Advancing the Clinical Translation of Cyst Fluid Assays for Early Detection of Pancreatic Cancer

PROJECT SUMMARY Pancreatic cystic lesions (PCLs) represent an opportunity for early detection of pancreatic adenocarcinoma. The accurate identification of cysts with high grade dysplasia or invasive adenocarcinoma, together referred to as “Advanced Neoplasia” (AN), that warrant surgical intervention represents a critical unmet need in the management of PCLs. Most pancreatic cysts with the potential to develop AN are mucinous; in contrast, non- mucinous PCLs have little or no malignant potential. Biochemical and cytological analysis of aspirated cyst fluid are important tools in the diagnosis and risk stratification of PCLs. However, sensitivity of the only clinical biomarker for mucinous cysts, carcinoembryonic antigen (CEA), is insufficient to allow clinicians to confidently remove patients from surgical consideration. Moreover, the most commonly performed CEA assay requires 500 L of fluid, which is often unavailable. For over 50% of patients that undergo invasive trans-gastric cyst fluid aspiration, inadequate biospecimen is obtained to run the standard of care biochemical tests. In an effort to improve the sensitivity and applicability of cyst fluid analysis, we used our novel multiplex mass spectrometry technology to identify the protease gastricsin which accurately identifies mucinous cysts with an AUC of 0.98 and requires only 5 L fluid. Despite reliance by clinicians on these analyses to guide clinical decision-making, little effort has been directed toward optimization of biospecimen processing for pancreatic cyst fluid. There are no standardized pathways for cyst fluid processing and, unlike other biospecimens such as serum, pancreatic cyst fluid has variable viscosity and contamination with blood and proteinaceous material that could interfere with assay reproducibility. It is unknown if the variability we observe clinically in cyst fluid CEA and cytology reflects true biological differences or inconsistent preanalytical biospecimen processing. The overall objective of this proposal is to improve completeness, reproducibility, and accuracy in pancreatic cyst fluid diagnostic evaluation in order to improve the early diagnosis of pancreatic cancer while avoiding the burdens of overdiagnosis and overtreatment. To achieve our objective, we will systematically evaluate the impact of preanalytical variables on cyst fluid biochemical and cytological analysis (Aim 1) and identify strategies to mitigate the small volumes of available cyst fluid (Aim 2) in order to develop a streamlined, reproducible protocol (Aim 3) that improves the reliable early detection of pancreatic cancer. We will then validate the performance of our streamlined protocol using prospectively - collected clinical samples, and we will evaluate inter-assay variability by implementing the protocols at two independent sites. By improving the reliability of our assays, clinicians will be able to direct surgical intervention appropriately to patients with incipient pancreatic cancer while sparing others unnecessary risks of mortality, substantial morbidity, and health care costs.