Comprehensive profiling of cancer associated cells in the blood of breast cancer patients to predict pathological complete response.

Abstract

e12603 Background: Goal of neoadjuvant chemotherapy (NAC) has expanded from increasing the rate of breast conservation to in vivo assessment of response for use of second line therapies that increase survival. Pathological complete response (pCR) is a desirable goal of NAC; however, prediction of pCR for individualized therapeutic decisions remains a vexing problem. Cancer associated cells in the blood provide a promising venue to research response to therapy. Circulating tumor cells (CTCs) are of epithelial origin [CTCs (E+)] that undergo epithelial to mesenchymal transition [CTCs (E+M+)] to become motile and invasive; leading to fully invasive mesenchymal [CTCs (M+)] cells that seed in distant organs, causing metastasis. In addition, cancer associated macrophage-like cells (CAMLs) in the blood of cancer patients are reported to either engulf or aid the transport of cancer cells to distant organs. A comprehensive evaluation of these cancer associated circulating cells through liquid biopsy has a potential to provide a much needed biomarker to predict pCR. Methods: Patients treated with NAC with an intention to cure were enrolled in the study. Blood was collected at 3 time points: pre, mid and post NAC. Microfluidics-based label-free Labyrinth technology was used to isolate CTCs and CAMLs from the blood. Data was collected for patient demographics, biological tumor markers, and treatment response from medical records including pathology reports. Residual Cancer Burden Index (RCBI) and class (pCR vs. residual disease) were analyzed as dependent variables. Independent variables included CTCs, CAMLs and tumor receptor status for univariate and multivariate models to assess the combined effects using R. Results: A total of 57 samples from 19 patients were analyzed. Univariate regression models predicting RCBI or pCR using CTCs and CAML did not show any significant associations. However, in multivariate regression analysis predicting RCBI, an increase in post-NAC CAML and a decrease in post-NAC CTCs were associated with a decrease in RCBI (p=0.005, p=0.035, respectively). In a multivariate logistic regression model predicting pCR, CAML was positively associated with the log-odds of pCR (OR=1.60 [1.01, 2.54]; p=0.047), while CTCs showed a trend in negatively predicting pCR (OR=0.33 [0.10, 1.06]; p=0.063). When controlled for each other, increased CAML and decreased CTCs were associated with increased pCR. While increased M+ was associated with the decreasing log-odds of pCR (OR=0.27 [0.08, 0.99]; p=0.049), neither E+ nor E+M+ was associated with pCR. Conclusions: Increased CAMLs in circulation after treatment combined with lowered CTCs are associated with pCR and lower RCBI. M+ cells could be a factor governing association between CTCs and pCR.