ABC | Volume 110, Nº2, February 2018

Original Article Gripp et al Global Longitudinal Strain Accuracy for Cardiotoxicity Arq Bras Cardiol. 2018; 110(2):140-150 maintaining a statistically significant association in the multivariate models, even when the variables selected on univariate regression were tested two by two. ROC curves to predict cardiotoxicity by use of LV GLS To define the most accurate cutoff point of the absolute LV GLS value on the third month to predict cardiotoxicity on the sixth month, a ROC curve was built (Figure 3A). The LV GLS value of -16.6 showed sensitivity of 80% Table 3 – Echocardiographic characteristics of the patients treated with anthracyclines and trastuzumab - Association with cardiotoxicity. ECHO Variable* Cardiotoxicity P Sim Não n = 5 n = 44 Baseline ECHO EF (%) 64 ± 4.8 68.3 ± 7.7 0.120 GLS (%) -19.3 ± 1.2 -20.5 ± 2.0 0.100 E/E' 8.9 ± 2.5 7.9 ± 1.6 0.450 LV S (cm/s) 7.8 ± 1.1 8.3 ± 1.1 0.380 RV S (cm/s) 12.6 ± 2.1 12.9 ± 2.0 0.760 ECHO 3 months EF (%) 57.6 ± 12.3 67.2 ± 6.4 0.006 EF Dif. 3 months (%) 6.4 ± 16.2 1.1 ± 7.2 0.190 GLS (%) -15.2 ± 2 -19.6 ± 2.1 0.005 GLS Dif. 3 months (%) 4.1 ± 1.6 0.8 ± 1.6 0.008 E/E' 7.1 ± 1.6 8.6 ± 1.9 0.230 LV S (cm/s) 8 ± 0.8 8.5 ± 1.6 0.600 RV S (cm/s) 12.6 ± 2.2 12.9 ± 2.3 0.800 ECHO 6 months EF (%) 52 ± 5.1 67.4 ± 6.6 0.001 EF Dif. 6 months (%) 12 ± 5.2 0.9 ± 9.8 0.004 GLS (%) -15.6 ± 1.1 -19.4 ± 2 < 0.001 GLS Dif. 6 months (%) 3.7 ± 1.8 1 ± 1.6 0.026 E/E' 9 8.2 ± 2.4 0.040 LV S (cm/s) 6.3 ± 0.5 7.8 ± 1.4 < 0.001 RV S (cm/s) 11.8 ± 1.6 13 ± 2 0.200 * Means ± SD; ECHO: Echocardiography; EF: Ejection Fraction (Simpson’s); EF Dif.: Ejection Fraction Difference; GLS Dif.: Global Longitudinal Strain Difference; E/E’: Ratio between E and E’ wave values on Doppler echocardiography; LV S: S wave of the left ventricle; RV S: S wave of the right ventricle. The echocardiographic variables were compared by using paired Student t test, p value ≤ 0.05. Table 4 – Description of the cases with cardiotoxicity. Cases EF LV GLS Baseline 3 Months 6 Months Baseline 3 Months (% ∆GLS) 6 Months 1 66% 52% 58% -19.4% -12.9% (33.50) -17.0% 2 65% 69% 50% -18.7% -16.0% (14.40) -14.3% 3 56% 69% 49% -19.2% -16.5% (14.10) -15.5% 4 64% 58% 45% -21.2% -17.4% (17.90) -15.0% 5 69% 40% 53% -18.0% -13.3% (26.10) -16.4% EF: Ejection Fraction (Simpson’s method); GLS: Global Longitudinal Strain; % ∆GLS: Percentage Variation of Global Longitudinal Strain. and specificity of 95% to predict cardiotoxicity on the sixth month. Similarly, a second ROC curve was built to define the most accurate cutoff point of the percentage reduction of the LV GLS capable of predicting cardiotoxicity (Figure 3B). The LV GLS value of -14% showed sensitivity of 80% and specificity of 99% for that diagnosis. The accuracy of the percentage drop of 14% of the GLS (strain of the third month in regard to that of baseline) was assessed by use of its sensitivity and specificity (100% and 93%, respectively). 145