Conflict of interest: none declared

Transcription

1 The value of left ventricular global longitudinal strain assessed by three-dimensional strain imaging in the early detection of anthracycline-mediated cardiotoxicity C. Mornoş, A. Ionac, D. Cozma, S. Pescariu, L. Petrescu Institute of Cardiovascular Diseases, Timişoara, Romania Victor Babeş University of Medicine and Pharmacy, Timişoara, Romania Conflict of interest: none declared

2 Background Cardiotoxicity induced by anthracycline remains a significant and unresolved issue in cancer therapy. Histologic myocardial changes can be detected even after low doses. Motoki H, Eur Heart J Cardiovasc Imaging 2012;13(1): Current approaches to surveillance are often inadequate to detect myocardial disease, which can delay medical therapy.

3 Background Despite its feasibility, LVEF is not sensitive enough to reveal subclinical myocardial dysfunction. Tsai HR, Am J Cardiol 2011;107(3):472-7 Strain imaging might provide earlier detection of myocardial dysfunction. 3D-strain analysis improves information on LV deformation by avoiding loss of speckles as it is the case in monoplane 2D-strain analysis. Saito K, J Am Soc Echocardiogr 22 (9):

4 Purpose To assess if early 3D - strain analysis could predict later anthracycline-induced cardiotoxicity.

5 Methods - We analysed 92 patients in sinus rhythm, referred for evaluation of cardiac function before therapy with anthracyclines. - Thirty age-matched healthy subjects who presented for a routine physical evaluation and had normal echocardiographic findings were included as a control group.

6 Methods Exclusion criteria: - prior treatment with cardiotoxic drugs/radiation therapy, - known heart disease, - inadequate echocardiographic images, - LVEF <50%, - renal failure, - trastuzumab required after anthracycline during follow-up 33 pts were excluded. Only 59 pts formed our study group.

7 Methods Clinical examination, ECG, echocardiography (including 3D-strain), troponin T and NTproBNP determination were performed before and after 12 and 36 weeks of treatment.

8 Methods GLS, GRS and GCS were determined using 3D-strain. The method has been previously described. Hansegard J, Cardiovasc Ultrasound 7:18 3D matrix array transducer (Vivid E9, 3V-D probe, GE). A frame rate of 20 to 30 Hz was used. 3D full-volume data sets were acquired in real time. Data analysis was performed using an EchoPAC software workstation (version BT11, 4D Auto LVQ; GE) for semiautomated endocardial surface detection.

9 Methods Alignment with presentation of 4C-, 2C- and 3C apical views, as well as the transversal plane, was performed. Orientation was performed automatically. On each apical view, automatic endocardial border delineation was processed at end diastole and end systole after positioning two landmarks on the mitral annulus and one on the LV apex. Manual correction was performed for an optimal LV delineation. The correct alignment of the contours with the endocardium during the cardiac cycle was controlled. A second semiautomated delineation was made for the epicardium. Last, 3D-strain analysis calculated GLS, GCS and GRS.

10 Methods Cardiotoxicity was defined as a reduction of the LVEF of 5% to <55% with symptoms of HF or an asymptomatic reduction of the LVEF of 10% to <55%. Plana JC, Rev Esp Cardiol 2011;64(5): Procentual changes between baseline and 12 weeks after initiation of chemotherapy versus baseline were calculated (Δ). The cumulative anthracycline dose was calculated using the previously reported formula, including following ratios: doxorubicin = 1.0 daunorubicin = 0.5 idarubicin = 1.6 epirubicin = 0.6 Herait P, Eur J Cancer 1992;28A: Casazza AM, Cancer research 1970;20:16 16

11 *P < 0.05 vs the control group. Baseline characteristics Variable Control Study group (n = 30) (n = 59) Age (years) 50 ± ± 10 Female/male gender, n (%) 19 (59.2%)/13 (40.8%) 35 (59.3%)/24 (40.7%) Body surface area (m 2 ) 24.6 ± ± 3.8 Heart rate (beats/min) 66 ± 7 73 ± 10* Systolic blood pressure (mmhg) 119 ± ± 17 Diastolic blood pressure (mmhg) 75 ± ± 9 Cardiac risk factors Blood pressure >140/>90 mmhg, n (%) 4 (13.3%) 7 (11.8%) Total cholesterol >200 mg/dl, n (%) 2 (6.6%) 4 (6.7%) Diabetes mellitus, n (%) 0 ( 0%) 1 (1.6%) Smoker, n (%) 5 (16.6%) 10 (16.9%) Family history of CAD, n (%) 4 (13.3) 8 (13.5%) Type of cancer Breast cancer, n (%) - 26 (44.1%) Non-Hodgkin s lymphoma, n (%) - 12 (20.3%) Hodgkin s lymphoma, n (%) - 10 (16.9%) Acute lymphoblastic leukaemia, n (%) - 8 (13.6%) Acute myeloblastic leukaemia, n (%) - 2 (3.4%) Osteosarcoma, n (%) - 1 (1.7%)

12 *P < 0.05 vs the control group. Baseline characteristics Variable Control Study group (n = 30) (n = 59) Age (years) 50 ± ± 10 Female/male gender, n (%) 19 (59.2%)/13 (40.8%) 35 (59.3%)/24 (40.7%) Body surface area (m 2 ) 24.6 ± ± 3.8 Heart rate (beats/min) 66 ± 7 73 ± 10* Systolic blood pressure (mmhg) 119 ± ± 17 Diastolic blood pressure (mmhg) 75 ± ± 9 Cardiac risk factors Blood pressure >140/>90 mmhg, n (%) 4 (13.3%) 7 (11.8%) Total cholesterol >200 mg/dl, n (%) 2 (6.6%) 4 (6.7%) Diabetes mellitus, n (%) 0 ( 0%) 1 (1.6%) Smoker, n (%) 5 (16.6%) 10 (16.9%) Family history of CAD, n (%) 4 (13.3) 8 (13.5%) Type of cancer Breast cancer, n (%) - 26 (44.1%) Non-Hodgkin s lymphoma, n (%) - 12 (20.3%) Hodgkin s lymphoma, n (%) - 10 (16.9%) Acute lymphoblastic leukaemia, n (%) - 8 (13.6%) Acute myeloblastic leukaemia, n (%) - 2 (3.4%) Osteosarcoma, n (%) - 1 (1.7%)

13 Results No significant changes during the first 12 weeks of chemotheraphy: Left atrium and LV dimensions, LVEF, Pulmonary artery systolic pressure, Transmitral flow wave velocities (E and A), Myocardial performance index (GMI), NTproBNP.

14 Clinical, echocardiographic and biomarker characteristics Variable Control Study group Baseline 12 weeks Δ (%) Cumulative doses (mg/m 2 ) ± 62* - Heart rate (beats/min) 66 ± 7* 73 ± ± ± 6.2 Systolic blood pressure (mmhg) 117 ± ± ± ± 5.8 Diastolic blood pressure (mmhg) 72 ± ± 9 72 ± ± 4.0 Left atrial diameter (cm) 3.5 ± ± ± ± 5.7 LV end-diastolic diameter (cm) 4.6 ± ± ± ± 4.6 LV end-diastolic volume (ml) 79 ± ± ± ± 9.1 LV end-systolic volume (ml) 32 ± ± ± ± 6.5 LV ejection fraction (%) 60 ± ± ± ± 7.3 Myocardial performance index 0.45 ± ± ± ± 8.2 PASP (mmhg) 25 ± 7 24 ± 9 27 ± ± 3.2 Isovolumic relaxation time (ms) 74 ± ± ± 18* 26.9 ± 19.3 E wave (cm/s) 89 ± ± ± ± 11.2 A wave (cm/s) 74 ± ± ± ± 3.4 LV 3DGLS (%) ± ± ± 2.4* 10.1 ± 6.3 LV 3DGRS (%) 43.2 ± ± ± 5.4* 11.2 ± 6.1 LV 3DGCS ( ) 21 ± ± ± 1.7* 2.2 ± 2.9 NTproBNP (pg/ml) 89 ± ± ± ± 42.1 Troponin T (μg/l) < 0.01 < ± 0.08* 592 ± 857 *P < 0.05 vs baseline

15 Clinical, echocardiographic and biomarker characteristics Variable Control Study group Baseline 12 weeks Δ (%) Cumulative doses (mg/m 2 ) ± 62* - Heart rate (beats/min) 66 ± 7* 73 ± ± ± 6.2 Systolic blood pressure (mmhg) 117 ± ± ± ± 5.8 Diastolic blood pressure (mmhg) 72 ± ± 9 72 ± ± 4.0 Left atrial diameter (cm) 3.5 ± ± ± ± 5.7 LV end-diastolic diameter (cm) 4.6 ± ± ± ± 4.6 LV end-diastolic volume (ml) 79 ± ± ± ± 9.1 LV end-systolic volume (ml) 32 ± ± ± ± 6.5 LV ejection fraction (%) 60 ± ± ± ± 7.3 Myocardial performance index 0.45 ± ± ± ± 8.2 PASP (mmhg) 25 ± 7 24 ± 9 27 ± ± 3.2 Isovolumic relaxation time (ms) 74 ± ± ± 18* 26.9 ± 19.3 E wave (cm/s) 89 ± ± ± ± 11.2 A wave (cm/s) 74 ± ± ± ± 3.4 LV 3DGLS (%) ± ± ± 2.4* 10.1 ± 6.3 LV 3DGRS (%) 43.2 ± ± ± 5.4* 11.2 ± 6.1 LV 3DGCS ( ) 21 ± ± ± 1.7* 2.2 ± 2.9 NTproBNP (pg/ml) 89 ± ± ± ± 42.1 Troponin T (μg/l) < 0.01 < ± 0.08* 592 ± 857 *P < 0.05 vs baseline

16 Results The mean follow-up period: 36 weeks Averaged cumulative anthracycline dose: 256 mg/m 2 Eight patients (13.5%) met the criteria for cardiotoxicity Not associated with future cardiotoxicity: - age, sex, cardiac risk factors, Δ blood pressure, Δ heart rate, - ΔLV and Δ left atrial dimensions, Δ PASP, - Δ myocardial performance index, - Δ isovolumic relaxation time, ΔE, ΔA, - Δ3DGRS, Δ3DGCS, - ΔNTproBNP.

17 Results Variable Univariate logistic regression analysis Multiple logistic regression analysis Odds Ratio P-value 95% CI Odds Ratio P-value 95% CI Cumulative anthracycline doses at 12 weeks ΔLV 3DGLS ΔLVEF ΔTroponin T Δ = procentual changes between baseline and 12 weeks after initiation of chemotherapy versus baseline; EF = ejection fraction; 3DGLS = three-dimensional global longitudinal strain; LV = left ventricle; 95%CI = 95% confidence interval.

18 Results The ROC analysis found Δ3DGLS as the best echocardiographic predictor of future cardiotoxicity: Δ3DGLS (AUC = 0.90, 95%CI = , p <0.001) cumulative dose (AUC = 0.82, 95%CI = , p <0.001) ΔLVEF (AUC = 0.81, 95% CI = , p <0.001). The optimal cut-off for Δ3DGLS was 13.7% with a sensitivity of 88% and specificity of 71%.

19 Discussion The results of our study have demonstrated that in patients treated with anthracyclines, 3D parameters that explore myocardial deformation deteriorated before any LVEF decrease. An early deterioration of 3DGLS predicts the later occurrence of cardiotoxicity.

20 Discussion Our results confirm that LVEF is not acurate enough for early detection of myocardial dysfunction. Fallah-Rad N, J Am Coll Cardiol 2011;57(22): Motoki H, Eur Heart J Cardiovasc Imaging 2012;13(1): Tsai HR, Am J Cardiol 2011;107(3):472-7 Jurcut R, J Am Soc Echocardiogr 2008;21(12): Stoodley PW, Eur J Echocardiogr 2011;12(12): Sawaya H, Am J Cardiol 2011;107(9): Similarly with previous data, we found that isovolumic relaxation time prolongation precedes a significant decrease in LVEF. Stoddard MF, J Am Coll Cardiol 1992; 20 (1):62 9 A recent study sugested that isovolumic relaxation time might be useful when strain imaging is not available. Motoki H, Eur Heart J Cardiovasc Imaging 2012;13(1):

21 Discussion The superiority of strain over LVEF could be explained by the regional pattern the cardiotoxicity (in the early stages the function of some myocardial segments may compensate for others, leading to a preserved LVEF) and by the lower variability. Sawaya H, Am J Cardiol 2011;107(9): In a recent study, 2DGLS had a good reproducibility and accuracy, but 2DGCS and 2DGRS were less reliable, which limits their use in clinical practice. Reant P, J Am Soc Echocardiogr 2012; 25 (1): D-strain improves LV function analysis by avoiding loss of speckles ( 2Dstrain). Reant P, J Am Soc Echocardiogr 2012; 25 (1):68-79 Jenkins C, J Am Coll Cardiol 2004; 44 (4): The reproducibility of 3D-strain > 2D-strain for GCS and GRS analyses. 3D-strain reduced the time of analysis by 25% compared with 2D-strain imaging. Reant P, J Am Soc Echocardiogr 2012; 25 (1):68-79

22 Discussion In our study, the 3D-strain imaging was a sensitive tool for the early detection of subclinical myocardial damage. Δ3DGLS emerged as the only independent predictor of later cardiotoxicity. Δ3DGRS and Δ3DGCS did not reach statistical significance, probable because the reduction in longitudinal function could precede the reduction in radial and circumferential function, as reported in previous studies. Singal PK, N Engl J Med 1998; 339 (13):900 5 Takeda S, Heart 2001;86:52-6 Henein MY, Br Heart J 1994;72:237-42

23 Study limitations The study was limited to a population with good image quality, which explains the high rate of patients excluded from analysis. Variability of strain measurements is a limitation of the technique. Reant P, J Am Soc Echocardiogr 2012; 25 (1):68-79 Jenkins C, J Am Coll Cardiol 2004; 44 (4): Saito K, J Am Soc Echocardiogr 2009; 22 (9): Single center study its reproduction in other centers or by multicenter studies would argue for its validity. A relatively small group of patients Short-term follow-up

24 Conclusions This preliminary study indicates that anthracycline therapy induced early deterioration of 3DGLS, 3DGCS and 3DGRS. Δ3DGLS can be a good predictor of future development of anthracycline - induced cardiotoxicity.

25

26

27 Discussion Toxicity is cumulative and dose dependent. Yeh ET, J Am Coll Cardiol 2009;53(24): LVEF decrease in 16%, 38%, and 65% of patients receiving doxorubicin cumulative doses of 300 mg/m 2, 450 mg/m 2, and 550 mg/m 2, respectively. Yoon GJ, J Am Coll Cardiol 2010;56(20): In our group, the cumulative anthracycline dose was 256 mg/m 2 and the dosis administred at 12 weeks was asociated with later cardiotoxicity on univariate analysis. Of our patients, 13.5% met the criteria for cardiotoxicity.