Developing A Predictive Model For Disease Status In Breast Cancer Patients Using Th1 And Th2 Serum Cytokine Profiles.

10078 Background: Using the Luminex multiplex assay, we have reported significant correlations between serum levels of MCP-1, eotaxin, and IL-6 and disease characteristics in breast cancer (BCa) patients. We now examine the potential of utilizing a general cytokine profile to develop a statistical model to predict certain disease states in these patients. Methods: Sera from 36 BCa patients (24 node-negative, 12 node-positive, 12 normals) were analyzed using the Luminex assay for levels of 21 cytokines (IL-1α, 1b, 2, 4, 5, 6, 8, 10, 12, 13, 15, 17, IFN-γ, G-CSF, GM-CSF, TNF-α, IP-10, MIP-1α, RANTES, MCP-1, eotaxin). Logistic regression models were used to assess if a binary outcome variable (Y) can be predicted by using serum cytokine levels (X). The area under a receiver operating characteristic (ROC) curve (c) was used to assess the potential utility of a biomarker. The larger the value of c, the better the biomarker. Results: MCP-1 was found to be a possible predictor of the presence of BCa while other potential biomarkers were IL-13, MIP-1α and eotaxin. The higher the MCP-1 level, the greater the likelihood that the patient would have BCa. Similar relationships applied to the other potential biomarkers. Among BCa patients, GM-CSF seemed to be a good predictor of nodal status with lower levels of GM-CSF predicting positive nodes. Other potential biomarkers with a similar expression pattern for nodal status were MCP-1, IL-6 and IL-5. Due to small sample sizes, we were unable to examine a potential “panel” of cytokines to develop a prognostic algorithm based on serum analysis. Conclusions: MCP-1, which was previously shown to be elevated in BCa, may also have some predictive value linking the presence of disease and disease severity as measured by nodal status. Other prominent cytokines from earlier studies (MIP-1α, eotaxin, IL-6) also displayed some possible predictive value. Our results warrant studying a larger population in order to establish a unified prognostic formula for BCa based on serum cytokine levels. [Table: see text] No significant financial relationships to disclose.