ABC | Volume 112, Nº3, March 2019

Review Article Jorge et al Imaging for chemotherapy-related ventricular dysfunction Arq Bras Cardiol. 2019; 112(3):309-316 Echocardiography is the method of choice for serial evaluation of patients undergoing CT, considering availability, cost, and risk. Accurate LVEF calculation should be carried out with the best available method in the echocardiography laboratory, ideally 3D ECHO. In the event that 2D ECHO is used, Simpson’s is the preferred method. 8 When utilizing 2D ECHO, contrast should be associated to improve the accuracy and reproducibility of the measurements of LVEF and volumes. 24 Myocardial two-dimensional strain echocardiography Two-dimensional speckle tracking echocardiography (2D STE) is an accurate technique, which is easy to perform and which has good reproducibility. It measures myocardial deformation through strain and strain rate in the three spatial planes, i.e. longitudinal, circumferential, and radial. Left ventricular systolic global longitudinal strain (LVGLS) has an intra- and inter-observer variability of 4.9% to 8.6%, which is lower than the that of the measurement of LVEF. 25 Its prognostic value is incremental to LVEF for the prediction of all-cause mortality in the general population. 26 Strain and strain rate are attributed to have the capacity to detect subclinical LV dysfunction and cardiac dysfunction related to CT, especially with the use of ANT and TZU. In the study by Charbonnel et al., 27 LVGLS measured before treatment and after low doses (150 mg/m²) of ANT was shown to be a predictor of alterations in LVEF 12 months after the initial dose; CTX occurred in 6 of the 86 patients (7.0%). Pre‑CT evaluation of LVGLS was observed to have a predictor value with an area under the ROC curve of 0.76, with a predictor threshold of (-)19.95% for LVGLS; in the evaluation with low doses, LVGLS provided incremental predictive information on CTX with an area under the ROC curve of 0.82, for a predictor threshold of (-)17.45% for LVGLS. The study by Mousavi et al. 28 demonstrated that LVGLS was a strong predictor of symptomatic HF in pre-CT evaluation of patients undergoing CT with ANT. Age and LVGLS were also predictors of all-cause mortality. LVGLS greater than (-)16% was associated with a 4.7-fold increase in the occurrence of symptomatic HF. Our cardio-oncology research group studied pre-CT LVGLS as a predictor of LV dysfunction and found data (yet to be published) similar to those of the above-cited authors. 27,28 In 52 patients with breast cancer undergoing CT with ANT, who were studied 1 week before CT and after the complete cycle, a significant decrease in LVGLS occurred, which was not accompanied by a decrease of more than 10% in LVEF, which suggests that LVGLS is more sensitive than LVEF in detecting alterations in systolic function. 17 In patients with non-Hodgkin's lymphoma, followed for 6 months, there was a significant reduction in LVGLS and an increase in cardiac troponin T (cTnT); LVEF, however, remained within reference values. When alterations in LVGLS and cTnT were included in the multivariate analysis, only the decrease in LVGLS between the baseline and the third cycle of CT was an independent predictor of CTX; there was, however, no long-term follow-up to determine the clinical relevance of these findings in predicting late cardiac events. 29 Sawaya et al. 30 followed 43 patients with breast cancer treated with ANT and TZU, performing 2D STE and measuring biomarkers at baseline and at 3 and 6 months. High-sensitivity cardiac troponin I (hsTnI) and LVGLS were early markers of subsequent decreases in LVEF. Patients who presented neither a decrease of more than 10% in LVGLS nor an elevation in hsTnI during the third month had a 3% probability of a decreased LVEF in the sixth month. Patients who had a decrease of more than 10% in LVGLS and/or an elevation in hsTnI presented a risk of CTX during the sixth month 9 times greater than those without either of these alterations. The authors infer that the presence of these markers does not mandate the suspension of CT but rather alerts that there is a necessity for closer cardiac monitoring, earlier initiation of “cardioprotective” measures, or the use of alternative, less cardiotoxic therapies. 30 Sawaya et al. 31 followed 81 patients treated with ANT, taxanes, and TZU, 26 of which developed CTX, 5 with symptoms of HF. LVGLS and hsTnI were predictors of CTX, and LVGLS greater than (-)19% was present in all of those who developed symptoms of HF. According to Negishi et al. 32 a relative reduction in LVGLS of more than 11%, with respect to the pre-CT exam, is the best predictor of CTX. In the absence of a pre-CT exam, LVGLS greater than (-)20,5% would also be a predictor of CTX, albeit, with a lower area under the ROC curve (0.67); they thus suggest using relative reduction rather than absolute value. Forty-two patients with ANT and TZU regimes were followed for 12 months, and 10 of them developed CTX. LVGLS, left ventricular systolic global radial strain (GRS), and S’ detected a pre-clinical reduction in the third month, and all 10 who developed CTX had a reduction in LVEF at 6 months. 33 Even though there is a consensus regarding the need for cardiac function monitoring via non-invasive imaging after exposure to CT, the method to be utilized and the monitoring time have yet to be defined. Evidence suggests that a reduction in LVGLS precedes a decrease in LVEF; however, long-term follow-up is necessary to determine the clinical relevance of these findings. Radionuclide ventriculography RNV was the first scintigraphic method for serial evaluation of LVEF in patients undergoing CT. 34,35 Its advantages include excellent accuracy, an estimated intra- and inter-variability of less than 5%, independence from LV geometry, and the possibility of being performed in patients with large body surfaces. 36,37 Gated single-photon emission computed tomography (SPECT) has currently been used to measure LVEF; nevertheless, the American Society of Nuclear Cardiology recommends that serial monitoring be performed by RNV as a Class I indication, whereas the gated blood-pool SPECT technique is a Class IIb indication. 36 Limitations to the routine use of nuclear imaging include patients with arrhythmias; incompleteness of information on the right ventricle, the atria, the valves, and the pericardium; repeated exposure to ionizing radiation; and higher costs compared to echocardiography. In 2016, the American Society of Clinical Oncology published guidelines on the prevention and monitoring of cardiac 311