Cyst Fluid Biosignature to Predict Intraductal Papillary Mucinous Neoplasms of the Pancreas with High Malignant Potential

Abstract

Background: Current standard-of-care technologies, such as imaging and cyst fluid analysis, are unable to consistently distinguish intraductal papillary mucinous neoplasms (IPMNs) of the pancreas at high risk of pancreatic cancer from low-risk IPMNs. The objective was to create a single-platform assay to identify IPMNs that are at high risk for malignant progression.

Study design: Building on the Verona International Consensus Conference branch duct IPMN biomarker review, additional protein, cytokine, mucin, DNA, and microRNA cyst fluid targets were identified for creation of a quantitative polymerase chain reaction-based assay. This included messenger RNA markers: ERBB2, GNAS, interleukin 1β, KRAS, MUCs1, 2, 4, 5AC, 7, prostaglandin E2R, PTGER2, prostaglandin E synthase 2, prostaglandin E synthase 1, TP63; microRNA targets: miRs 101, 106b, 10a, 142, 155, 17, 18a, 21, 217, 24, 30a, 342, 532, 92a, and 99b; and GNAS and KRAS mutational analysis. A multi-institutional international collaborative contributed IPMN cyst fluid samples to validate this platform. Cyst fluid gene expression levels were normalized, z-transformed, and used in classification and regression analysis by a support vector machine training algorithm.

Results: From cyst fluids of 59 IPMN patients, principal component analysis confirmed no institutional bias/clustering. Lasso (least absolute shrinkage and selection operator)-penalized logistic regression with binary classification and 5-fold cross-validation used area under the curve as the evaluation criterion to create the optimal signature to discriminate IPMNs as low risk (low/moderate dysplasia) or high risk (high-grade dysplasia/invasive cancer). The most predictive signature was achieved with interleukin 1β, MUC4, and prostaglandin E synthase 2 to accurately discriminate high-risk cysts from low-risk cysts with an area under the curve of up to 0.86 (p = 0.002).

Conclusions: We have identified a single-platform polymerase chain reaction-based assay of cyst fluid to accurately predict IPMNs with high malignant potential for additional studies.

Figure 1.

Workflow of cyst fluid preparation for input into the bioinformatics model to predict level of cyst dysplasia.

Figure 2.

Removal of confounders. A Pearson correlation matrix was constructed between each pair of gene markers. GNAS, miR106B, miR155, miR24, miR92A, and miR532 were removed from the analysis due to high correlation of gene expression that confounded machine learning algorithms in identifying predictive markers for the biosignature.

Figure 3.

Intraductal papillary mucinous neoplasm (IPMN) cyst fluid biosignature can differentiate high from low-risk cysts. Lasso penalized logistic regression with cross validation identified a 3-gene cyst fluid signature with optimal accuracy to predict the risk of pancreatic malignancy in IPMN. In this model, low-risk (low and moderate grade dysplasia) vs high-risk (high-grade dysplasia and invasive cancer) cysts were predicted with an accuracy, as measured by AUC, of 86%; y=0.36+ (−0.06IL1β) + (−0.17MUC4) + (−0.50PTGES2); AUC=0.86, p-value=0.002.