PROJECT SUMMARY/ABSTRACT: Intraductal papillary mucinous neoplasms (IPMN) of the pancreas are cystic tumors that represent a radiographically identifiable precursor lesion of pancreatic cancer. Currently, IPMN represent our best opportunity for intervention - and interception - prior to the development of an incurable cancer. The challenge however, is current laboratory, endoscopic, and imaging technologies are unable to distinguish between IPMN that is at low-risk (low-grade dysplasia) or at high-risk (high-grade dysplasia) of becoming an invasive cancer. In addition, our ability to improve diagnostic accuracy - particularly through biomarker development - has been difficult, as IPMN is very heterogeneous with multiple histologic sub- types and grades of dysplasia existing within the same lesion. Over the past 18 months we have designed and performed pilot studies utilizing a novel spatial transcriptomics platform as a means to spatially identify precise, sub-type specific, epithelial markers of high-grade dysplasia in the tissue from patients resected for IPMN. These studies have found significant differences in gene expression between both low-grade and high-grade dysplasia - and for the first time - have clearly defined gene expression differences between intestinal and pancreaticobiliary sub-type IPMN. The importance of this latter finding cannot be overemphasized, as intestinal sub-type IPMN will typically progress to invasive colloid cancer, which is a much less aggressive malignancy. Furthermore, we have identified differences in cyst fluid protein abundances between intestinal and pancreaticobiliary sub-type IPMN that mirror these gene expression differences. In this project, we will utilize this spatial platform to expand our work and identify sub-type specific epithelial markers of high-grade dysplasia. We will then assess cyst fluid for these proteins using a novel proteomics platform (Olink Focus), and further develop a prediction model for high-risk disease (validated during the previous project period, AUC approximately 0.8). Finally, we will attempt to rapidly validate this model(s) by applying the model(s) to two separate cyst fluid repositories developed by our group that contain cyst fluid from hundreds of patients who have been managed for pancreatic cysts - and have mature follow-up. We believe this approach will allow us to identify very specific cyst fluid protein markers to both epithelial sub-type and grade of dysplasia and develop a very accurate test for high-risk IPMN. Patients with IPMN represent a population in which the identification of accurate markers of high-grade dysplasia will allow for curative intervention (resection) prior to the development of an incurable disease, while sparing those with low-grade lesions the morbidity – and even mortality – of operation.

Project Narrative: Intraductal papillary mucinous neoplasms (IPMN) of the pancreas are radiographically identifiable cystic precursor lesions of pancreatic cancer that evolve from low-grade dysplasia to high-grade dysplasia to carcinoma. Through the use of spatial transcriptomics, we have identified sub-type specific (intestinal vs. pancreaticobiliary) epithelial markers of high-grade dysplasia. In this proposal we utilize this novel technology to precisely identify sub-type specific markers of high- grade dysplasia; to incorporate these cyst fluid markers into a recently developed preoperative prediction model; and to validate this improved prediction model on two large clinical and cyst fluid repositories that have been created by our group.