Previous studies have shown 3D echocardiography to have lower interobserver and intraobserver variability probability in comparison to 2D echocardiography [6, 7].

The reason is that 3D echo lowers the probability of chamber foreshortening and allows for accurate identification of the true apex of the heart.

LV volume can be quantified without any preestablished assumption of LV geometry. This is of special interest in dilated ventricles with or without segmental alterations.

Jacobs et al. compared 2D and 3D echocardiography against CMR.

Three-dimensional echo correlated highly with CMR imaging values of EDV, ESV, and EF (r = 0.96, r = 0.97 and r = 0.93 for EDV, ESV, and EF, respectively) [13] (Fig. 4.2).

Sugeng proved that 3D echocardiography is also an accurate modality for serial measurements of systolic function [14].

Furthermore, the test-retest variability by a separate operator has shown lower underestimation of LV volumes by 3D echo than by 2D echo when CMR is the reference technique [15, 16].

In a study of Thavendiranathan, non-contrast 3D echo showed minimal detectable EF differences of 6 %, ESV differences of 14 ml, and EDV differences of 34.8 ml.

In this study, contrast administration did not improve variability in measurements of EF or volumes in this population [17].

3D measures are of big interest for follow-up examinations for which exact plane duplication (what is needed in 2D measures) is almost impossible.

Some underestimation of LV volumes due to suboptimal resolution of trabecularization remains in comparison to MRI. This error can be minimized tracing the endocardium underneath the trabeculations and not on top of them [18, 19].

As a general recommendation, 3D echocardiography should be considered in patients with good acoustic windows.

In the past years, abnormalities of diastolic parameters were described.

Stoddard et al. prospectively evaluated 26 patients before administration of doxorubicin and three weeks after cumulative doses. He described an early enhancement of filling time and isovolumetric relaxation time that preceded and predicted a drop in LVEF of >10 %, occurring up to three months later [20].

Other studies confirmed these early changes of diastolic parameters but were not linked to a predictive value [21].

It should be noted that side effects associated with chemotherapy (e.g., diarrhea) might influence the diastolic parameters due to altered loading conditions in follow-up examinations.

A reduction of transmitral E/A ratio is associated with a reduction in longitudinal strain in patients with preserved LV EF late after treatment.

In summary, early chances of diastolic parameters after chemotherapy might not be useful in predicting late-onset systolic dysfunction.

Nevertheless, evaluation of diastolic function should be a part of comprehensive echocardiographic evaluation in the question of cardiotoxicity.

The measurement of LVEF by 2D echo is prone to a high variability, which can be higher than the thresholds used to define cardiotoxicity [17].

Furthermore, reduction of LVEF is often a late phenomenon and can be irreversible. Strain imaging seems to hold promise to detect chemotherapy-induced alterations very precisely and at a very early stage.

Myocardial deformation (strain) and the rate of deformation (strain rate) reflect intrinsic contractility of the myocardium.

The decrease in myocardial systolic function induced by anthracyclines could be documented already 2 h after the first anthracycline dose [2].

Histological studies have shown that cardiomyocyte apoptosis is not restricted to a single myocardial layer.

This is in line with the fact that anthracyclines seem to alter global, circumferential, and radial strains (Fig. 4.3).

Strain imaging requires a good 2D acoustic window to track endocardial borders correctly. In these cases, strain imaging provides an excellent interobserver and test-retest variability [22].

In a study of Sawaya, 81 women with newly diagnosed breast cancer were treated with anthracyclines followed by taxanes and trastuzumab.

Every three months during their cancer therapy (total of 15 months), a follow-up examination by echocardiography was investigated.

Left ventricular ejection fraction, peak systolic longitudinal, radial, and circumferential myocardial strain were calculated. Patients with a pathological longitudinal strain (GLS <19 %) developed later heart failure.

So in this study, global longitudinal strain predicted subsequent cardiotoxicity, whereas LVEF and radial and circumferential strain could not predict cardiotoxicity [23].

In a study of Fallah-Rad et al., patients who received Herceptin had a decrease in global longitudinal and radial strain but not LVEF as early as three months in patients who later developed cardiotoxicity [24].

Strain rate also decreased in half of patients receiving Herceptin in a study of Hare; only 16 % of those with a decrease in strain rate had a drop in EF ≥10 % [25].

However, whether early deformation measurements reflect heart failure in the subsequent years has to be proven by further studies.

Negishi tried to determine the best means of early detection of subsequent reduction of EF in patients with breast cancer treated with trastuzumab (with or without anthracyclines). The strongest predictor of cardiotoxicity was ΔGLS. An 11 % reduction was defined as the optimal cutoff which predicts a subsequent reduction of EF with sensitivity of 65 % and specificity of 94 %.

Patients with a reduction of GLS of <8 % compared with initial measurements were classified as not to be clinically meaningful, whereas those >15 % are very likely to be of clinical significance [26].

Although these findings suggest strain and strain rate as useful parameters with regard to screening for cardiotoxicity, it should be considered that many of these parameters should be used with caution in patients with comorbidities (e.g., LV hypertrophy, obesity, and myocardial infarction) that can affect LV strain. Age and gender also should be taken into consideration when interpreting strain values [27, 28].

Nevertheless, the most important limitation is intervendor variability. Actually, there is now consistent normal value for all vendors. This should be considered in follow-up examinations [29].

Stress echocardiography is helpful in the evaluation of patients with suspected coronary heart disease who will receive regimens that may cause ischemia (fluorouracil, bevacizumab, sorafenib, and sunitinib).

Stress echocardiography has also been used to identify chemotherapy-induced subclinical cardiac dysfunction. Jardfeld documented a reduced ejection fraction at stress compared with rest in ten out of 23 patients who received anthracyclines at childhood due to acute lymphatic leukemia. These young adults were examined median 21 years after remission, whereas 12 healthy controls had no reduced LVEF [30].

Bountioukos evaluated whether repetitive systolic and diastolic cardiac function evaluated by low-dose stress echocardiography can predict anthracycline cardiotoxicity. In this study, the assessment of contractile reserve had no incremental value for the early detection of cardiotoxicity [10].

In another study, high-dose dobutamine stress echocardiography detected subclinical anthracycline-induced cardiotoxicity at lower cumulative doses.

In conclusion stress, echocardiography can detect ischemia whether due to a coronary artery disease as a bystander or induced by chemotherapy.

Stress echocardiography might allow the earlier identification of disease based on recognizing a reduced cardiac reserve, but the results remain unclear.

Echocardiography is regarded as the standard imaging modality to detect cardiotoxicity.

LVEF assessed by 2D echo might not be accurate enough to detect small changes in LV contractility.

Especially in the case of any border line cases, the best available method should be used.

Three-dimensional echocardiography (if available or feasible) or switching to another imaging modality like cardiac MRI can help to identify pathologies.

Contrast echocardiography should be used if a clear endocardial delineation of two continuous segments is not possible.

Strain imaging is a very promising technique that helps to detect subclinical dysfunction.

Global longitudinal strain is the most often used parameter.

Cardiac magnetic resonance (CMR) measures ventricular volumes, mass, and function very accurately; it is reproducible and it is today regarded as the gold standard for these measurements. CMR provides also much information about the morphology and function of the tissues of the heart and important information about paracardiac structures. It shows tissue alterations like myocardial edema and fibrosis. Finally, it helps to identify cardiotoxicity at a very early stage.

Furthermore, CMR is a radiation-free method.

CMR is therefore a very valuable tool especially if a prior echo examination is non-conclusive or shows borderline results.

Radionuclide imaging evaluates functional and structural parameters of the heart and might be used to evaluate cardiotoxicity. Multiple-gated cardiac blood pool imaging (MUGA) is a widely accepted modality to evaluate LVEF due to its high reproducibility and low interobserver and intraindividual variability. Instead to quantify cardiac blood pool, SPECT detects myocardial contours and is less accurate than MUGA, but it can be regarded as an alternative technique for LV measurement. Generally, nuclear imaging offers with different approaches the possibility to detect cardiotoxic alterations at early stages. ¹²³I-MIBG scintigraphy has been shown to be an excellent early predictor of cancer treatment-induced cardiac dysfunction.

Radiation exposition is the main concern for the use of radionuclide imaging, and in pediatric populations, these imaging techniques are not generally used.