Background: Doxorubicin and trastuzumab have been described to cause clinical heart failure and asymptomatic declines in left ventricular ejection fraction (LVEF). Initial studies of doxorubicin in the 1970’s defined cardiotoxicity by the clinical syndrome of heart failure without imaging for LVEF determination. Trials with adjuvant trastuzumab used various LVEF cut-offs that were arbitrarily determined. This study aims to compare the predictive value of commonly used definitions of cardiotoxicity associated with chemotherapy for the development of clinical heart failure. Methods: A retrospective chart review was performed on an IRB approved cohort of 638 patients with breast cancer who received doxorubicin, trastuzumab, or both and who had baseline and follow up echocardiograms, collecting LVEF values for each echocardiogram. Clinical heart failure was determined by formal cardiology evaluations denoting clinical heart failure in patients. Four different definitions of cardiotoxicity were compared; American Society of Echocardiography (ASE), Cardiac Review and Evaluation Committee (CREC), Alexander et al, and Schwartz et al (Table). Results: Only 8 patients (1.25%) developed clinical heart failure. The sensitivity, specificity, positive predictive value, and negative predictive value of each definition of cardiotoxicity are listed in the table. Conclusions: Overall there is a low threshold for detection of clinical heart failure in breast cancer therapy. While the ASE definition has the highest combination of sensitivity, specificity, and positive predictive value, it has a sensitivity of only 62.5%. Thus, cardiologists and oncologists should collaborate to develop a definition of cardiotoxicity better correlated to clinical outcomes.