The development of a predictive model to estimate cardiotoxic risk for patients with metastatic breast cancer receiving anthracyclines

Background Cardiac toxicity from anthracyclines (ACH) can lead to therapy discontinuation, hospitalization or congestive heart failure (CHF). Since such risk may vary by patient, we developed and tested a risk-prediction tool for cardiac toxicity in metastatic breast cancer (MBC) patients receiving chemotherapy with doxorubicin, either in its traditional (DOX) or pegylated liposomal (PLD) formulation. Methods Data was obtained (n = 509) from a randomized clinical trial of MBC patients assigned either DOX (60 mg/m(2) every 3 weeks) or PLD (50 mg/m(2) every 4 weeks) (O'Brien Ann Oncol 15, 440-449, 2004). Patient, disease and treatment factors were identified for each cycle of therapy. Factors with a P-value of <= 0.25 with <= grade 2 cardiac toxicity following a cycle were retained and included in a generalized estimating equations (GEE) regression model A risk scoring algorithm (range 0-62) was then developed from the final model. Results Factors predictive of cardiac toxicity included an interaction effect between DOX and the number of cumulative cycles, patient age and weight, previous ACH exposure and poor performance status. A ROC analysis had an area under the curve (AUC) of 0.84 (95% CI: 0.79-0.89). A precycle risk score cutoff of >= 30 to < 40 was identified to optimally balance sensitivity (58.5%) and specificity (89.0%). Patients with a score in a given cycle, within or above this threshold, would be considered at high risk for cardiac toxicity. Conclusion Our model provides patient specific risk information that could be helpful in assessing the risks and benefits of anthracyclines in the MBC patients.