, the quality of these images and the frequency of their acquisition has not been previously defined. We performed monoplane transesophageal echocardiograms in lOO consecutive patients to determine the quality and frequency of acquisition of 10 standard, transverse views of the heart and aorta. Each image was graded for quality by ascribing a value from 0 (poor quality or not obtained) to 3 (excellent quality). The effect of age on image acquisition and quality was determined for patients less than or equal to 60 years (n = 54) versus greater than 60 years old (n = 46). For the 10 image planes, the average image quality score and the frequency of image acquisition(%) was (l) 2.23 and 100% for the four-chamber view, (2) 1.96 and 98% for the right ventricular inflow view and coronary sinus, (3) 2.17 and 99% for the five-chamber view, (4) 1.80 and 84% for the aortic valve short axis, (5) 1.54 and 82% for the mitral valve short axis, (6) 1.84 and 91% for the mid-left ventricular short axis, (7) 1.02 and 59% for the left ventricular apical short axis, (8) 1.79 and 93% for the ascending aorta, (9) 1.23 and 77% for the aortic arch, and (10) 1.88 and 92% for the descending thoracic aorta view. Ascending aorta and aortic valve short-axis views were obtained with a significantly higher frequency in patients less than or equal to 60 years old. In addition, image quality was significantly better for views of the left ventricular apical short axis, aortic arch, mitral valve short axis, mid-left ventricular short axis, and the right ventricular inflow view in the younger population. Image quality for most basal transesophageal images is high, higher than that of transgastric images, which are less frequently obtained. Although the frequency of image acquisition is usually not affected by age, image quality is often better in patients who are 60 years of age or younger. (JAM Soc EcHOCARDIOGR 1993;6: 577-82.)
Monoplane transverse transesophageal echocardiography generates a limited number of images because the transducer is confined to the esophagus and
From the Department of Cardiology of The Cleveland Clinic Foundation. 'Current address for M.D. is St. Andrews' War Memorial Hospital, Brisbane, Australia. Data presented at the Second Scientific Session of the American Society of Echocardiography, June 1991. Reprint requests: Gerald I. Cohen, MD, The Cleveland Clinic Foundation, Department of Cardiology F-15, 9500 Euclid Ave., Cleveland, OH 44106. Copyright© 1993 by the American Society ofEchocardiography. 0894-7317/93$1.00 + .10 27/l/48557
stomach and because the ultrasound plane cannot be rotated. 1-3 Although newer biplane and multiplane probes have increased the number of views,4- 7 the advantages of these new developments are difficult to evaluate without first defining the limitations of single-plane imaging. Neither the frequency nor the quality of image acquisition have been previously described for monoplane transesophageal echocardiography. Furthermore, the impact of patients' increased age on transesophageal echocardiography has not been previously examined, although this adversely affects the quality of transthoracic images. 8 Therefore, the purpose of our study was ( l) to determine how often standard transverse transesophageal views can be obtained, (2) to assess the quality 577
Journal of the American Sociery of Echocardiography November-December 1993
578 Cohen et al.
Table 1 The demographics and M-mode measurements of the two groups studied (patients s60 and >60 years old)
n
Age Sex OPD/OR LA RT LVd LVs FS(%)
:s60
>60
54 44±11 27F:27M 26/28 47 ± lO 32 ± 6 56± 11 36 ± 10 36 ± 8
46 68 ± 5 14F:32M 27/19 49 ± 10 33 ± 7 56± l l 39 ± 11 31 ± l l
M-mode measurements of the two groups (reported as mean ± 2 SD) were compared by use of unpaired student t tests; chi-square analysis was used for sex and examination location ratios. Except for age, sex (p = 0.05), and fractional shortening (p = 0.01), the two groups are not significantly different. OPD I OR = the number of studies performed on an ambulatory versus an intraoperative basis. M-mode measurements are of the left atrium (lA), aortic root (RT), and of the left ventricle in diastole (LVd) and systole (L Vs). FS(%) = fractional shortening.
of these images, and (3) to determine the effect of age on image acquisition.
METHODS Patient Population
Monoplane transesophageal echocardiograms were performed prospectively on 100 consecutive patients to demonstrate standard views of the heart and aorta. Fifty-four of the patients were less than or equal to 60 years of age, and 46 patients were more than 60 years old (Table 1). Patients were referred for transesophageal imaging to assess the mitral valve (n = 43), the aortic valve (n = 14), or both valves (n = 9), to exclude endocarditis (n = 13) or an intracardiac source of embolism (n = 7), and to evaluate suspected aortic or pulmonary disease (n = 4), atrial septal defects (n = 3), masses (n = 3), constrictive pericarditis (n = 2), and hypertrophic cardiomyopathy (n = 2). Equipment
Machines used for transesophageal imaging included the Hewlett Packard 1000 (Andover, Mass.) and Acuson 128 (Moutainview, Calif.). Informed consent was obtained before the examination, which was performed in a standard manner that has been previously described. 1-3 During the entire examination, machine settings (depth, gain, and compress) were optimized. A 5 MHz single-plane probe was used. Studies were performed on an ambulatory basis
(n = 53) or before surgery (n = 47) in the surgical suite with the patient under general anesthesia. Before the series was started, we identified, by consensus, 10 standard transverse views of the heart and thoracic aorta that merited evaluation. In addition to routine image acquisition performed to answer clinical questions at hand, the physician attempted to obtain and optimize these views: basal views of (1) the left ventricular inflow or four-chamber view, (2) the right ventricular inflow, which includes an image of the coronary sinus, (3) the left ventricular outflow or five-chamber view, and (4) the short axis of the aortic valve (specific structures were also sought at or near this level, including the atrial septum, left atrial appendage, pulmonary arteries, and coronary arteries); standard transgastric views of the short axis of the (5) mitral valve and the (6) mid -left ventricle and (7) apical left ventricle, and views of the (8) ascending aorta, (9) arch, and (10) descending thoracic aorta. Probe positioning and manipulation to obtain these images was performed in a standard manner that has been previously described. 3•5 On average the intraoperative study took 20 minutes to perform whereas the outpatient study took 30 minutes to perform. The study time varied with the complexity of the cases. Although all standard views were attempted in a timely manner, there was never a clinical event that led to an abbreviated or incomplete study in any patient.
Analysis The videotaped studies were subsequently reviewed by two observers blinded to the patient's clinical data to determine the frequency of image acquisition and the quality of images. Image quality was graded by consensus and scored from 0 to 3. The score assigned was 0 when a meaningful image was not obtained, 1 when the image was comprehensible but border definition was only fair, 2 when most but not all borders had good definition, 3 when all borders had excellent definition. For each view, the grades for image quality were averaged as an overall quality score for the patient population. The obliquity of short-axis images of the aortic and mitral valve and the left ventricle were evaluated as absent for a perfectly round short-axis image, minimal for an imperfect short-axis or oval image, and marked for a very oblique short-axis or very oval image. The effect of age (::S versus >60 years) on the frequency of image acquisition and quality was determined by use of the Fisher's exact test. The
Journal of the American Society of Echocardiography
Cohen et al.
Volume 6 Number 6
579
Table 2 The quality of standard images of the heart and aorta indicated by a grade of 0 (not obtained) to 3 (excellent quality) and the overall quality score. Numbers represent the number of patients or percentages in each category of image quality (N = lOO) Quality Scale Image
Basal views Four-chamber RV inflow Five-chamber Aortic valve short-axis Transgastric MV short-axis LV (mid) short-axis LV (apex) short-axis Aorta Ascending Arch Descending
0
1
2
3
Quality score (all patients)
0 2 16
4 17 6 6
69 64 68 60
27 17 25 18
2.23 1.96 2.17 1.80
18 9 41
21 20 26
50 51 23
20 10
11
1.54 1.84 1.02
7 23 8
23 43 9
54 32 70
16 2 13
1.79 1.23 1.88
l
RV, Right ventricle; MV, mitral valve; LV, left ventricle.
Table 3 Image quality (0 = not obtained to 3 = excellent) and overall score of the left atrial appendage, atrial septum, and left, main, and right pulmonary arteries. Numbers represent the number of patients or percentages in each category of image quality (N = lOO) Quality scale Image
0
1
2
3
Quality score (all patients)
Specific structures Atrial septurn LA appendage LPA MPA RPA
15 55 9 9
l
10 22 20 32 27
61 45 22 47 50
28 18 3 12 14
2.16 1.66 0.73 1.62 1.69
LA, Left atrial; LPA, left pulmonary artery; .MPA, main pulmonary artery; RPA, right pulmonary artery.
obliquity of views of the aortic valve, mitral valve, and mid-left ventricle were compared by chi-square analysis. RESULTS Image Quality Versus Standard View
Table 2 summarizes the frequency of image acquisition and image quality of the l 0 standard image planes. In the majority of patients, most standard basal views and views of the thoracic aorta were obtained and graded l to 3 for image quality; failure to obtain an image (grade 0) occurred infrequendy. Good or excellent transgastric views were less often obtained than the basal transesophageal views (Figure l). Short-axis views of the aortic and mitral valves
were obtained in most patients, though usually oblique to some extent. A perfect short axis of the left ventricle could usually be obtained (Figure 2). In most patients, it was possible to image most of the thoracic aorta. The best images were obtained of the descending thoracic aorta. In contrast, the arch was more difficult to image and was seen adequately in only 77% of patients. Visualization of Specific Structures
Good images of the atrial septum, left atrial appendage, and main and right pulmonary artery were usually acquired (Table 3); however, the left pulmonary artery was usually not seen well. Coronary artery visualization was good but not universal. The left main trunk was seen in 74% of patients whereas its branches, the left anterior descending artery and cir-
580
Journal of the American Society of Echocardiography November-December 1993
Cohen et al.
Image Quality of Standard Views
(n=lOO)
"100 80
60
40 20
.,
¥
,
Basal Views
Figure 1 Bar graph representation of the percentage of patients with good or excellent quality images of standard basal, transgastric, and aortic views. 4C, Four-chamber; RV, right ventricle inflow; SC, five-chamber; AV-SX, aortic valve short-axis; MV, mitral valve; LV, left ventricle; Asc, ascending; Desc, descending.
cumflex artery, were visible in 43% and 67%, respectively. The right coronary artery could be seen in only 51% of patients. Effect of Age on Image Acquisition and Quality
Aging had more of an adverse effect on image quality than on frequency of image acquisition (Table 4). Only the short-axis view of the aortic valve and the view of the ascending aorta were less frequently obtained in the older population. For the other imaging planes, the frequency of acquisition was not significantly different for the two age groups. However, image quality was significantly worse in older patients for the views of the right ventricular inflow, transgastric short-axis view of the mitral valve and left ventricle, and the view of the aortic arch. DISCUSSION
This study documents the frequency with which transverse transesophageal echocardiography generates high-quality images. Compared with transthoracic imaging, it is well known that transesophageal echocardiography has the advantage of being in close
proximity to the heart with less intervening tissue to impede ultrasound penetration. Thus good to excellent definition of edges and endocardium was obtainable in most images. Inadequate visualization of other structures was likely due to their far-field position to the transducer, especially for transgastric views, or due to tracheal interposition, as with the aortic arch. 1-3 In the case of the left pulmonary artery, this structure quickly moves out of the plane of imaging and was often not well seen. Transesophageal echocardiography imaged the posterior heart best. Basal views had a better quality score than transgastric images in which the heart was more in the far field of the ultrasonic beam. Despite their small diameter, coronary arteries could usually be seen, especially the left main and circumflex arteries. The right coronary artery was also often not well seen, at least in part because of its position in the ultrasonic far field. Our success rate was not as high as Zwicky et al., 9 who imaged the proximal right and left coronary arteries in 50 consecutive patients. This discrepancy may be due to differences of technique (including length of time devoted to imaging the coronary arteries), equipment, and definition of adequate coronary visualization. The inability to rotate the transducer or move it beyond the confines of the esophagus and stomach is the most likely cause of the frequent inability to obtain a perfect short-axis view of the aortic and mitral valves and of the left ventricle. This is an important limitation because it may prevent one from localizing pathologic lesions of the leaflet and from diagnosing wall motion abnormalities. Without an adequate short axis, it is more difficult to determine the medial versus lateral relations of mitral pathologic lesions or the site of regurgitation along the coaptation line of the leafiets. 1°For the aortic valve, an adequate short axis is important to determine the number of cusps present. 6 The Effect of Age
This is the first study to show that patient characteristics, such as age, have an impact on transesophageal image quality and acquisition. We speculate that this may be due to a greater incidence of esophageal pathology (such as esophageal dilatation and hiatal hernia) with reduced transducer contact with the mucosa. 11 Dilatation of cardiac chambers and tortuosity of the aorta may have distorted the position of these structures relative to the esophagus and stomach, though M-mode measurements did not show a significant difference between the two age groups. However, it is well known that aging has a
Journal of the American Society ofEchocardiography Volume 6 Number 6
581
Cohen et al.
Figure 2 Bar graph comparison of the obliquity of the aortic and mitral valves (v) and the mid left ventricle (LV) during transverse transesophageal imaging. Views of the left ventricle were significantly less oblique than those of the aortic valve (p < 0.0001) and mitral valve (p < 0.006). Numbers represent the number of patients or percentage in each category of obliquity.
Table 4 The effect of age on image acquisition (%) and quality score for patients < versus > 60 years of age Frequency of acquisition (%) Image plane
:s60years
>60years
Four-chamber RV inflow Five-chamber AoV Short Axis MV Short Axis LV (mid) Short Axis LV (apex) Short Axis Ascending Aorta (Asc) Aortic Arch Descending Aorta (Desc)
lOO lOO lOO
lOO
92 87 95 65 98 80 94
96 98 75 77 87 52 87 74 89
Quality score
p
* *
:s60years
>60years
2.24 2.07 2.20 1.85 1.72 1.92 1.20 1.91 1.20 1.98
2.22 1.83 2.13 1.67 1.33 1.65 0.80 1.65 1.04 1.85
p
* * * * *
*Significant (p :s 0.05) differences between the two age groups for image acquisition frequency (left) and quality (right).
much more profound effect on image quality during transthoracic imaging because of obesity, thickening of the chest wall, and lung disease. 8 Limitations
This study does not establish whether variation in image quality of obliquity during single-plane transesophageal echocardiography has an adverse affect on patient outcome. This question is a major issue because of the greater physician time commitment and expense of transesophageal echocardiography. Our findings represent a reference point for an evolving technology because newer probes may generate better images. Muitiple frequency probes may improve near-field and far-field resolution and multiplane probes may help offset problems with shadowing and image obliquity.H Other factors that af-
feet image quality include machine type, technology, and settings, transducer contact with the esophagus, and operator skill. Though these factors were not specifically examined by our study, utilization of multiple machine types under different clinical settings and the participation of many skilled echocardiographers should make our averaged findings more widely applicable. Single-plane transesophageal echocardiography was able to provide high-quality and detailed views of the heart and aorta in the majority of our patients. REFERENCES l. Gussenhoven EJ, Taams MA, Roelandt JR, et al. Transesophageal two-dimensional echocardiography: its role in solving clinical problems. JAm Coil Cardiol 1986;8:975-9.
582 Cohen et al.
2. Mitcheil MM, Sutherland GR, Gussenhoven EJ, Taams MA, Roelandt JR. Transesophageal echocardiography. J AM Soc ECHOCARDIOGR 1988;1:362-77. 3. Seward JB, Khandheria BK, Oh JK, et al. Transesophageal echocardiography: technique, anatomic correlations, implementation, and clinical applications. Mayo Clin Proc 1988;63:649-80. 4. Omoto R, Kyo S, Matsumura M, et al. Recent technological progress in transesophageal color Doppler flow imaging with special reference to newly developed biplane and pediatric probes. Heidelberg: Springer-Verlag, 1989:21. 5. Seward JB, Khandheria BK, Edwards WD, Oh JK, Freeman WK, Tajik. AJ. Biplanar transesophageal echocardiography: anatomic correlations, image orientation, and clinical applications. Mayo Clin Proc 1990;65:1193-213. 6. Cohen GI, Chan KL. Biplane transesophageal echocardiography: clinical applications of the long-axis plane. J AM Soc ECHOCARDIOGR 1991;4:155-63.
Journal of the American Society of Echocardiography November-December 1993
7. FlachskampfFA, Hoffmann R, Hanrath P. Experience with a transesophageal echo transducer allowing full rotation of the viewing plane: the ornniplane probe [Abstract]. J Am Coil Cardiol1991;17:34A. 8. Bansal RC, Tajik. AJ, Seward JB, Offord KP. Feasibility of detailed two-dimensional echocardiographic examination in adults. Prospective study of 200 patients. Mayo Clin Proc 1980;55:291-308. 9. Zwicky P, Daniel WG, Mugge A, Lichtlen PR. Imaging of coronary arteries by color-coded transesophageal Doppler echocardiography. Am J Cardiol 1988;62:639-40. 10. Stewart WJ, Currie PJ, Salcedo EE, et al. Evaluation of mitral leaflet motion by echocardiography and jet direction by Doppler color flow mapping to determine the mechanism of mitral regurgitation. JAm Coil Cardiol1992;20:l353-6l. ll. Altman DF. The effect of age on gastrointestinal function. Philadelphia: WB Saunders, 1989:162.