Abstract 5679: Predicting prognosis in papillary thyroid carcinoma: Clues in the tumor microenvironment

Abstract

Abstract The incidence of thyroid carcinoma is dramatically increasing in the U.S. Papillary thyroid carcinoma (PTC) represents 85% of thyroid cancer and has a ~15% recurrence rate and a 5-10% metastasis rate. Current molecular assays fail to predict the biologic behavior of these tumors. Here we perform a computational immunogenomic evaluation of 568 PTCs and correlate immune infiltrate with tumor behavior. Computational immunogenomic analysis was performed on RNA sequencing data from 568 PTCs in The Cancer Genome Atlas (TCGA) using the CIBERSORT deconvolution analytic tool for characterizing cellular composition (Newman et al. Nature Methods 2015). Immunologic cell fractions from tumors with global p values <0.05 were characterized as having “high confidence” and those with p values ≥0.05 were characterized as having “low confidence.” The immune cell fractions were correlated with tumor pathologic T stage and lymph node (LN) stage. Immune cell fractions for activated dendritic cells (DCs) were categorized as low (<0.5%), medium (0.5-3.0%) and high (>3.0%); and for follicular helper T cells (TFH) were low (<1%), medium (1-5%), and high (>5%). A high CD8/T regulatory cell (CD8/Treg) ratio was defined as ≥4. For the logistic regression, pathologic tumor T stage was divided into low stage (T1 and T2) and high stage (T3 and T4), and LN stage was classified as no metastasis (N0) and LN metastasis (>N0). Among tumors with high CIBERSORT confidence, there is a significant negative correlation between high TFH fraction and advanced pathologic tumor stage (OR=0.12, 95%CI=0.02-0.62, p = 0.01) and a significant positive correlation between high activated DC fraction and advanced T stage (OR= 8.75, 95%CI= 1.7-45, p=0.009). The observed odds ratio for cases with a medium infiltrate of activated DCs is 2.31 (CI: 0.93-5.7, p=0.07). Finally, among tumors with high CIBERSORT confidence, a high CD8/Treg ratio is negatively associated with the presence of LN metastases (OR=0.39, 95%CI=0.19-0.81, p=0.01). Here we present a robust computational analysis of immune cell infiltrates and their correlation with tumor and lymph node stage. We demonstrate that increased DCs and decreased TFH cells correlate with increased local disease. We also show that decreased CD8/Treg ratios correlate with LN metastases. Additional studies are needed to confirm these computational findings in PTCs, evaluate for predictors of tumor recurrence, and expand this analysis to include other thyroid neoplasms. In conclusion, immunologic evaluation of thyroid tumors may provide more information on the biologic behavior of these lesions and inform both diagnosis and treatment. Citation Format: Donna C. Ferguson, William D. Dupont, Thomas Stricker, Vivian L. Weiss. Predicting prognosis in papillary thyroid carcinoma: Clues in the tumor microenvironment [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 5679.