Abstract B065: Biomarker approach to define tumor subtype in pancreatic ductal adenocarcinoma
Carson Ho,Michael Horowitz,Lyanne Oblein,Shayan Sarkar,Soma Kobayashi,Karen Bai,Natalia Marchenko,Luisa Escobar-Hoyos,Kenneth Shroyer
DOI: https://doi.org/10.1158/1538-7445.panca2023-b065
IF: 11.2
2024-01-18
Cancer Research
Abstract:Background: Pancreatic ductal adenocarcinoma (PDAC) is one of the most lethal tumors with a predicted five-year survival rate in 2023 of 12%. Building on prior observations that molecular subtypes correlate with survival and response to treatment, our goal has been to understand the patterns of expression and prognostic correlations of three reciprocally expressed proteins previously identified as surrogate biomarkers of PDAC subtypes. Keratin 17 (K17) and Keratin 5 (K5) have both been identified as biomarkers of the basal subtype whereas GATA6, a transcription factor necessary for normal pancreatic development, is a surrogate marker of the classical subtype associated with longer-term survival. Here, we explore the expression of these three diagnostic biomarkers in PDAC and their specificity to tumor cells. Methods: Immunohistochemical analyses of K17 (KDx Diagnostics), GATA6 (R&D Systems), and KT5 (Leica) were performed on archival primary PDAC specimens from Stony Brook University Hospital (n= 30), by routine immunoperoxidase methods on serial sections. IHC stained slides were scanned using an Olympus VS200 slide scanner and images were imported into QuPath for digital analysis. Regions of interest (ROIs) were defined by the presence of tumor cells and were identified by histological assessment of representative sections. Digital scoring was performed on identified ROIs using the integrated "positive cell detection" function in QuPath and classified into three different stain intensities (1+, 2+, 3+). The threshold for each intensity is as follows: 1+, 0.2; 2+, 0.4; 3+, 1.0. According to the number of detections of each intensity, tumor cells were categorized into negative (0 and 1+) or positive (2+ and 3+). The percentage positive of analyzed biomarker expression within the tumor ROI was calculated by dividing the total number of positive cells by the total number of detected cells. Results: K17 was detected in both well and poorly differentiated PDAC components and in a few benign proliferative ducts of chronic pancreatitis but no staining was detected in stromal cells, benign ductal epithelium, or acinar cells. K5 was detected in a smaller proportion of cells from glandular components of the tumor as well as in most diffusely infiltrative single cells, benign ductal, and acinar cells but not in stromal cells. By contrast, GATA6 was detected in the nuclei of most cancer epithelial cells and was also detected in benign ductal cells, acinar cells, and stromal cells. The proportion of K17 expressing cells exceeded that of K5 positive cells in the great majority of ROIs and across cases. Conclusions: Keratin 17 detects a higher proportion of tumor cells than K5 and shows the lowest level of expression in benign cell types. By contrast, GATA6 IHC shows variable patterns of expression related to K17 or K5 expression and is expressed in the highest proportion of tumor cells, including both well-differentiated, diffusely infiltrative components, and benign tissue components. Citation Format: Carson Ho, Michael Horowitz, Lyanne Oblein, Shayan Sarkar, Soma Kobayashi, Karen Bai, Natalia Marchenko, Luisa Escobar-Hoyos, Kenneth Shroyer. Biomarker approach to define tumor subtype in pancreatic ductal adenocarcinoma [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) nr B065.
oncology