Background: Despite advances in supportive care, patients undergoing standard chemotherapy for breast cancer are at significant, but unequal, risk for the development of treatment-induced side effects. Current methods for predicting risk are inconsistent and unreliable. Because the pathobiology of many side effects are regulated by the synergistic interaction of simultaneously expressed genes, we reasoned that a novel approach to risk prediction could be achieved by the development of clusters of SNPs defined by BNs. The present study was designed to assess the feasibility of identifying BNs that were associated with regimen-related diarrhea, oral mucositis, nausea and vomiting, fatigue, cognitive dysfunction and neuropathy in patients with breast cancer treated with dose-dense AC+T. The primary endpoint was an area under the receiver operating characteristic curve (aROC) >0.80.

Methods: After informed consent, 78 patients with breast cancer who had received at least 3 cycles of (DD) AC+T were enrolled and saliva samples were collected. Salivary DNA was isolated and analyzed on Illumina Omni microarrays (2.5 × 106 SNPs). Standard supportive therapy for prevention and management of side effects was provided. The frequency and severity of diarrhea, oral mucositis, fatigue, cognitive dysfunction, neuropathy<INS dateTime=2012-09-25T18:35 cite=“mailto:CJordan”>,</INS> and nausea and vomiting were determined using Patient Care Monitor©, a validated patient-reported symptom assessment instrument. BNs were developed for each of the six side effects using Bayesian methods including ten-fold internal cross-validation.

Results: The incidence of clinically significant side effects was higher than anticipated by the literature and are reported in Table 1 along with the accuracy and aROC.

Conclusions: We identified SNP-BNs associated with moderate to severe common side effects of AC+T regimens used to treated breast cancer. These BNs, sourced from salivary SNPs, should provide </DEL> clinicians with accurate actionable information which predicts side effect risk for patients who are scheduled to receive dose-dense doxorubicin/cyclophosphamide and paclitaxel for the treatment of breast cancer.

Citation Information: Cancer Res 2012;72(24 Suppl):Abstract nr P1-15-12.