Surgical inspection identified 37 and 25 patients with and without chordal rupture, respectively. RT3DTEE correctly diagnosed the presence or absence of chordal rupture in 59 of these patients (95.2%) as opposed to 43 patients (69.4%) using 2D TEE (difference in proportions = 25.8%, CI [16.6%, 37.9%], P < 0.001). Agreement of TEE and surgical reference regarding the scallops of origin of ruptured chordae was almost perfect using RT3DTEE (κ = 0.82, CI [0.68, 0.93]) and fair using 2D TEE (κ = 0.27, CI [0.14, 0.44]). Chordal rupture of the posterior leaflet is compared in Figures 3 and 4. Online Figure 2 (see Supplemental Digital Content 2, http://links.lww.com/AA/A433; see Appendix for online figure legends) shows chordal rupture of the anterior leaflet.
The analysis by scallop is listed in Table 6. The accuracy of RT3DTEE was significantly higher for the most frequent lesions, scallop P2 (difference in proportions = 24.2%, CI [13.1%, 36.7%], P < 0.001), and for A2 (difference in proportions = 6.5%, CI [0.2%, 15.4%], P = 0.046). There were 2 cases of papillary muscle rupture and both were correctly identified by either method.
The surgeon described 22 and 40 patients with and without leaflet clefts, respectively. RT3DTEE was feasible for the detection of surgically relevant MV leaflet clefts (Fig. 5) with a substantial agreement of RT3DTEE and intraoperative findings (κ = 0.65, CI [0.44, 0.81]). Five clefts reported by surgical inspection were not detected by RT3DTEE. Conversely, 7 clefts reported by RT3DTEE were not reported by surgical inspection (Table 6), which equals a false-positive rate of 17.5%. All true-positive clefts detected with RT3DTEE were localized accurately.
Three-dimensional zoom recordings had an average of 10.3 ± 2.1 frames per second compared with 42.8 ± 8.4 in 2D recordings. Image quality was rated excellent in 64.5%, satisfactory in 21.0%, and poor in 14.5% of the 3D examinations compared with 74.2% excellent, 24.2% satisfactory, and 1.2% poor in 2D. The mean image quality rating in 2D TEE was better than in RT3DTEE (P = 0.03). Poor 3D image quality was attributable to (a) an incomplete capture of the MV structures in a single 3D clip (n = 3), (b) interruption of acquisition by the surgical procedures (n = 2), and (c) poor echocardiographic signal quality (n = 4). None of these patients was excluded from the statistical evaluation.
Inter- and Intraobserver Agreement
Inter- and intraobserver agreement was substantial for both methods with a median κ value of 0.75 (range [0.35; 0.83]) and 0.80 (range [0.44; 0.89]) for inter- and intraobserver interpretations, respectively.
Mitral Valve Prolapse
RT3DTEE was more accurate than 2D TEE for the diagnosis of MVP in patients with MR. However, the predominant prolapse was correctly identified in the majority of cases using both RT3DTEE and 2D TEE. Although 2D TEE was reliable for the localization of a major prolapse of the MV, RT3DTEE was more accurate in detecting all prolapsing scallops in complex cases or excluding the involvement of adjacent scallops. The close visual resemblance of the 3D images to the surgical inspection is an advantage of RT3DTEE and was confirmed by the κ analysis, which illustrated a large difference of 2D versus 3D agreement with the surgeon when evaluating the entire valve for prolapsing scallops (difference in proportions = 33.9%) and for chordal ruptures (difference in proportions = 25.8%) (Figs. 3 and 4). However, superiority of RT3DTEE was not proven for all scallops. This effect is related to the small number of lesions in less frequently involved scallops (Table 6). Differences were significant in scallops P1 and P2, which accounted for two-thirds of all prolapsing scallops. Similar results were obtained for chordal rupture with significance for scallop P2 and a trend for P1. The frequent incidence of P2 lesions is consistent with the published literature22 and may reflect surgical patient selection.23,24
The differentiation of P1 versus P2 or a combined involvement is challenging in 2D TEE images (Figs. 3 and 4). Slight differences in patient anatomy or probe position may cause alterations in 2D sections, leading to intersection with the scallop that is adjacent to the expected one. The challenge increases when 2 scallops appear simultaneously in one 2D section (Fig. 3a). Diagnosis in these cases was better using RT3DTEE, in which the physiological interscallop indentations were observed in 3D en-face views, which facilitated the localization of MVP.
Three-dimensional images obviate these difficulties by including the entire information of the valve that is necessary for spatial orientation in 1 image (Fig. 2). MPR planes, which are very similar to 2D sections, can be obtained from the same heartbeat and can be arranged freely, to intersect the MV annulus anywhere.
Although 2D TEE can detect chordal ruptures sensitively,25 localization remains challenging (Fig. 3) (online Fig. 2, http://links.lww.com/AA/A433; see Appendix for online figure legends) because these fine structures only partially appear in the image, often without visualization of their leaflet insertion. Updating the results on RT3DTEE in previous studies,7,26 we found that the insertion at the leaflet of the majority of ruptured chordae can be detected accurately (Fig. 4) (online Fig. 2, http://links.lww.com/AA/A433). The visualization of multiple ruptured chordae in one “bird’s eye view” provides an advantage for the precise description of complex MV pathology.
Even though 22 patients in our study were reported to have MV leaflet clefts, the anatomical definition of leaflet clefts is not well defined and their clinical significance is uncertain. In addition to visualization, it is challenging to distinguish physiological interscallop indentations27 from clefts in the mitral leaflets that cause regurgitation and require surgical intervention. These clefts can arise from interscallop indentations that exceed what Victor and Nayak27 refer to as anatomically regular “slits” (online Fig. 1, see Supplemental Digital Content 1, http://links.lww.com/AA/A432; see Appendix for online figure legends), and intrascallop clefts in abnormal places20 (Fig. 5). We found substantial agreement of RT3DTEE with the surgical inspection. Seventeen clefts were correctly visualized using RT3DTEE, providing new information that may be used for surgical planning. Two reasons likely account for the number of false-positive clefts in RT3DTEE: the overestimation of interscallop indentations at low gain settings, and the exceptionally frail leaflet tissue that sometimes presents as a tissue defect in RT3DTEE (Fig. 6) and leads to a false-positive diagnosis. When imaging with a low number of frames per second, a cleft may be visible in one single frame only, making it difficult to determine true anatomical clefts from imaging artifacts.
Limitations of 2D TEE
Intraoperative 2D TEE is the standard method for the evaluation of MV pathology and provides accurate data with proven feasibility in the operating room.2,25,28 However, spatial orientation is often difficult with 2D TEE, and interpretation of 2D images remains challenging and requires a high level of expertise. Clinicians limited to 2D imaging are often confronted with ambiguous 2D recordings, in which the affection of adjacent scallops cannot be distinguished with certainty. Information regarding the position of the probe at the moment of acquisition and the impact of manual adjustments on images is essential for correct interpretation, thus making it highly operator-dependent. Ahmed et al. suggested that the limitations of conventional 2D TEE lie in its inability to display the entire surface of the mitral leaflets in the short axis, and true localizations may differ from the established standard 2D sections.29,30 Further factors leading to these aberrations are anatomical differences of patients, the left lateral positioning in minimally invasive surgery, and cases with complex MV disease.6 Sensitivity has been reported between 50% and 96% for the diagnosis of MVP with 2D TEE.2,6,9,28 The accuracy of 2D TEE in our study is consistent with these numbers and reflects the aforementioned challenges. Color Doppler helps with orientation in addition to 2D imaging of the MV apparatus.
Limitations of RT3DTEE
Although the spatial resolution of RT3DTEE technology has advanced notably, temporal resolution remains limited. Three-dimensional zoom recordings offer approximately 25% of the frame rate of 2D imaging, which may lead to an inability to diagnose scallop pathologies associated with motion, such as ruptured chordae and leaflet clefts. RT3DTEE also has a greater tendency to misidentify normal structures as pathology (Fig. 6), especially when the image gain is reduced to suppress image noise.27,29 In our study, this led to a false-positive rate of 17.5% for leaflet clefts with respect to the surgical findings.
RT3DTEE acquisition of one 3D zoom-mode recording takes approximately 1 to 2 minutes. Considering that one recording has to be made compared with a full set of 2D recordings, acquisition of RT3DTEE is considerably faster. However, thorough interpretation of a 3D zoom-mode recording requires gain changes, rotation, and MPR analysis after acquisition, which is not necessary with 2D TEE interpretation.
In this study, the average image quality was rated higher with 2D TEE than RT3DTEE. This difference may have been attributable to a longer experience with 2D acquisition. RT modes such as the 3D zoom mode can obviate the problem of artifacts caused by the stitching of subvolumes with faulty alignment caused by arrhythmia or movements.31 Unfortunately, it is restricted by its maximum image window width, and anatomically significant structures may be cut off in large valves.
At present, there is only one vendor who offers RT3DTEE for intraoperative use, and the analysis of RT3DTEE data requires specific software, which limits widespread clinical availability.
Limitations of This Study
The comparison of 2D versus RT3DTEE was performed after surgery to ensure proper blinding of the interpreters. The relative diagnostic inferiority of 2D TEE, as shown in our study, might have been influenced by the offline analysis of 2D images. In everyday clinical practice, the intraoperative physician is using color flow Doppler along with 2D imaging. Two-dimensional interpretation also depends on the experience of the echocardiographers and the knowledge of the probe depth or right-left turn. This information was not recorded and could have impaired 2D interpretations, whereas it is not necessary for RT3DTEE interpretation because the images contain all of the information.
In several cases, more leaflets were reported as prolapsing in RT3DTEE images than during surgical inspection. Using TEE before institution of cardiopulmonary bypass, MV function can be evaluated in a relatively physiological state (assuming normal ventricular loading conditions). However, the surgeon is evaluating the anatomical features in a resting heart devoid of blood during the surgical inspection period. Consequently, smaller lesions may be missed by the surgeon despite a correct diagnosis with TEE.
In this prospective study of intraoperatively acquired TEE imaging, RT3DTEE proved feasible and more accurate than 2D for the detection, localization, and description of MVP and ruptured chordae tendineae. The new generation of RT3DTEE imaging facilitates the detection of leaflet clefts. RT3DTEE provides a full view of the MV annulus, which can be further “dissected” offline so that any component can be interrogated. Two-dimensional TEE is an indispensable tool in MV surgery, but RT3DTEE offers distinct improvements in spatial orientation and visualization of valvular pathology. The increasing sophistication of MV surgery requires that the surgeon be supplied with the best available imagery. We recommend RT3DTEE as a routine supplement to intraoperative MV imaging.
Appendix: Online Figure Legends
Online Figure 1: Complex mitral valve prolapse (MVP) involving all scallops. Severe mitral valve (MV) bi-leaflet prolapse disease that involved all scallops. All images are from the same patient. Red lines mark the annular level in both 2-dimensional (D) and real-time 3-dimensional transesophageal echocardiography (RT3DTEE) images. (a-d) Standard 2D ME sections. It seems as if there is no A1 prolapse in (a) and no A2 prolapse in (d). In (b), A2 prolapse is visible, in (c), A1 and A2 prolapse is visible, which would contradict (a) and (d). (c): in this complex case involving bi-leaflet prolapse of different extents, it is difficult to specify which scallops are intersected. An MPR would be helpful. (e-f): RT3DTEE images. (e) The annular level and the commissural line are marked in white. (*) marks PL interscallop indentations. In (e), all scallops, with the exception of A2, are clearly prolapsing. (f) The same image turned clockwise and cropped from the PMC shows that A3 and A2 are also prolapsing above the annular level (red line). [AL: anterior leaflet; ALC: anterolateral commissure; Ao: aortic valve; LV: left ventricle; ME: midesophageal; MPR: multiplanar reconstruction; PL: posterior leaflet; PMC: posteromedial commissure]
Online Figure 2: Anterior prolapse with chordal ruptures. (a-d) Four standard midesophageal (ME) views of the mitral valve (MV) in 2-dimensional (D) transesophageal echocardiography (TEE) reveal prolapse and chordal rupture of the anterior leaflet. (e) Four Chordae (black arrows) can be distinguished in the real-time (RT) 3D en-face view at end systole, including their exact origin, corresponding to the surgeon’s view. (f) Surgeon’s view (minimally invasive access) with prolapse and broad tear-off of chordae in A2 (black arrows).
Name: Maximilian Dominik Hien, MD.
Contribution: This author assisted in the study design, conducted the study, analyzed the data, and wrote the manuscript.
Attestation: Maximilian Dominik Hien has seen the original study data, reviewed the analysis of the data, approved the final manuscript, and is responsible for archiving the study files.
Name: Helmut Rauch, MD.
Contribution: This author assisted in the study design, conducted the study, and supervised the performance of the study.
Attestation: Helmut Rauch has seen the original study data, reviewed the analysis of the data, and has approved the final manuscript.
Name: Artur Lichtenberg, MD.
Contribution: This author assisted in the conduct of the study and the revision of the manuscript.
Attestation: Artur Lichtenberg has seen the original study data, reviewed the analysis of the data, and has approved the final manuscript.
Name: Raffaele De Simone, MD.
Contribution: This author conducted the study, assisted in the writing and revising of the manuscript, and provided valuable advice.
Attestation: Raffaele De Simone has seen the original study data, reviewed the analysis of the data, and has approved the final manuscript.
Name: Marc Weimer, DSc.
Contribution: This author assisted in data analysis, calculated the statistics, and acted as a consultant.
Attestation: Marc Weimer has seen the original study data, reviewed the analysis of the data, and has approved the final manuscript.
Name: Oriana Amanda Ponta, MSc.
Contribution: This author assisted in data analysis, calculated the statistics, and acted as a consultant.
Attestation: Oriana Amanda Ponta has seen the original study data, reviewed the analysis of the data, approved the final manuscript, and is responsible for archiving the study files.
Name: Christian Rosendal, MD, DESA.
Contribution: This author assisted in the study design, conducted the study, analyzed the data, and wrote the manuscript.
Attestation: Christian Rosendal has seen the original study data, reviewed the analysis of the data, and has approved the final manuscript.
This manuscript was handled by: Martin J. London, MD.
The authors thank Thomas Müller, Tanja Sacconi, Christoph Schramm, and Johann Motsch for their assistance in data acquisition. They also thank the Research Training Group 1126 and the German Research Foundation (DFG) for supporting the research project.
a Two-day intensive training course on 3D TEE and clinical practice.
b Tanja Sacconi, Maximilian Hien, Christoph Schramm, Thomas Müller, and Johann Motsch.
1. Shanewise JS, Cheung AT, Aronson S, Stewart WJ, Weiss RL, Mark JB, Savage RM, SearsRogan P, Mathew JP, Quinones MA, Cahalan MK, Savino JS. ASE/SCA guidelines for performing a comprehensive intraoperative multiplane transesophageal echocardiography examination: recommendations of the American Society of Echocardiography Council for Intraoperative Echocardiography and the Society of Cardiovascular Anesthesiologists Task Force for Certification in Perioperative Transesophageal Echocardiography. Anesth Analg. 1999;89:870–84
2. F1ter GP, Isselbacher EM, Rose GA, Torchiana DF, Akins CW, Picard MH. Accurate localization of mitral regurgitant defects using multiplane transesophageal echocardiography. Ann Thorac Surg. 1998;65:1025–31
3. Johnson ML, Holmes JH, Spangler RD, Paton BC. Usefulness of echocardiography in patients undergoing mitral valve surgery. J Thorac Cardiovasc Surg. 1972;64:922–34
4. Goldman ME, Mora F, Guarino T, Fuster V, Mindich BP. Mitral valvuloplasty is superior to valve replacement for preservation of left ventricular function: an intraoperative twodimensional echocardiographic study. J Am Coll Cardiol. 1987;10:568–75
5. Onnasch JF, Schneider F, Falk V, Mierzwa M, Bucerius J, Mohr FW. Five years of less invasive mitral valve surgery: from experimental to routine approach. Heart Surg Forum. 2002;5:132–5
6. Beraud AS, Schnittger I, Miller DC, Liang DH. Multiplanar reconstruction of threedimensional transthoracic echocardiography improves the presurgical assessment of mitral prolapse. J Am Soc Echocardiogr. 2009;22:907–13
7. Agricola E, Oppizzi M, Pisani M, Maisano F, Margonato A. Accuracy of realtime 3D echocardiography in the evaluation of functional anatomy of mitral regurgitation. Int J Cardiol. 2008;127:342–9
8. Ben Zekry S, Nagueh SF, Little SH, Quinones MA, McCulloch ML, Karanbir S, Herrera EL, Lawrie GM, Zoghbi WA. Comparative accuracy of two- and threedimensional transthoracic and transesophageal echocardiography in identifying mitral valve pathology in patients undergoing mitral valve repair: initial observations. J Am Soc Echocardiogr. 2011;24:1079–85
9. Manda J, Kesanolla SK, Hsuing MC, Nanda NC, AboSalem E, Dutta R, Laney CA, Wei J, Chang CY, Tsai SK, Hansalia S, Yin WH, Young MS. Comparison of real time twodimensional with live/real time threedimensional transesophageal echocardiography in the evaluation of mitral valve prolapse and chordae rupture. Echocardiography. 2008;25:1131–7
10. Pothineni KR, Inamdar V, Miller AP, Nanda NC, Bandarupalli N, Chaurasia P, Kirklin JK, McGiffin DC, Pajaro OE. Initial experience with live/real time threedimensional transesophageal echocardiography. Echocardiography. 2007;24:1099–104
11. Ma N, Li ZA, Meng X, Yang Y. Live three-dimensional transesophageal echocardiography in mitral valve surgery. Chin Med J (Engl). 2008;121:2037–41
12. Sugeng L, Shernan SK, Salgo IS, Weinert L, Shook D, Raman J, Jeevanandam V, Dupont F, Settlemier S, Savord B, Fox J, Mor-Avi V, Lang RM. Live 3-dimensional transesophageal echocardiography initial experience using the fully-sampled matrix array probe. J Am Coll Cardiol. 2008;52:446–9
13. Grewal J, Mankad S, Freeman WK, Click RL, Suri RM, Abel MD, Oh JK, Pellikka PA, Nesbitt GC, Syed I, Mulvagh SL, Miller FA. Real-time three-dimensional transesophageal echocardiography in the intraoperative assessment of mitral valve disease. J Am Soc Echocardiogr. 2009;22:34–41
14. Sugeng L, Shernan SK, Weinert L, Shook D, Raman J, Jeevanandam V, DuPont F, Fox J, Mor-Avi V, Lang RM. Real-time three-dimensional transesophageal echocardiography in valve disease: comparison with surgical findings and evaluation of prosthetic valves. J Am Soc Echocardiogr. 2008;21:1347–54
15. Aubert S, Acar C. Gaping cleft or commissure—an under-rated cause of residual mitral insufficiency following valve repair: case reports. J Heart Valve Dis. 2009;18:290–1
16. Amin A, Davis M, Auseon A. Isolated cleft posterior mitral valve leaflet: an uncommon cause of mitral regurgitation. Eur J Echocardiogr. 2009;10:173–4
17. Quill JL, Hill AJ, Laske TG, Alfieri O, Iaizzo PA. Mitral leaflet anatomy revisited. J Thorac Cardiovasc Surg. 2009;137:1077–81
18. Shah PM. Current concepts in mitral valve prolapse: diagnosis and management. J Cardiol. 2010;56:125–33
19. Levine RA, Triulzi MO, Harrigan P, Weyman AE. The relationship of mitral annular shape to the diagnosis of mitral valve prolapse. Circulation. 1987;75:756–67
20. Tango T. Equivalence test and confidence interval for the difference in proportions for the pairedsample design. Stat Med. 1998;17:891–908
21. Cohen J. Weighted kappa: nominal scale agreement with provision for scaled disagreement or partial credit. Psychol Bull. 1968;70:213–20
22. Muller S, Muller L, Laufer G, Alber H, Dichtl W, Frick M, Pachinger O, Bartel T. Comparison of three-dimensional imaging to transesophageal echocardiography for preoperative evaluation in mitral valve prolapse. Am J Cardiol. 2006;98:243–8
23. Woo YJ, Seeburger J, Mohr FW. Minimally invasive valve surgery. Semin Thorac Cardiovasc Surg. 2007;19:289–98
24. DiBardino DJ, ElBardissi AW, McClure RS, Razo-Vasquez OA, Kelly NE, Cohn LH. Four decades of experience with mitral valve repair: analysis of differential indications, technical evolution, and long-term outcome. J Thorac Cardiovasc Surg. 2009;139:76–83
25. Hozumi T, Yoshikawa J, Yoshida K, Yamaura Y, Akasaka T, Shakudo M. Direct visualization of ruptured chordae tendineae by transesophageal two-dimensional echocardiography. J Am Coll Cardiol. 1990;16:1315–9
26. Salcedo EE, Quaife RA, Seres T, Carroll JD. A framework for systematic characterization of the mitral valve by real-time threedimensional transesophageal echocardiography. J Am Soc Echocardiogr. 2009;22:1087–99
27. Victor S, Nayak VM. Definition and function of commissures, slits and scallops of the mitral valve: analysis in 100 hearts. Asia Pac J Thorac Cardiovasc Surg. 1994;3:10–6
28. Grewal KS, Malkowski MJ, Kramer CM, Dianzumba S, Reichek N. Multiplane transesophageal echocardiographic identification of the involved scallop in patients with flail mitral valve leaflet: intraoperative correlation. J Am Soc Echocardiogr. 1998;11:966–71
29. Ahmed S, Nanda NC, Miller AP, Nekkanti R, Yousif AM, Pacifico AD, Kirklin JK, McGiffin DC. Usefulness of transesophageal three-dimensional echocardiography in the identification of individual segment/scallop prolapse of the mitral valve. Echocardiography. 2003;20:203–9
30. Lambert AS, Miller JP, Merrick SH, Schiller NB, Foster E, Muhiudeen-Russell I, Cahalan MK. Improved evaluation of the location and mechanism of mitral valve regurgitation with a systematic transesophageal echocardiography examination. Anesth Analg. 1999;88:1205–12
31. Fischer GW, Salgo IS, Adams DH. Realtime three-dimensional transesophageal echocardiography: the matrix revolution. J Cardiothorac Vasc Anesth. 2008;22:904–12
Supplemental Digital Content
© 2013 International Anesthesia Research Society