By Lisa Newkirk
Pancreatic cancer is one of the deadliest forms of cancer, but patients who are diagnosed at an early stage can survive five years or longer.
A team at Mayo Clinic is developing patient-centered solutions for early detection of pancreatic cancer using artificial intelligence and biomarkers.
This year, over 57,000 Americans will be diagnosed with pancreatic cancer, and 47,000 Americans will die of the disease, according to American Cancer Society estimates. More than 50% of patients with pancreatic cancer are diagnosed at a late stage, when the disease has spread beyond the pancreas.
Pancreatic cancer is challenging to catch early because most tumors don’t produce symptoms and elude diagnosis. There are no tests or tools for low-cost, convenient, safe and accurate screening.
Mayo Clinic researchers are creating an early-detection strategy to change that and transform patient care.
“As Mayo Clinic physicians and scientists it is our responsibility to purposefully innovate with a focus on improving outcomes for our patients, a shared responsibility to create the best-in-class care models for the future generation,” says Shounak Majumder, M.D., a Mayo Clinic gastroenterologist and pancreatologist.
Dr. Majumder and Gloria Petersen, Ph.D., a Mayo Clinic scientist and the Purvis and Roberta Tabor Professor, are leading the research team that is developing the early-detection strategy. Using multidisciplinary team science and artificial intelligence tools, the team will build state-of-the-art algorithms to systematically identify people at high risk for pancreatic cancer, and discover and validate novel imaging and molecular biomarkers for accurate early detection.
Systematic identification for high-risk patients isn’t available yet as a standard of care in clinical practice, but Dr. Petersen has shown that people with a family history of pancreatic cancer and those who are carriers of a predisposing gene mutation have higher risk, as well as people who develop new onset of diabetes mellitus later in life.
Dr. Petersen and Dr. Majumder’s team will develop algorithms that make it possible to systematically identify people at high risk, using information from electronic health records. The team includes artificial intelligence researchers in the Department of Health Sciences Research. They will leverage decades of well-characterized clinical data to develop the algorithms and build a comprehensive approach.
The research team will also develop a machine learning algorithm for precise, image-based identification of early pancreatic cancer and advanced precancerous lesions. They will collaborate with Mayo Clinic radiologists who have expertise in artificial intelligence and pancreatic imaging. Mayo has begun this work with an artificial intelligence tool that can quickly scan CT images and accurately outline the pancreas — requiring minimal assistance from humans. The next step is to develop deep learning algorithms using images from patients with pancreatic cancer and people who have risk factors for pancreatic cancer. The goal is to reliably detect early-stage pancreatic cancer and identify changes that are harbingers of the disease.
The early detection strategy calls for combining imaging techniques with blood tests and other noninvasive, cost effective tools. These approaches require discovery and validation of molecular biomarkers. Dr. Majumder has shown that DNA markers in tissue and pancreatic juice can accurately detect pancreatic cancer and advanced precancerous lesions. Studies are ongoing. The research team is also in the early stages of developing a blood-based molecular test.
Dr. Majumder has also initiated a high-risk pancreas clinic for patients on Mayo Clinic’s Rochester campus that uses an innovative clinical model. Dr. Majumder and his team collaborate with primary care physicians, genetic counselors and other practitioners to identify patients at high risk; develop accurate, cost-effective and practical approaches to screening and surveillance; and treat advanced precancerous lesions and early-stage disease. The clinic will serve as a translational research hub for early detection.
The Centene Charitable Foundation is supporting the research team’s work to create the early-detection strategy, which Drs. Majumder and Petersen seek to bring to the clinic within five years. (Read related news release.)
Mayo Clinic cares for nearly 1,100 patients with pancreatic cancer each year. Mayo is active in several areas of research, leads national initiatives such as the Mayo Clinic Pancreatic Cancer Specialized Program of Research Excellence and the Pancreatic Cancer Genetic Epidemiology Consortium, and has created one of the largest research registries in the world of pancreatic cancer patients and families.
“We so appreciate the many Mayo patients whose data and biospecimens have already helped us to make major advances and now enable our research team to take our work to the next level with this support,” says Dr. Petersen.
Tags: About, artificial intelligence, biomarkers, cancer genomics, cancer screening, collaboration, CT, DNA, Gloria Petersen, Innovations, Mayo Clinic Cancer Center, News, pancreatic cancer, patient experience, Shounak Majumder