Background: Early recognition of anthracycline-induced cardiomyopathy may reduce morbidity and mortality in children, but risk stratification tools are lacking. This study evaluates whether electrocardiogram (ECG) changes precede echocardiographic abnormalities in children with anthracycline-induced cardiomyopathy. Methods: We performed a retrospective analysis of 589 pediatric cancer patients who received anthracyclines at a tertiary referral center. ECG endpoints were sum of absolute QRS amplitudes in the 6 limb leads (ΣQRS(6 L)) and corrected QT interval (QTc). Cardiomyopathy was defined by echocardiogram as ejection fraction < 50%, shortening fraction < 26%, or left ventricular end-diastolic diameter z-score > 2.5. Results: Median age at start of therapy was 7.8 years (IQR 3.7-13.6); median follow-up time was 3.6 years (IQR 1.1-5.8). 19.5% of patients met criteria for cardiomyopathy. Male sex, race, older age at first dose, and larger body surface area were associated with development of cardiomyopathy. A 0.6 mV decrease in ΣQRS(6 L) and 10 ms increase in QTc were associated with an increased risk of developing cardiomyopathy with hazard ratios of 1.174 (95% CI = 1.057-1.304, p = 0.003) and 1.098 (95%CI = 1.027-1.173, p = 0.006) respectively. Kaplan-Meier estimates showed a lower chance of cardiomyopathy-free survival for QTc ≥ 440 ms and ΣQRS(6 L) ≤ 3.2 mV over time. After controlling for confounders, total anthracycline dose predicted a decrease in ΣQRS(6 L) and an increase in QTc independent of cardiomyopathy status (p = 0.01 and p < 0.001 respectively). Cardiotoxic radiation did not predict changes in ECG parameters. Cardiomyopathy was associated with increased mortality (34% versus 12%, p < 0.001). Conclusion: In children receiving anthracyclines, decrease in ΣQRS(6 L) and QTc prolongation are associated with increased risk of developing cardiomyopathy. ECG is a potential non-invasive risk stratification tool for prediction of anthracycline-induced cardiomyopathy and requires prospective validation.