Gene expression profiling of pancreatic ductal adenocarcinomas in response to neoadjuvant chemotherapy

• Published in: Cancer Medicine
• Main Authors: Sumit Sahni, Christopher Nahm, Mahsa S. Ahadi, Loretta Sioson, Sooin Byeon, Angela Chou, Sarah Maloney, Elizabeth Moon, Nick Pavlakis, Anthony J. Gill, Jaswinder Samra, Anubhav Mittal
• Format: Article
• Language: English
• Published: Wiley 2023-09-01
• Subjects: biomarkers; chemotherapy response; gene expression analysis; neoadjuvant chemotherapy; pancreatic ductal adenocarcinoma
• Online Access: https://doi.org/10.1002/cam4.6411

Abstract Aim Pancreatic ductal adenocarcinoma (PDAC) has the lowest survival rate of all major cancers. Chemotherapy is the mainstay systemic therapy for PDAC, and chemoresistance is a major clinical problem leading to therapeutic failure. This study aimed to identify key differences in gene expression profile in tumors from chemoresponsive and chemoresistant patients. Methods Archived formalin‐fixed paraffin‐embedded tumor tissue samples from patients treated with neoadjuvant chemotherapy were obtained during surgical resection. Specimens were macrodissected and gene expression analysis was performed. Multi‐ and univariate statistical analysis was performed to identify differential gene expression profile of tumors from good (0%–30% residual viable tumor [RVT]) and poor (>30% RVT) chemotherapy‐responders. Results Initially, unsupervised multivariate modeling was performed by principal component analysis, which demonstrated a distinct gene expression profile between good‐ and poor‐chemotherapy responders. There were 396 genes that were significantly (p < 0.05) downregulated (200 genes) or upregulated (196 genes) in tumors from good responders compared to poor responders. Further supervised multivariate analysis of significant genes by partial least square (PLS) demonstrated a highly distinct gene expression profile between good‐ and poor responders. A gene biomarker of panel (IL18, SPA17, CD58, PTTG1, MTBP, ABL1, SFRP1, CHRDL1, IGF1, and CFD) was selected based on PLS model, and univariate regression analysis of individual genes was performed. The identified biomarker panel demonstrated a very high ability to diagnose good‐responding PDAC patients (AUROC: 0.977, sensitivity: 82.4%; specificity: 87.0%). Conclusion A distinct tumor biological profile between PDAC patients who either respond or not respond to chemotherapy was identified.