Early Detection and Prediction of Cardiotoxicity in Chemotherapy Treated Patients

Editor's note: This article is based on Sawaya et al, Early Detection and Prediction of Cardiotoxicity in Chemotherapy-Treated Patients, Am J Cardiol. 2011 Mar 1.

Identification of patients who develop myocardial dysfunction during treatment with chemotherapy is an important issue. In their study, the authors sought to assess whether early echocardiographic measurements of myocardial deformation (strain rate imaging) and biomarkers (cardiac TnI and NT-proBNP) could predict the development of chemotherapy-induced cardiotoxicity in patients treated with the combination of anthracyclines and trastuzumab1.

A total of 43 patients diagnosed with HER-2-overexpressing breast cancer either scheduled to be treated with anthracyclines/trastuzumab (n=33) or trastuzumab after previous anthracycline treatment (n=10) were enrolled at four institutions. Patients with LVEFs <50% were excluded.

Echocardiogaphy and biomarker sampling were performed at baseline, three and six months. LVEF was measured using a modified Simpson's biplane method. Peak systolic radial and circumferential strain was measured by averaging the peak systolic strain values in all six segments of the short-axis view. In five patients for radial strain and 11 patients for circumferential strain, >2/6 segments could not be reproducibly analyzed at one time point, and these values were excluded. In addition, in 60% to 70% of the patients, the apex was not analyzable at all time points, so only the basal and midwall segments of the apical views we used to calculate global peak systolic longitudinal strain.

NT-proBNP and troponin I were measured using the Siemens cTnI Ultra assay (lower limit detection 0.015 μg/L, 10% CV at 0.04 ug/L). Values of NT-proBNP >125 pg/ml and of troponin >0.015 μg/L were considered elevated. Cardiotoxicity was defined as either a change in LVEF of ≥5% to <55% with heart failure symptoms, or an asymptomatic reduction of the LVEF of ≥10% to <55%.

Baseline measurements were normal for all 43 patients. After treatment, LVEF decreased significantly by a mean of 8% at six months. Longitudinal, circumferential and radial all decreased significantly from baseline to six months (mean change, 11-17%). No change in mean cTnI or NT-proBNP was noted when all patients were considered together.

Nine patients (21%) met the criteria for cardiotoxicity, one at three months and eight at six months. Echocardiographic parameters that predicted cardiotoxicty at six months included decreases in longitudinal strain and radial strain, but not circumferential strain or echocardiographic indexes of diastolic function. An elevated cTnI at three months was predictive of cardiotoxicity at sixmonths. In contrast, neither the change in NT-proBNP between baseline and three months nor an NT-proBNP level higher than normal limits at three months predicted cardiotoxicity.

An elevated cTnI at three months (p <0.02) and a decrease in longitudinal strain between baseline and three months (p <0.02) were the only independent predictors of later cardiotoxicity. When both a 10% decrease in longitudinal strain and an elevated cTnI was present at three months, the SN was 55%, with a SP of 97% and a PPV of 83% for identifying cardiotoxicity. If either a 10% decrease in longitudinal strain or elevated cTnI was present at three months, SN increased to 89% and NPV from 89 to 97%, but SP fell to 65%.The authors concluded that an early decrease in myocardial strain or elevation in cTnI predicted the later occurrence of cardiotoxicity.


Identification of patients undergoing chemotherapy who will ultimately develop heart failure as a result of treatment remains an important goal, especially as newer chemotherapeutic agents are developed. Traditionally, assessment of LVEF has been used to determine the cardiac impact of chemotherapy. However, it has important limitations. Measurement is susceptible to changes in loading conditions, inability to visualize all segments, variability in reproducibility, and the fact that deterioration in LVEF represents a relatively late stage of cardiac dysfunction. Earlier detection of subclinical cardiac dysfunction could lead to identifying, intervening, and possibly preventing late adverse cardiac outcomes. Either the frequency and/or dose of treatment could be altered, and supplemental treatments that could potentially mitigate or prevent further damage could be added.

Although previous studies have investigated both echocardiography and biomarkers for identifying those at risk for later cardiotoxicity, there is very little information comparing both techniques in the same cohort of patients. Therefore, information from this study is welcome.

Elevations in troponin after chemotherapy has been found not uncommonly, occurring after numerous types of treatment. Elevations correlated with cardiac events in large patient cohorts2. The frequency of detection will depend on the timing of the sampling, as well as the assay and cut off values used3. The use of a more sensitive assay, as was used in the current study1, will have the advantage of detecting more minor degrees of injury. The frequency of cTnI elevation in the current study, 28% appears consistent with prior studies. Using more sensitive cTn assays does have the complication of making comparisons with prior studies difficult, and therefore the predictive value of each new generation of assay will have to be validated.

Although other studies have found elevations in natriuretic peptides identify subsequent cardiotoxity3, in the current study, neither the baseline value nor the change in NT-proBNP between zero and three months was predictive of a later decrease in the LVEF. This may relate to timing of when NT-proBNP was sampled. In one study, sustained elevations of NT-proBNP at 72 hours was predictive of later systolic and diastolic abnormalities4.

Newer echocardiographic indicies, such as tissue Doppler imaging5 as well myocardial strain and strain rate imaging (6) are sensitive measures of tissue deformation and have the ability to detect subclinical myocardial dysfunction in a number of different diseases6. In the current study, changes in strain were predictive of subsequent cardiotoxicity. Although circumferential strain was not predictive, this may be a result of excluding 11 patients due to suboptimal images, and points out a potential limitation of this technique.

Several factors may explain the increased sensitivity of strain compared to the LVEF in the detection of early cardiotoxicity. It is possible that the chemotherapy-induced cardiotoxicity has a regional pattern such that the function of abnormal myocardial segments is compensated for others, leading to a preserved LVEF. Measurement of strain, at least in experienced labs, may be less variable (although both techniques clearly depend on an adequate imaging window).

There are important limitations in this study. The number of patients included was relatively small, and the follow-up period was fairly short. Often cardiac dysfunction is not evident for years after treatment. In addition, more frequent marker sampling could have identified more patients with injury, improving cTnI’s predictive ability, as elevations found on serial sampling was a better predictor of subsequent cardiotoxicity2. The cut-off value used for this study, 0.015 ug/L, is at the lower limit of detectability, which may affect reproducibility.

Given the relative inexpensive nature of biomarkers compared to imaging, a higher negative predictive value is more important. If sufficiently high, patients with negative markers could potentially avoid frequent imaging, while those with positive cTn would be identified as requiring more intensive monitoring. The presence of subclinical echocardiographic abnormalities could further select out a group of patients in whom greater surveillance is necessary. Hopefully, larger studies, such as the ongoing LITE Trial7, will provide more information on multi-modality identification of patients at high risk for cardiotoxicity.


  1. Sawaya H, Sebag IA, Plana JC, et al. Early Detection and Prediction of Cardiotoxicity in Chemotherapy-Treated Patients. Am J Cardiol. 2011 Mar 1.
  2. Cardinale D, Sandri MT, Colombo A, et al. Prognostic value of troponin I in cardiac risk stratification of cancer patients undergoing high-dose chemotherapy. Circulation. 2004;109:2749-54.
  3. Cardinale D, Sandri MT. Role of biomarkers in chemotherapy-induced cardiotoxicity. Prog Cardiovasc Dis. 2010;53:121-9
  4. Sandri MT, Salvatici M, Cardinale D, et al. N-terminal pro-B-type natriuretic peptide after high-dose chemotherapy: a marker predictive of cardiac dysfunction? Clin Chem. 2005;51:1405-10.
  5. Tassan-Mangina S, Codorean D, Metivier M, et al. Tissue Doppler imaging and conventional echocardiography after anthracycline treatment in adults: early and late alterations of left ventricular function during a prospective study. Eur J Echocardiogr 2006; 7:141–146.
  6. Marwick TH. Measurement of strain and strain rate by echocardiography: ready for prime time?, J Am Coll Cardiol. 2006;47:1313–1327.
  7. Lotrionte M, Palazzoni G, Natali R, Comerci et al. Appraising cardiotoxicity associated with liposomal doxorubicin by means of tissue Doppler echocardiography end-points: rationale and design of the LITE (Liposomal doxorubicin-Investigational chemotherapy-Tissue Doppler imaging Evaluation) randomized pilot study. Int J Cardiol. 2009;135:72-7

Keywords: Anthracyclines, Antibodies, Monoclonal, Humanized, Biomarkers, Breast Neoplasms, Cardiotoxins, Echocardiography, Female, Heart Failure, Natriuretic Peptide, Brain, Systole, Troponin, Troponin I

< Back to Listings