Background: Breast cancer (BC) is a heterogeneous disease and difficult to treat once metastasis occurs. Thus, determining blood molecular biomarkers that allow for early diagnosis would help to identify BC patients at higher risk. MicroRNAs (miRs) are short sequences (18-22 base pairs) and promising candidates as biomarkers due to their high stability in blood. Using a direct blood cell-free miR (cfmiR) next generation sequencing (NGS) -based assay that requires no extraction, we assessed 2083 cfmiRs in plasma samples from BC patients. We hypothesize that specific cfmiR signatures detected in plasma from BC patients are associated with early BC diagnosis. Those cfmiRs represent unique patterns that may allow for monitoring BC tumor burden and early detection.Methods: Two cohorts of pre-operative plasma samples from BC patients (n= 158), one cohort of pre-operative plasma samples from patients with BC brain metastasis (BCBM, n= 5), and a cohort of female normal donors’ plasma samples (n=20) were assessed by HTG EdgeSeq miR whole transcriptome assay. All primary BC tumors were histopathology staged (TNM, AJCC 8th) after surgery and divided into two different cohorts: the BC-1 cohort (n= 80, including stage I= 39, II= 37, and III= 4 patients), and the BC-2 cohort (n= 78, including stage I= 42, II= 33, and III= 3 patients). All of the samples were analyzed using DESeq2 normalization and interrogated for differentially expressed (DE) miRs using a cut-off of 30 counts, FC>1.2 or <-1.2, and FDR (p<0.05). We utilized the BC-1 group as the training cohort and the BC-2 and the BCBM cohorts were used to validate the observations in BC-1.

Results: In a retrospective study the cfmiR profiles obtained from BC-1 patients were compared to the cfmiR profiles obtained from normal donors’ plasma samples. Briefly, 328 cfmiRs were DE in plasma samples from the BC-1 cohort compared to normal donors’ plasma samples, of which 184 were upregulated and 144 were downregulated. To validate our findings, we screened the plasma samples from the BC-2 cohort (n=78). The results showed that 181 of 361 DE cfmiRs were downregulated and 180 were upregulated. By comparing both BC cohorts we found 269 DE cfmiRs consistently changing, which included 82% and 74.5% of the cfmiRs identified in BC-1 and BC-2 cohorts, respectively. In addition, we compared the cfmiR expression of BCBM (n=5) and normal donors’ plasma samples. Of the 300 DE cfmiRs, 30 were downregulated and 270 were upregulated. By integrative analysis, a 59 cfmiR signature observed in BCBM plasma samples was consistently identified in both BC-1 and BC-2 cohorts. Also, 172 cfmiRs were found exclusively in BCBM and not in primary BC tumors. Receiving operative characteristic curve analysis showed that the 59 cfmiR signature was able to distinguish primary BC from female normal donors’ (BC-1 AUC= 0.910, p<0.05 and BC-2 AUC= 0.922, p<0.05, respectively). Using the 59 cfmiR signature, the first five components of the principal component analysis showed 85.26% and 85.27% mean cumulative variance for BC-1 and BC-2 respectively.Conclusions: Specific cfmiR patterns were associated with primary BC tumors and may have potential utility for early BC diagnosis, including disease recurrence. A 59 cfmiR signature detected in primary BC patients was able to detect BCBM and may have potential clinical utility in monitoring patients at higher risk of metastasis. Further analysis are in progress to validate our observations.

Citation Format: Matias Alberto Bustos, Negin Rahimzadeh, Xiaoqing Zhang, Kevin D. Tran, Grace Chang, Carol Nishikubo, Dave S.B. Hoon. Identification of cell-free microRNA signatures in pre-operative plasma from breast cancer patients [abstract]. In: Proceedings of the 2020 San Antonio Breast Cancer Virtual Symposium; 2020 Dec 8-11; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2021;81(4 Suppl):Abstract nr PS2-17.