Diagnostic accuracy of electrocardiography and NT-proBNP for mild cardiac dysfunction among long-term breast cancer survivors

<h3>Context:</h3> Long-term consequences of breast cancer treatment may include left ventricular systolic dysfunction (LVSD). This ideally should be detected in primary care, since most specialist follow-up care for breast cancer is discontinued after five years. While echocardiography is the gold standard to diagnose LVSD, its limited accessibility in primary care poses challenges. Therefore, electrocardiography (ECG) and N-terminal pro-brain natriuretic peptide (NT-proBNP) are often used as diagnostic tests by general practitioners (GPs), however their diagnostic accuracy in breast cancer survivors remains unclear. <h3>Objective:</h3> To assess the accuracy of index tests ECG and NT-proBNP, compared to the reference test echocardiography in long-term breast cancer survivors. <h3>Study Design and Analysis:</h3> A cross-sectional diagnostic accuracy study was conducted. <h3>Setting or Dataset:</h3> A dataset was used, consisting of 350 long-term female breast cancer survivors included between 2013 and 2016 in general practices in the Netherlands. <h3>Population Studied:</h3> Breast cancer survivors diagnosed with stage I-III breast cancer who had been disease-free for at least 5 years. Patients were treated with chemo- and/or radiotherapy. <h3>Intervention/Instrument:</h3> We used 12-lead resting ECG and NT-proBNP as index tests. Any abnormality in ECG was documented and classified based on the Novacode criteria. For NT-proBNP a cutoff value of 125 pg/mL was applied. Mild LVSD was defined as left ventricular ejection fraction (LVEF) &lt; 54%. <h3>Outcome Measures:</h3> Sensitivity and specificity. <h3>Results</h3> The median age was 63 years with a median follow-up after breast cancer diagnosis of 10 years. Among the participants, 15.4% had a LVEF &lt; 54%. ECG was abnormal in 50% participants. Regarding NT-proBNP, 35.1% had levels ≥ 125 pg/ml. An abnormal ECG demonstrated a sensitivity of 63.0% and a corresponding specificity of 51.7% for detecting a LVEF &lt; 54%. For NT-proBNP, a sensitivity of 50.9% and a specificity of 66.6% were found. <h3>Conclusions</h3> ECG and NT-proBNP are inadequate diagnostic tools to screen for mild LVSD among asymptomatic long-term breast cancer survivors. Therefore, it seems unnecessary to use ECG and NT-proBNP to screen every woman who underwent potential cardiotoxic treatment for cardiac dysfunction. This differs from the current guideline from the European society of cardiology on cardio-oncology.