Haab Laboratory

Tests to detect and diagnose pancreatic cancer

The successful treatment of pancreatic cancer critically depends on achieving an accurate and early diagnosis, but this can be frustratingly difficult. Conventional methods of evaluating patients—assessing scans, visual inspection of cells from a biopsy, and weighing behavioral, health, and demographic data—do not have the detail necessary to distinguish between benign and malignant disease or between cancers with vastly different behaviors. Sometimes a physician can see a mass or other unusual feature in the pancreas but is unsure what it is. Is it benign or cancerous? And if it is cancer, what is the best course of treatment? 

Our research builds on the concept that molecular-level information will provide details about a condition that are not observable by conventional methods. Molecular biomarkers could provide such information and enable physicians to make accurate diagnoses and develop optimal treatment plans. We are making progress toward this goal for pancreatic cancer. For example, in recent publications in Molecular and Cellular Proteomics and the Journal of Proteome Research, we disclosed carbohydrate-based biomarkers in the blood serum that improve upon the widely used blood test called CA19-9. By using a panel of three or more independent biomarkers, we detected a greater percentage of cancers than we could with any individual biomarker. We are seeking to substantiate those findings and to evaluate their clinical value using serum samples from several clinical sites. 

Other research is aimed at further improving the biomarker tests. The results so far suggest that each individual biomarker arises from a distinct subpopulation of cancer patients and from a characteristic cell type. This finding is important because the biomarkers may reveal differences between subgroups of tumors—a possibility we are exploring in the research described below. For the purpose of improving our blood tests, determining the characteristics of the cells that produce each biomarker, as well as of the cells that do not produce any of our biomarkers, will help to optimize a blood test to accurately identify cancers across the entire spectrum of patients. 

The ultimate goal is to get the new tests established in clinical laboratories in order to benefit patients. To that end, we are working with industry partners to transfer our biomarker assays to the clinical laboratory setting and to begin analyzing patient samples received consecutively from clinical sites. If we have good results, we hope to initiate clinical trials for the diagnosis of pancreatic cancer and, eventually, for evaluations of surveillance among people at elevated risk for pancreatic cancer. 

Better treatment through subtyping
Pancreatic cancer characteristics, such as the cell types within the tumor, the amount of metastasis, the responses to treatments, and overall outcomes, vary greatly among patients. So far, identifying the underlying causes of such differences and predicting the behavior of individual tumors have not been possible. If we could determine what drives the differences between the tumors or identify molecules that help predict the behavior of each tumor, we could establish better treatment plans for each patient or determine the drugs that work best against each subtype. 

Our research is revealing major groupings of tumors based on the carbohydrates on the surface of, and in the secretions from, cancer cells. The carbohydrates are related to the CA19-9 antigen and have distinct biological functions. In current research we want to determine the molecular nature of the subgroups of cells and whether the subgroups have different levels of aggressiveness or different responses to particular drugs. We are using new approaches for measuring carbohydrates and proteins in tumor tissue, and we are employing powerful new software—introduced in our recent publication in Analytical Chemistry—to examine the cell types that produce each carbohydrate-based biomarker. We are using that information to evaluate whether certain types of cells predict clinical behavior. As advances and new options in treatments become available, this type of research is increasingly important for guiding clinical decisions. We are working closely with our physician collaborators to evaluate on a case-by-case basis the value of the molecular information and to guide our research toward improving the tests. Ultimately, physicians could use the molecular tests on material from biopsies, surgical resections, or blood samples.