My name is Anil Rustgi, and I would like to provide an overview of two things. The first is the exciting research that is ongoing between new-onset diabetes mellitus and pancreatic cancer. Secondly, as an illustration of that, a clinical prediction model in which I am a coauthor, the primary work being done by Dr. Ben Boursi and Dr. Yu-Xiao Yang, both colleagues here at the University of Pennsylvania in collaboration with colleagues at MD Anderson Cancer Center.
Every year in the United States there are about 53,000 or 54,000 new cases of pancreatic cancer. Unfortunately, there are about 44,000 or 45,000 deaths related to either local complications from the disease or, more frequently, when the disease has spread to distant organs such as liver and lung. As of now, the median survival is about 6 months, and the 5-year survival is improving through intensive efforts in clinical trials to 7%, where it had been 5%. Unfortunately, it is estimated that by the year 2020, pancreatic cancer will be the second leading cause of cancer-related mortality. Currently, it is third. The reasons for the poor outcome with pancreatic cancer are likely two-fold. One, the vast majority of patients present with metastatic disease. Secondly, the disease is intrinsically quite resistant to chemotherapy for a variety of reasons. That being said, significant advances have been made in the field. First, experimental animal models are used iteratively and applied translationally for new approaches in therapeutics. Secondly, an appreciation of approaches to new therapies in pancreatic cancer, especially in immunotherapy, and thirdly, an ongoing emphasis on those individuals and families that are at increased risk for pancreatic cancer. It is estimated that about 5% to 10% of cases every year have a hereditary or familial basis. There is a cohort of individuals that are at higher risk based upon genetic predisposition. There is also an appreciation that there may be a cohort of patients in the general population at increased risk due to new-onset diabetes mellitus, which really is to be distinguished from type 1, which is insulin-dependent diabetes mellitus due to an autoimmune etiology, typically very early onset in life. We can see it in childhood and adolescence, and then, there is a long-term sequela from that. Secondly, type 2 diabetes mellitus which may not be insulin dependent and may evolve to be an insulin dependent, and these are typically adults. What I am going to refer to as type 3C, which is relatively new-onset diabetes mellitus and believed to be possibly linked to the diagnosis of pancreatic cancer, typically that diagnosis is within a range or window of 6 months to 3 years of the diagnosis of pancreatic cancer. From the viewpoint of those individuals who are genetically predisposed, we often are able to identify such individuals and families based upon family history, pursue genetic testing and counseling, and modify clinical screening as needed, especially under clinical studies, for instance, in which I am involved, called CAPS5. With the potential relationship of new-onset diabetes mellitus in pancreatic cancer, we are not at the stage of screening the general population, but that is the desire to try to come up with a risk model based upon clinical parameters to target pancreatic cancer screening with new-onset diabetes.
This study, led by Dr. Ben Boursi and Dr. Yu-Xiao Yang, my colleagues, in which I am a collaborator and a coauthor, conducted a retrospective cohort study in a population-represented database from the United Kingdom. Now, the advantage of using this database is that for the last 30 years or so, millions and millions of patients have been followed longitudinally, demographically, clinically, diagnostically, and therapeutically in the United Kingdom. One is able to access these databases and distill from that parameters that one may be interested in. From that, individuals with incident diabetes after the age of 35 and greater than or equal to 3 years of followup after the diagnosis of diabetes, were eligible for inclusion. Predictors consisted of epidemiological and clinical characteristics that were available at the time of diagnosis of diabetes. Biostatistical analyses were conducted in great detail. The study included almost 110,000 patients with new-onset diabetes. Among them, 390 or 0.4% were diagnosed with pancreatic cancer within 3 years. A model emerged from that which included age, body mass index change, smoking, hemoglobin A1c - a parameter for diabetes, cholesterol, hemoglobin, creatinine, alkaline phosphatase, and the use of certain medications (proton pump inhibitors and antidiabetic medications). Then, a specific type of modeling was done called bootstrapping, and if the predicted risk threshold for definitive pancreatic cancer screening was set at 1% over 3 years, only a little over 6% of the new-onset diabetes population would undergo definitive screening, and sensitivity specificity of positive predictive values were calculated. From this comprehensive study, a risk model based on widely available clinical parameters can help in targeting pancreatic cancer screening in patients with new-onset diabetes. The hope is that this can be done in other databases in other countries first, and secondly, that a prospective study could be done as well. Very exciting research, very exciting results and the hope is that at some point in the future that additional studies may allow for risk model development, as is the case in certain other disease states for pancreatic cancer in patients with new-onset diabetes mellitus, so called type 3C diabetes mellitus. Thank you.