Abstract B110: Molecular subtypes from endoscopy ultrasound-guided biopsy samples collected at initial diagnosis predict therapeutic response to neoadjuvant chemotherapy in pancreatic cancer

Abstract

Abstract Introduction Pancreatic cancer has a poor prognosis, despite the availability of various therapeutic approaches. Recent evidence suggests that precision strategies tailored to the molecular profiles of patients’ tumors can improve oncologic outcomes in solid cancers. This study aimed to identify clinically relevant molecular subtypes based on transcriptomic data of samples collected through ultrasound-guided fine needle aspiration biopsies (EUS-FNA) performed before neoadjuvant chemotherapy (NACT) in patients with borderline resectable pancreatic cancer (BRPC) or locally advanced pancreatic cancer (LAPC). Methods Total RNA sequencing was performed on tumor samples collected using EUS-FNA prior to NACT. The response to NACT as well as the clinicopathologic results of patients were recorded. To determine molecular subtypes based on transcriptomic data, an unbiased approach based on non-negative matrix factorization (NMF) was utilized. Gene set enrichment analysis (GSEA) and ingenuity pathway analysis (IPA) were performed to describe the molecular characteristics of subgroups. Clinicopathologic and genetic characteristics of subgroups were validated using a published cohort database, the cancer genome atlas for pancreatic cancer (TCGA-PAAD). In addition, spatial transcriptomic profiling was conducted to investigate the interactive molecular characteristics the tumor microenvironments specific to each subtype. Results Sequencing data was obtained from samples collected from 35 patients. Two molecular subtypes were identified. GSEA showed that subgroup 2 enhanced expression of gene sets pertaining to G2M checkpoint, E2F targets, KRAS signaling, and mitotic spindle, all of which are related to cancer-specific proliferation. IPA results revealed the subtype-specific upregulation of WNT11 and ASCL1, which are implicated in the aggressive behavior of pancreatic cancer in this subtype. Notably, patients in subtype 2, characterized by aggressive cancer behavior, exhibited poorer outcomes following neoadjuvant treatment than those in subtype 1. Similar results were found in the TCGA-PAAD cohort. Using spatial transcriptomic analysis, we revealed that a set of genes representing subtype 2 were highly expressed in a tumor area of a patient belonging to subtype 2, but showed low expression in a patient belonging to subtype 1. Conclusion Two molecular subtypes defined based on transcriptomic data from EUS-FNA samples obtained during the initial diagnosis of pancreatic cancer possessed distinct molecular characteristics and were related to different clinical outcomes, especially regarding the therapeutic efficacy of neoadjuvant treatment. These results demonstrate that, following robust validation with corroborative evidence, precision strategies based on tumor biology should be provided in a clinical setting to improve treatment outcomes. Citation Format: Woohyung Lee. Molecular subtypes from endoscopy ultrasound-guided biopsy samples collected at initial diagnosis predict therapeutic response to neoadjuvant chemotherapy in pancreatic cancer [abstract]. In: Proceedings of the AACR Special Conference in Cancer Research: Pancreatic Cancer; 2023 Sep 27-30; Boston, Massachusetts. Philadelphia (PA): AACR; Cancer Res 2024;84(2 Suppl):Abstract nr B110.