Background
The aim of this study was to test the feasibility of the assessment of right ventricular (RV) volumes and function using real-time three-dimensional (3D) transesophageal echocardiographic (TEE) imaging in patients undergoing cardiac surgery.
Methods
One hundred-fifty surgical patients were enrolled: 65 undergoing mitral valve repair, 10 undergoing mitral valve and tricuspid valve repair, four with congenital heart disease, two undergoing Jarvik implantation, 13 undergoing aortic valve surgical replacement, and 56 undergoing transcatheter aortic valve implantation. Real-time 3D TEE acquisition for RV evaluation was performed before and after the surgical procedure and compared with standard two-dimensional multiplane TEE measurements. In a subgroup of 81 patients, 3D transthoracic echocardiographic imaging was also performed. RV volumetric quantification was performed for all data using dedicated software.
Results
Three-dimensional RV analysis was feasible in 98.7% in the preoperative TEE data set and in 92.7% in the postoperative TEE data set. Agreement between 3D transthoracic and transesophageal echocardiography for end-diastolic volume ( r = 0.98; 95% confidence interval [CI], −0.2 ± 13.6 mL), end-systolic volume ( r = 0.97; 95% CI, −2.1 ± 10.2 mL), ejection fraction ( r = 0.77; 95% CI, 1.8 ± 8.2%), and stroke volume ( r = 0.91; 95% CI, 2.0 ± 12.9 mL) was significant. RV parameters were highly reproducible in patients with both normal and dilated RV volumes.
Conclusions
Intraoperative 3D TEE assessment of RV volumes and function is feasible in patients with normal and dilated right ventricles, with good correlation between 3D transthoracic echocardiographic and TEE RV parameters. These measurements could improve the quantitative evaluation of RV function during cardiac surgery.
Right ventricular (RV) dimensions and function are known to be of diagnostic and prognostic importance in patients with cardiac diseases, particularly in subjects undergoing cardiac surgery. In fact, RV failure is associated with a significant decrease in survival in the postoperative period in valvular and vascular surgery, as well as heart transplantation. For these reasons, the accurate assessment of RV function and the early diagnosis of its failure might improve outcomes in the most critical patients undergoing cardiac surgery, leading to better prevention and management strategies. Currently used two-dimensional (2D) echocardiographic RV assessment is challenging because of the peculiar morphology of this ventricular cavity. Recently, a new RV three-dimensional (3D) echocardiographic analysis has been proposed and validated in normal and pathologic subjects using dedicated software applied to 3D transthoracic echocardiographic (TTE) acquisitions. Moreover, a new generation of transesophageal echocardiographic (TEE) probes with a novel matrix-array technique has been introduced, allowing 3D TEE representation of the cardiac structures in real time. The utility of this imaging technique for intraoperative evaluation and guidance in several cardiac surgical and interventional procedures has been demonstrated. The aim of this study was to assess the feasibility and reproducibility of 3D RV analysis using full-volume TEE acquisitions in an intraoperative setting during several cardiac surgical procedures, in patients with both normal and dilated right ventricles.
Methods
The study population consisted of 150 nonconsecutive patients who underwent 2D TEE monitoring during cardiac surgery between June 2009 and January 2011: 65 undergoing mitral valve (MV) repair for prolapse, 10 undergoing MV and tricuspid valve (TV) repair, four with congenital heart diseases, two undergoing Jarvik implantation, 13 undergoing aortic valve surgical replacement (six with aortic regurgitation and seven with aortic valve stenosis), and 56 undergoing transcatheter aortic valve implantation (TAVI) for severe aortic stenosis. Clinical indication for TEE monitoring was based on consensus by anesthesiologists and surgeons. Exclusion criteria were contraindications to transesophageal echocardiography and atrial fibrillation.
The study was approved by the ethics committee of Centro Cardiologico Monzino IRCCS, and informed consent was obtained from all patients.
Data Acquisition
Complete 2D and 3D TEE exams were performed in the operating room after general anesthesia and endotracheal intubation under stable hemodynamic conditions, immediately before and after the surgical procedure. In patients undergoing open-heart surgery, preoperative image acquisition preceded sternotomy. All TEE data sets were acquired using the iE33 echocardiographic imaging platform (Philips Medical Systems, Andover, MA) equipped with a real-time 3D probe (model X7-2t), allowing real-time 3D and full-volume TEE acquisitions in addition to 2D multiplane TEE and Doppler modalities. In our protocol, we acquired TEE data sets in the full-volume modality from seven consecutive cardiac cycles during temporary interruption of ventilator support. Two-dimensional multiplane TEE examinations were also performed in all patients according to a standard protocol, advancing the probe from the upper to the lower esophagus and finally to the transgastric position. The best visualization of the right ventricle, both for 2D multiplane TEE measurements and for volumetric acquisitions, was achieved from standard lower esophageal four-chamber view with appropriate adjustments of the probe through posterior tilting and clockwise rotation. Before the acquisition of 3D data set in full-volume modality, the real-time biplane imaging modality was used to obtain the simultaneous visualization of two orthogonal planes, sagittal (four-chamber view) and coronal (including the outflow tract) ( Figure 1 ), and therefore to confirm complete RV visualization.
In addition, the day before surgery, in a subgroup (group 1) of 81 patients (65 undergoing MV repair, 10 undergoing MV and TV repair, four with congenital heart disease, and two undergoing Jarvik implantation), 3D TTE exams were also performed using the same iE33 echocardiographic imaging platform, equipped with an X3-1 matrix-array probe. These exams were performed to provide more accurate and reliable measurements of chamber size and function and to improve the delineation of valvular and congenital abnormalities, according to routine clinical practice at our institution with regard to the management of these patients. Three-dimensional TTE data sets were acquired from the four-chamber apical view, adapted to improve the visualization of RV cavity.
Data Analysis
The volumetric TTE and TEE data sets were digitally stored and transferred to a workstation with dedicated software platform for offline postprocessing analysis (Research Arena 2.0; TomTec Imaging Systems, Inc., Unterschleissheim, Germany). RV 3D reconstruction was obtained using the corresponding software modality (4D RV-Function; TomTec Imaging Systems) usually applied to 3D TTE acquisitions. Once the RV analysis starts, the software automatically displays a short-axis, an apical four-chamber, and a coronal view independently from the TTE or TEE acquisition modality. Thanks to the selection of TV and MV center and left ventricular (LV) apex, the software was able to orient the volumetric data set in the correct way. End-diastolic volume (EDV), end-systolic volume (ESV), stroke volume (SV), and ejection fraction (EF) were computed as previously described. The difference in EDV between preoperative and postoperative TEE data was also computed (ΔEDV). Examples of the results obtained by the RV analysis software in a patient with a normal RV and in a patient with a dilated RV are shown in Figures 1 and 2 , respectively.
In the midesophageal four-chamber view, 2D TEE RV end-diastolic area (EDA) and end-systolic area were measured. Differences in EDA between preoperative and postoperative TEE data (ΔEDA) and RV fractional area change (FAC), defined as (EDA − end-systolic area)/EDA × 100, were calculated.
Tricuspid annular plane systolic excursion (TAPSE) was also measured as the distance at diastole compared with systole between the apex and the lateral tricuspid annulus.
The feasibility of 3D RV reconstruction was evaluated for both preoperative and postoperative TEE data separately. Moreover, combined feasibility in preoperative and postoperative TEE acquisitions for each patient was measured. As previously described, the quality of 3D reconstructed right ventricles, judged on the basis of RV morphology and on the presence of artifacts throughout the cardiac cycle, was rated as follows: 1 = insufficient (inadequate visualization), 2 = sufficient (sufficient quality or good with artifacts), 3 = good (good images without artifacts), or 4 = optimal (optimal images without artifacts).
Statistical Analysis
Data are presented as mean ± SD for continuous variables (assessed using the Kolmogorov-Smirnov test) and as absolute numbers and relative percentages for categorical variables. In group 1, to test for correlation and agreement between the TTE and TEE measurements, linear regression (Pearson’s correlation coefficient) and Bland-Altman analysis (bias and 95% limits of agreement) were performed, respectively. To assess the relationship between 2D TEE preoperatively and postoperatively, as well as 3D TEE RV parameters, Pearson’s correlation coefficient was computed.
In a random subset of 22 patients with normal RV volumes and function, RV analysis was repeated by a second operator ≥1 week later. During these repeated analyses, investigators were blinded to each other’s and prior measurements. Intraobserver and interobserver variability were assessed as coefficient of variation, defined as the ratio between the standard deviation and the mean of the two measurements, expressed as a percentage. Moreover, Bland-Altman analysis was applied to evaluate the limits of intraobserver and interobserver agreement.
Intraobserver and interobserver variability were also computed in a selected subgroup of 22 patients with RV dilatation (EDV > 180 mL, chosen as an arbitrary cutoff value).
Results
Clinical characteristics of the study population and the feasibility of 3D TEE RV volume assessment are shown in Table 1 , separately for preoperative and postoperative TEE data. The reconstruction of the RV preoperative and postoperative TEE volumes was feasible in 98.7% ( n = 148) and 92.7% ( n = 139) of patients, respectively. The time required for each real-time 3D acquisition was about 38 ± 20 sec, whereas for each analysis, the time required was up to 5 min. The mean quality scores of preoperative and postoperative TEE images were 3.0 ± 0.8 and 2.7 ± 0.8, respectively. In 139 patients (92.7%), the RV reconstructions were obtained on both preoperative and postoperative transesophageal echocardiography. The quality of 3D RV reconstructions for quantitative analysis was judged as inadequate in three of 150 preoperative TEE images (2.0%) and in 10 of 150 postoperative TEE images (6.7%). Three-dimensional reconstruction was scored as sufficient in 32 of 150 (21.3%) and 39 of 150 (26.0%), as good in 62 of 150 (41.3%) and 67 of 150 (44.7%), and as optimal in 53 of 150 (35.3%) and 34 of 150 (22.7%) on preoperative and postoperative transesophageal echocardiography, respectively.
| Procedure | n (women/men) | Age (y) | BSA (m 2 ) | Feasibility (%) | ||
|---|---|---|---|---|---|---|
| TEE imaging, preoperative | TEE imaging, postoperative | TEE imaging, combined preoperative and postoperative | ||||
| MV repair | 65 (21/44) | 59 ± 12 | 1.8 ± 0.2 | 96.9% | 89.2% | 89.2% |
| MV and TV repair | 10 (4/6) | 64 ± 11 | 1.9 ± 0.2 | 100% | 100% | 100% |
| Congenital heart disease | 4 (3/1) | 44 ± 18 | 1.7 ± 0.2 | 100% | 100% | 100% |
| Jarvik implantation | 2 (1/1) | 66 ± 14 | 1.7 ± 0.1 | 100% | 100% | 100% |
| TAVI | 56 (41/15) | 82 ± 5 | 1.7 ± 0.2 | 100.0% | 98.2% | 98.2% |
| AVR | 13 (4/9) | 65 ± 14 | 1.9 ± 0.1 | 100.0% | 76.9% | 76.9% |
| Overall | 150 (74/76) | 68 ± 15 | 1.8 ± 0.2 | 98.7% | 92.7% | 92.7% |
The RV measurements obtained by 3D TTE and TEE imaging are outlined in Table 2 , separately for group 1 ( n = 81), patients with normal right ventricles ( n = 139), and those with dilated right ventricles ( n = 11).
| Variable | Group 1 | Normal RV size | Dilated RV size | |||
|---|---|---|---|---|---|---|
| TTE imaging | Preoperative TEE imaging | Preoperative TEE imaging | Postoperative TEE imaging | Preoperative TEE imaging | Postoperative TEE imaging | |
| EDV (mL) | 119.7 ± 33.8 | 119.8 ± 34.0 | 107.2 ± 23.8 | 110.1 ± 25.3 | 231.2 ± 38.5 | 224.2 ± 38.1 |
| EDV/BSA (mL/m 2 ) | 65.2 ± 17.8 | 66.9 ± 21.0 | 60.2 ± 12.2 | 60.4 ± 14.2 | 133.6 ± 28.7 | 129.0 ± 25.1 |
| ESV (mL) | 56.4 ± 21.1 | 58.6 ± 21.0 | 52.2 ± 14.8 | 54.2 ± 16.3 | 128.2 ± 33.2 | 140.3 ± 38.7 |
| ESV/BSA (mL/m 2 ) | 30.7 ± 11.3 | 32.3 ± 11.6 | 29.2 ± 7.7 | 29.6 ± 8.8 | 74.3 ± 22.3 | 81.5 ± 27. |
| SV (mL) | 63.3 ± 15.8 | 61.3 ± 15.8 | 55.0 ± 13.3 | 55.9 ± 13.0 | 102.5 ± 26.8 | 83.0 ± 23.9 |
| EF (%) | 53.5 ± 6.1 | 51.7 ± 6.1 | 51.6 ± 6.9 | 51.7 ± 7.0 | 44.5 ± 9.4 | 38.0 ± 10.6 |
High Pearson’s correlation coefficients (EDV, r = 0.98; ESV, r = 0.97; EF, r = 0.77; SV, r = 0.91; P < .001) between TTE and preoperative TEE RV parameters were observed in group 1, and relevant scatterplots are depicted in Figure 3 together with the results of Bland-Altman analysis, which show limited bias and narrow limits of agreement between the two techniques. However, we noticed that EDV in particular was underestimated by 3D TTE imaging compared with TEE imaging for greater volumes.
