Target capture RNA sequencing as a method of predicting response to neoadjuvant chemotherapy in borderline resectable pancreatic cancer: A multi-institutional study

Abstract

Background: Pancreatic ductal adenocarcinoma (PC) is an aggressively fatal disease. Patients classified as borderline resectable (BR) are often referred for neoadjuvant treatment prior to surgery, however predicting response remains challenging. To date, few studies have examined the transcriptome (gene expression profile) of BR patients or correlated these expression patterns to treatment success. Methods: We conducted a retrospective cohort study examining differential gene expression. Patients with BRPC who underwent surgery were identified from two high volume centers specialized in the management of PC. Patients were stratified into those who received neoadjuvant chemotherapy (nCTx) vs. those who did not. Formalin Fixed Paraffin Embedded tissue (FFPE) was obtained and microdissection was performed. Tissue was then analyzed on the HTG EdgeSeq assay. Differential gene expression analysis was performed on the DESeq2 tool using a negative binomial generalized linear model. Results: A total of 35 patients with BRPC were included in this study. Of the total cohort, 11/35 patients did not receive nCTx, 11/35 patients demonstrated a ≥25% pathological response and 13/35 patients demonstrated a <25% pathological response. A total of 2560 genes across multiple signaling pathways described in various cancers were sampled. Both oncogenes and tumor suppressor genes were analyzed. Log2 transformed differential gene expression pattern at the alpha = 0.05 and 0.01 value was observed in 604 genes and 279 genes respectively when comparing those who received nCTx vs. those who did not. Amongst those who responded to nCTx, 289 genes and 109 genes were differentially expressed with a log2 transformation at the alpha = 0.05 and 0.01 level, respectively, when compared to non-responders. These genes were enriched in 7 different signaling pathways. Statistically significant genes were further filtered by fold expression as seen in the two volcano plots. Conclusion: We have demonstrated a differential gene expression pattern between patients who receive nCTx vs. those patients who did not. Similarly, gene expression also varies between patients who demonstrated a pathological response vs. who those who did not. Future studies to determine whether such differential expression predicts treatment response or is a consequence of treatment is warranted.