124 - Tracking Changes in B Cell Antibodies in Cervical Cancer Patients Treated with Chemoradiation and Immunotherapy on GOG 9929
Presenter(s)
S. Kim1, D. Enserro2, H. A. Lankes3, D. DaSilva4, Y. Lin5, K. Moore6, S. Ghamande7, C. Aghajanian8, R. Schilder9, K. Matsuo10, M. J. Birrer11, A. Sharabi1, and J. S. Mayadev12; 1UC San Diego, Moores Cancer Center, Department of Radiation Medicine and Applied Sciences, La Jolla, CA, 2NRG stats, Buffalo, NY, 3NRG Oncology, Philadelphia, PA, 4University of Southern California, Los Angeles, CA, 5Genentech, South San Francisco, CA, 6University of Oklahoma, Oklahoma City, OK, 7Augusta University, Augusta, GA, 8Memorial Sloan Kettering Cancer Center, New York, NY, 9Thomas Jefferson University, Philadelphia, PA, 10University of Southern California Keck School of Medicine, Division of Gynecologic Oncology, Los Angeles, CA, 11University of Arkansas, Little Rock, AR, 12University of California San Diego, San Diego, CA
Purpose/Objective(s): Understanding the underlying immune mechanisms in node-positive cervical cancer remains an unmet need as 40% of patients recur after standard chemoradiation therapy (CRT). GOG9929 (NCT01711515) was a prospective NCI-funded phase I trial that studied the addition of the immune checkpoint inhibitor ipilimumab (ipi, anti-CTLA-4) following CRT and has completed long-term follow-up. Here, we assess changes in the B-cell antibody repertoires after CRT alone and correlate these findings to survival.
Materials/Methods:
34 women with FIGO stage IB2 to IVA node-positive cervical cancer enrolled from December 18, 2012, to August 31, 2016. Treatment consisted of standard CRT with weekly cisplatin followed sequentially by ipi every 21 days for 4 doses. Data on progression-free survival were obtained in collaboration with the NRG Oncology Data Sharing Project. Prospective serum samples were collected at three timepoints (TP): 1) pre-therapy (TP 1), 2) post-CRT but pre-ipi (TP 2), and 3) post-ipi (TP 3). A total of 41 samples across 17 patients were analyzed using the HuProt Human Proteome Microarray, which samples antibody binding to >20,000 human proteins, to define the antibody repertoire. Antigen specificity of IgG and IgM antibodies were sampled. Fluorescence (FL) data were quantile normalized; antigens were classified as being ‘detected’ if Z-score normalized FL was >0.5 for IgG or IgM, separately. Newly detected (ND) antigens were those detected at TP2 but not TP1. Fold-change (FC) was calculated as the ratio of fluorescence at TP2/TP1. Statistical analysis was performed using R software.Results:
14 patients had paired pre- (TP1) and post-CRT (TP2) samples (n=7 for patients with progressive disease and no evidence of disease, each). Using these data, we built an interactive tool to visualize changes to the antigen-specificity of antibodies and identify antigens with an IgG- vs IgM-skewed response. We detected a median of 18 (range: 11-26) and 422 (range: 259-856) ND antigens recognized by IgG and IgM, respectively. Additionally, we identified a median of 180 (range 34-692) and 28 (range: 7-324) antigens recognized by IgG and IgM, respectively, that had a post-CRT FC > 2. Patients with progressive disease had a greater number of ND antigens recognized by IgM (median 473 vs 406, P = 0.0262). There was a trend towards patients with progressive disease showing a greater number of antigens recognized by IgM with a high FC (median 45 vs 21, P = 0.2086).Conclusion: In patients with node-positive cervical cancer treated with CRT on GOG9929, we detected changes in the antigens recognized by the IgM antibody repertoire. These data add to the growing body of literature on the role of B cells in cancer outcomes, and support the further investigation of antibody profiling as a prognostic biomarker in cervical cancer.