Speckle Tracking for the Intraoperative Assessment of Right Ventricular Function: A Feasibility Study

Speckle Tracking for the Intraoperative Assessment of Right Ventricular Function: A Feasibility Study Claude Tousignant, MD, FRCPC, Matthias Desmet, MD, Richard Bowry, MBBS, FRCA, Alana M. Harrington, MSc, Jorge D. Cruz, MD, RDCS, and C. David Mazer, MD, FRCP Objectives: Speckle tracking is an ultrasound method that assesses B-mode features to measure tissue displacement and derive deformation parameters. The objective of this study was to assess the feasibility of using speckle tracking in the measurement of right ventricular (RV) longitudinal strain during cardiac surgery using transesophageal echocardiography (TEE). Design: This was a prospective, observational cohort study. Setting: A single university hospital setting. Participants: Twenty-one patients without valvular disease referred for coronary artery bypass graft surgery were studied. Interventions: None. Measurements and Main Results: After the induction of anesthesia and mechanical ventilation, transthoracic echocardiography (TTE) and transesophageal echocardiography (TEE) were used to obtain tricuspid annular plane systolic excursion (TAPSE), RV fractional area of change (FAC), and

2-dimensional strain analysis (speckle tracking) on 3 consecutive heart beats. There was a larger percentage of measurable segments achieved when using TEE. All segments could be analyzed per cardiac cycle in 73% of loops when using TEE and 38% when using TTE. The global strain value was similar using both methods (TEE: ⴚ20.4%, TTE: ⴚ20.1%). The TAPSE could be measured in only 52% of the segments using TTE and 100% using TEE. The FAC could be measured in 90.5% of the loops using TEE and in only 33.3% of the loops using TTE. Conclusions: Perioperative measurements of RV strain using TEE in ventilated patients is feasible. The success rate was higher using TEE in ventilated patients under anesthesia. Differences between the 2 methods were likely the result of differences in 2-dimensional image quality. © 2010 Elsevier Inc. All rights reserved.

R

ing that is not adjustable by the user. Speckle tracking has superior spatial resolution compared with TDI because it is not limited by the Doppler angle; however, its temporal resolution is limited by the 2-dimensional (2D) frame rate. Contrary to TDI, it depends on good 2D image quality. Lagrangian and natural (Eulerian) strain are not interchangeable. For clinical purposes, deformation is usually expressed as Eulerian strain. They are related according to the following relationship in which ␧N is natural strain, ␧ is Lagrangian strain, and t is time7: ␧N(t) ⫽ ln (1 ⫹ ␧[t]). The essential difference between the 2 depends on the point of reference. When describing a moving object, its distance from him/herself can be measured directly (Lagrange). Alternatively, two people (separated by a known distance) can measure its velocity as it passes them both (Eulerian). Unlike TDI, speckle tracking has not been studied extensively in the RV, more specifically in the perioperative period under mechanical ventilation. The hypothesis of this study was that TEE images would be more reliable than TTE images in providing strain measurements using speckle tracking analysis in ventilated patients. More specifically, the objective was to assess the feasibility of using speckle tracking in the measurement of RV longitudinal strain during cardiac surgery using TEE. This was compared with near-simultaneous TTE.

IGHT VENTRICULAR (RV) dysfunction is an independent risk factor and a significant contributor to perioperative morbidity and mortality, especially in the presence of left ventricular dysfunction.1-3 The perioperative assessment of RV function remains problematic. Volumetric or area-based methods such as ejection fraction or fractional area of change (FAC) are unreliable because of complex RV geometry and poor endocardial definition.4 Normal values for RV ejection fraction vary from 32% to 60%.5 Tricuspid annular plane systolic excursion (TAPSE) and velocity remain regional assessments and are influenced by translation and rotation. Deformation parameters such as strain and strain rate are more attractive. Myocardial strain using tissue Doppler imaging (TDI) is limited by the Doppler angle. Using TEE, limited segments align with the Doppler angle. Speckle tracking, however, with its Dopplerangle independence is a more viable alternative in assessing deformation parameters in multiple segments of the RV simultaneously. Strain (␧) is a measure of myocardial deformation. It can be derived using tissue velocity gradients (Eulerian or natural strain) using the following relationship in which V1 and V2 are tissue velocities measured simultaneously in the myocardium and separated by a known distance (L)6,7: t

␧N ⫽

兰 t0

V2 ⫺ V1

KEY WORDS: right ventricular function, transesophageal echocardiography, myocardial strain, transthoracic echocardiography, speckle tracking, intraoperative

L

Speckling patterns are the result of B-mode ultrasound imaging. The patterns created by these natural acoustic markers are stable and move along with tissue. Computerized search algorithms track myocardial speckling patterns frame by frame.8,9 Speckle tracking measures Lagrangian strain, which is based on direct length or displacement measurements. Frame intervals are used for calculations of velocities or strain rates. There is extensive automated postprocess-

From the Department of Anesthesia, The Keenan Research Center at the Li Ka Shing Knowledge Institute, St Michael’s Hospital, University of Toronto, Toronto, Ontario, Canada. Address reprint requests to Richard Bowry, MBBS, FRCA, Department of Anesthesia, St Michael’s Hospital, 30 Bond St, Toronto, ON M5B 1W8, Canada. E-mail: [email protected] © 2010 Elsevier Inc. All rights reserved. 1053-0770/10/2402-0010$36.00/0 doi:10.1053/j.jvca.2009.10.022

Journal of Cardiothoracic and Vascular Anesthesia, Vol 24, No 2 (April), 2010: pp 275-279

275

276

TOUSIGNANT ET AL

METHODS Because this was a prospective, observational cohort study, a convenience sample of 23 patients was used. After institutional review board approval and informed consent were obtained, patients referred for elective primary coronary artery bypass graft surgery were enrolled from October to November 2007. Exclusion criteria included the following: valvular disease or more-than-mild tricuspid regurgitation, redo surgery, pulmonary hypertension, ongoing ischemia, nonsinus rhythm, and contraindications to TEE. After the induction of anesthesia and mechanical ventilation, a pulmonary artery catheter (PAC) was inserted. The heart rate was recorded, and hemodynamic data were obtained with patients in the supine position using an indwelling radial arterial catheter for the mean arterial pressure and a PAC for the pulmonary artery pressures, pulmonary capillary wedge pressure, and the cardiac index using thermodilution. TTE and TEE were performed by using a GE Vivid 7 system (GE, Milwaukee, WI) with a phased array 1.7/3.4-MHz probe (TTE) and a 5.0-MHz transesophageal probe (TEE). Imaging was performed with the patients in the supine position and slightly tilted to the left after the insertion of the PAC. For TEE images, a midesophageal 4-chamber view was obtained with the 2D image sector narrowed to include the right ventricle and optimize frame rate. All TTE images were performed by a trained, certified transthoracic technician (JDC). For TTE images, an apical 4-chamber view was obtained with the 2D sector narrowed to image the right ventricle only and optimize frame rate. Several loops of 5 consecutive heartbeats were recorded for each view with special care taken to obtain the best quality images of the right ventricle without migration of any portion of the RV wall outside the 2D sector. All recordings were obtained during apnea. All images were stored digitally for later offline analysis. A transgastric short-axis view of the left ventricle also was obtained. The left ventricular FAC was calculated from the average of 3 loops. All images were analyzed offline using dedicated software (Quantitative Analysis; Echopac GE, Milwaukee, WI) by 1 observer. Using 2D strain analysis (speckle tracking), a region of interest (ROI) was manually traced along the RV endocardial border and anchored starting at the basal septum and ending at the lateral tricuspid annulus. Care was taken to maintain the ROI within the myocardium. The loops with the best 2D image quality were chosen for analysis. Of the 5 cardiac cycles, 3 consecutive beats were chosen for analysis (best quality). For each

cycle, the right ventricle was then divided into 6 segments (automatic): 3 in the septum (apical, mid, and basal) and 3 in the lateral wall (apical, mid, and basal) (Fig 1A). Quantitative analysis was performed, and a tracking score (pass/fail) was generated for each RV segment by the software (Fig 1A). The image was then assessed manually for proper tracking (Videos 1 and 2 [supplementary videos are available online]). If some segments failed, the ROI was readjusted by moving and/or adjusting the number of anchor points. Proper tracking was again verified manually. The process was repeated in an attempt to generate the largest number of pass scores per cardiac cycle. For each of the 6 RV segments, there were 3 measurements (3 cardiac cycles) per patient per method (TTE and TEE). The number of pass/fail scores was recorded for each RV segment. The strain was recorded for each RV segment producing a pass score. If all 6 segments in 1 cardiac cycle received a pass score, a global strain value was generated by the software and recorded. TAPSE also was measured for each chosen cardiac cycle in every patient using the anatomic M mode from the 2D image. The cursor was applied to the tricuspid annulus, and the orientation was adjusted to follow the annular motion. An average of 3 was used for each method in each patient. The RV FAC was measured from the 2D TEE and TTE images from the same 3 cardiac cycles used for the speckle tracking and TAPSE analysis. The endocardium was traced at end-diastole and end-systole (the smallest area). The FAC was calculated by subtracting the endsystolic area from the end-diastolic area and dividing this value by the end-diastolic area. The average RV FAC and the number of successfully measured FAC measurements were recorded. All values are expressed as mean ⫾ standard deviation. For each segment, there were 63 speckle tracking measurements for each technique (TEE and TTE). A chi-square analysis was performed to compare the proportion of measurable segments between techniques. A 2-way analysis of variance on strain also was performed on the patients in whom all segments could be measured for both techniques, the “matched” patients (n ⫽ 8). The TAPSE for both methods was compared by using the Mann-Whitney rank sum test on matched patients (n ⫽ 11). The RVFAC values were compared by using an unpaired t test on matched patients (n ⫽ 7). A p value ⬍0.05 was considered significant.

Fig 1. Examples of speckle tracking measurements in the right ventricle. (A) A midesophageal 4-chamber view of the RV with the outline of RV segments shown in different colors. A pass or fail score is indicated below for each segment. (B) The peak systolic strain is color coded and overlaid on each RV segment; the legend is on the right. (C) A graphic display of the RV strain values (left axis in %) for 1 cardiac cycle with each segment color coded as shown in A and B. The dotted white line represents the global strain. (D) A color carpet display of RV strain per segment from basal lateral (top) to basal septal (bottom) over 1 cardiac cycle. The center represents the apical portion, which in this case shows larger peak strain than the lateral segments.

SPECKLE TRACKING

277

Table 1. Demographic and Hemodynamic Data Mean ⫾ Standard Deviation

170.9 ⫾ 9.5 82.7 ⫾ 15.2 1.93 ⫾ 0.20 63 ⫾ 9 17/4 61 ⫾ 10 2.04 ⫾ 0.41 34 ⫾ 8 15 ⫾ 3 13 ⫾ 3 72 ⫾ 15 20 ⫾ 6 43.2 ⫾ 9.3

Height (cm) Weight (kg) BSA (m2) Age (y) M/F HR (beats/min) CI (L/min/m2) SVI (mL/m2) PCWP (mmHg) CVP (mmHg) MAP (mmHg) PAP mean (mmHg) LVFAC (%)

Abbreviations: BSA, body surface area; HR, heart rate (beats/min); CI, cardiac index; SVI, stroke volume index; PCWP, pulmonary capillary wedge pressure; CVP, central venous pressure; MAP, mean arterial pressure; PAP, pulmonary artery pressure; LVFAC, left ventricular fractional area of change.

RESULTS

A total of 23 patients were enrolled from October to November 2007. Two patients were excluded because no PAC was inserted after the induction of anesthesia; a final total of 21 patients to be analyzed. The demographics and hemodynamics are presented in Table 1. A total of 63 assessments per segment

were made per technique. The pooled data are presented in Table 2. There was a significant difference in success in acquiring adequate speckle tracking scores between both echocardiographic methods for each segment in which a larger percentage of measurable segments was achieved using TEE (Fig 2A). Overall, a global value for RV strain (when all segments could be analyzed per cardiac cycle) could be obtained in 73% of segments using TEE and 38% of the segments using TTE. TEE and TTE RV longitudinal strain values were compared in patients in whom all segments could be measured using both methods (matched patients n ⫽ 8) (Fig 2B). The septal segments were not significantly different between techniques (TEE v TTE). The lateral wall segments were significantly different between techniques (p ⬍ 0.023). Within the TEE technique, all lateral wall segments were significantly different from all septal segments (p ⬍ 0.004). All septal segments and all lateral segments were not significantly different from each other within each technique. The global strain value was similar using both methods (TEE: ⫺20.4%, TTE: ⫺20.1%). Other measures of RV function including the TAPSE was measured by both TEE and TTE using the anatomic M mode at the lateral wall of the right ventricle. The TAPSE could be measured in all of the segments using TEE and only in 52% of the segments using TTE (Fig 2A). The TAPSE values in both TEE and TTE for matched patients were not significantly different (TEE: 24 ⫾ 6 mm, TTE: 21 ⫾ 3 mm) (Fig 2B). The

Table 2. Strain Results by Segment for All Patients (Pooled Data) Segment

Strain (%) (All Segments, All Patients) Bas sept

TEE TTE Literature Literature Mean ⫾ Standard Deviation Mean ⫾ Standard Deviation Tissue Doppler (TTE) Speckle tracking (TTE) Reference

⫺13.68 ⫾ 4.03

⫺13.56 ⫾ 5.34

⫺18.5 ⫾ 2.1 ⫺21 ⫾ 5 ⫺22.22 ⫾ 2.58

⫺24.1 ⫾ 5.4 ⫺21.46 ⫾ 3.25

Mid sept

⫺16.70 ⫾ 4.46

⫺16.96 ⫾ 4.39

⫺18.8 ⫾ 2.1 ⫺21 ⫾ 5

⫺27.3 ⫾ 8.9

Ap sept

⫺15.21 ⫾ 8.80

⫺17.56 ⫾ 9.31

⫺19.6 ⫾ 3.2 ⫺32 ⫾ 6

⫺25.5 ⫾ 6.6

Ap lat

⫺23.28 ⫾ 6.96

⫺20.55 ⫾ 10.62

⫺30.6 ⫾ 4.2 ⫺32 ⫾ 6 ⫺34 ⫾ 10

⫺28.1 ⫾ 6.7

Mid lat

⫺29.80 ⫾ 5.15

⫺22.48 ⫾ 7.70

⫺30.7 ⫾ 3.3 ⫺27 ⫾ 6 ⫺28 ⫾ 11

⫺26.9 ⫾ 5.3

Bas lat

⫺31.46 ⫾ 8.12

⫺22.77 ⫾ 13.15

⫺30.1 ⫾ 4.9 ⫺19 ⫾ 6 ⫺34 ⫾ 13 ⫺24.42 ⫾ 5.84

⫺44.8 ⫾ 10.2 ⫺24.13 ⫾ 7.0

GLOBAL

⫺20.33 ⫾ 9.71

⫺20.01 ⫾ 3.24

17 14 15 16 17 14 15 17 14 15 17 14 15 20 17 14 15 20 17 14 15 20 16

NOTE. Data from other studies by segment and method (when available) are also presented with references. Abbreviations: Bas sept, basal septal; Mid sept, midseptal; Ap sept, apical septal; Ap Lat, apical lateral; Mid Lat, midlateral, Bas Lat, basal lateral.

278

TOUSIGNANT ET AL

Fig 2. (A) A histogram graph of the percentage of successful measurement of strain, TAPSE, and FAC per technique (TTE and TEE) for all patients. Values are shown below the graph. (B) Strain, TAPSE, and FAC values in matched patients (where both TTE and TEE techniques were successful). *A significant difference between techniques (TTE and TEE) for the RV segment (strain) (p < 0.023) and between techniques for RVFAC (p ⴝ 0.003).

FAC could be measured in 90.5% of the loops using TEE and in only 33.3% of the loops using TTE (Fig 2A). The RVFAC measured by TEE (45.5%) was significantly different from that measured by TTE (39.6%) (p ⫽ 0.003) in matched patients (Fig 2B). DISCUSSION

This study shows that speckle tracking analysis of RV longitudinal strain using TEE in ventilated patients undergoing cardiac surgery is feasible. There was a 73% success rate in obtaining all 6 segments simultaneously. TEE was more successful than using TTE (38% success) in this ventilated patient population. Longitudinal strain or shortening of the myocardium during systole may represent a component of the measure of work performed by the right ventricle.10 It is dependent on preload, afterload, and contractility. Strain rate, or the rate at which this shortening is performed, is an indicator of power or, alternatively, how quickly the work can be performed. Strain patterns may be helpful in the assessment of RV function such as strain distribution, the presence of dyskinesis, and postsystolic events.11,12 To the authors’ knowledge, no study has investigated the use of speckle tracking in the assessment of RV longitudinal strain in the perioperative period using TEE. In the present study, longitudinal strain values in the right ventricle were not distributed equally. The septal strain values were lower than those of the lateral wall (Table 2). The fibers in the lateral wall of the RV are predominantly longitudinal, whereas the architecture in the septum is more complex, including longitudinal and circular fibers.13 This may have contributed to the difference in strain between both regions. The strain values in the lateral wall were higher when measured using TEE. This may have been the result of better image quality using TEE in ventilated patients. No study has examined all the walls of the RV using tissue Doppler with TEE because of the limitations of the Doppler angle. Most studies examining RV strain have used tissue Doppler with TTE and have found overall that septal strain values were lower than

those of the lateral wall.14,15 The strain values were on the whole slightly lower than those in the literature, perhaps a result of a difference in patient population, clinical setting, or technique (Table 2). The patients were anesthetized, had cardiovascular disease, and were receiving medications such as ␤-blockers. Unfortunately, published values for strain in the RV can vary significantly even within normal controls and between techniques14-17 (Table 2). This may be caused by Doppler angle effects as well as technique including machine settings. Other measures of RV function such as TAPSE and RVFAC were consistent with normal values found in the literature.18,19 There was a difference in RVFAC between both techniques. This may have been the result of image quality, which was much poorer with TTE in this ventilated patient population. However, both results were within the normal range.5 Overall, measuring TAPSE and RVFAC was more successful using TEE, a reflection of the superior TEE image quality in these ventilated patients. In ventilated patients in the perioperative period, the most reliable method was TEE. The most reliable measurement was the TAPSE because it could be measured in 100% of instances. TAPSE, however, only measures lateral wall function. The FAC, measured in 90% of instances, assesses overall function but does not give information on segmental function. The strain by speckle tracking offers global and segmental information and was measurable in all segments simultaneously in up to 73% of instances. LIMITATIONS

Speckle tracking is a technology that is dependent on good 2D image quality. Poorer 2D image quality in TTE, for example, results in a poorer success rate. There is a great deal of postprocessing involved in speckle tracking. Furthermore, user input variability also may affect results. Compared with tissue Doppler, speckle tracking is independent of the Doppler angle. However, precision may still be affected by the direction of motion. Lateral resolution may play an important role in the accuracy of strain determined via

SPECKLE TRACKING

279

speckle tracking. In TTE, longitudinal RV wall motion is in the vertical axis; therefore, longitudinal strain occurs along the scanning plane (base to apex). In TEE, however, RV wall motion is more horizontal and crosses many scanning planes (lobes) (Fig 2A and B). There is poorer lateral resolution when measuring deep in the sector. Tissue dropout or stationary reverberations can contribute to poor tracking of speckles and may lead to erroneous strain measurements. Therefore, the placing of anchor points becomes very important. Furthermore, ensuring that the ROI remains in the myocardium and that proper tracking has been verified is very important. Optimizing frame rate also is important in acquiring images for speckle tracking analysis. If the frame rate is too low, too much time will have elapsed between sampling, and speckling

patterns will be “lost,” resulting in a failure of tracking. If the frame rate is too high, lateral resolution may be impaired. CONCLUSION

Perioperative measurement of RV strain using TEE is feasible and more reliably obtained than with TTE in ventilated patients undergoing cardiac surgery. Septal strain was lower than lateral wall strain, and values obtained with TEE in the lateral wall were higher than those obtained with TTE. Differences between the 2 methods may have been the result of differences in 2D image quality. RV strain measurements may be helpful for the clinical management of patients undergoing cardiac surgery or in the intensive care unit; however, there is significant offline user input to obtain results that may limit its applicability in the operating room.

REFERENCES 1. Ghio S, Tavazzi L: Right ventricular function in advanced heart failure. Ital Heart J 6:852-855, 2005 2. Ghio S, Gavazzi A, Campana C, et al: Independent and additive prognostic value of right ventricular systolic function and pulmonary artery pressure in patients with chronic heart failure. J Am Coll Cardiol 37:183-188, 2001 3. Maslow D, Regan M, Panzica P, et al: Precardiopulmonary bypass right ventricular function is associated with poor outcome after coronary bypass grafting in patients with severe left ventricular systolic dysfunction. Anesth Analg 95:1507-1518, 2002 4. Burgess MI, Bright-Thomas RJ, Ray SG: Echocardiographic evaluation of right ventricular function. Eur J Echocardiography 3:252262, 2002 5. Lang RM, Bierig M, Devereux RB, et al: Recommendations for Chamber Quantification: A Report from the American Society of Echocardiography’s Guidelines and Standards Committee and the Chamber Quantification Writing Group, Developed in Conjunction with the European Association of Echocardiography, a Branch of the European Society of Cardiology. J Am Soc Echocardiogr 18:14401463, 2005 6. Teske AJ, De Boeck BWL, Melman PG, et al: Echocardiographic quantification of myocardial function using tissue deformation imaging, a guide to image acquisition and analysis using tissue Doppler and speckle tracking. Cardiovasc Ultrasound 5:27, 2007 7. D’Hooge J, Heimdal A, Jamal F, et al: Regional strain and strain rate measurements by cardiac ultrasound: Principles, implementations and limitations. Eur J Echocardiogr 1:154-170, 2000 8. Amundsen BH, Helle-Valle T, Edvardsen T, et al: Noninvasive myocardial strain measurement by speckle tracking echocardiography: Validation against sonomicrometry and tagged magnetic resonance imaging. J Am Coll Cardiol 47:789-793, 2006 9. Leitman M, Lysysansky P, Shir V, et al: Two dimensional strain. A novel software for real-time quantitative echocardiographic assessment of myocardial function. J Am Soc Echocardiogr 17:1021-1029, 2004

10. Schenk S, Popovic ZB, Ochiai Y, et al: Preload-adjusted right ventricular maximal power: Concept and validation. Am J Physiol Heart Circ Physiol 287:H1632-H1640, 2004 11. Pislaru C, Abraham TP, Belohlavek M: Strain and strain rate echocardiography. Curr Opin Cardiol 17:443-454, 2002 12. López-Candales A, Dohi K, Rajagopalan N, et al: Right ventricular dyssynchrony in patients with pulmonary hypertension is associated with disease severity and functional class. Cardiovasc Ultrasound 3:23-33, 2005 13. Ho SY, Nihoyannopoulos P: Anatomy, echocardiography, and normal right ventricular dimensions. Heart 92:i2-i13, 2006 (suppl) 14. Pettersen E, Helle-Valle T, Edvardsen T, et al: Contraction pattern of the systemic right ventricle. Shift from longitudinal to circumferential and absent global ventricular torsion. J Am Coll Cardiol 49:2450-2456, 2007 15. Kowalski M, Kukulski T, Jamal F, et al: Can natural strain and strain rate quantify regional myocardial deformation? A study in healthy subjects. Ultrasound Med Biol 27:1097-1097, 2001 16. Stefani L, Toncelli L, Gianassi M, et al: Two-dimensional tracking and TDI are consistent methods for evaluating myocardial longitudinal peak strain in left and right ventricle basal segments in athletes. Cardiovasc Ultrasound 5:7, 2007 17. Pirat B, McCulloch ML, Zoghbi WA: Evaluation of global and regional right ventricular systolic function in patients with pulmonary hypertension using a novel speckle tracking method. Am J Cardiol 98:699-704, 2006 18. Lopez-Candales A, Rajagopalan N, Saxena N, et al: Right ventricular systolic function is not the sole determinant of tricuspid annular motion. Am J Cardiol 98:973-977, 2008 19. Lopez-Candales A, Rajagopalan N, Gulyasy B, et al: Comparative echocardiographic analysis of mitral and tricuspid annular motion: Differences explained with proposed anatomic-structural correlates. Echocardiography 24:353-359, 2007 20. Kjaergaard J, Sogaard P, Hassager C: Quantitative echocardiographic analysis of the right ventricle in healthy individuals. J Am Soc Echocardiogr 19:1365-1372, 2006