4L-2 EXPLORING THE EFFECTS OF FACTORS ASSOCIATED WITH THE OUTCOME OF PANCREATIC CANCER SCREENING IN HIGH-RISK INDIVIDUALS BY THE USE OF A MICROSIMULATION MODEL

Tuesday, October 21, 2014: 3:45 PM

Inge M.C.M. de Kok, PhD1, Femme Harinck, MD2, Ingrid C.A.W. Konings, MD2, Iris Lansdorp-Vogelaar, PhD3, Paul Fockens, MD, PhD4, Marco J. Bruno, MD, PhD2, Marjolein van Ballegooijen, MD, PhD1 and Sonja Kroep, MSc5, (1)Department of Public Health, Erasmus MC, University Medical Center Rotterdam, Rotterdam, Netherlands, (2)Department of Gastroenterology & Hepatology, Erasmus MC, University Medical Center Rotterdam, Rotterdam, Netherlands, (3)Erasmus MC, University Medical Center, Rotterdam, Netherlands, (4)Department of Gastroenterology and Hepatology, Academic Medical Center, University of Amsterdam, Amsterdam, Netherlands, (5)Erasmus MC, University Medical Center, Department of Public Health, Rotterdam, Netherlands

Purpose: To explore the uncertainties of early detection of pancreatic cancer in high-risk individuals and consequently highlight the areas to which further research should be directed.

Method: Effects of pancreatic cancer screening were estimated using the Microsimulation Screening Analysis (MISCAN)-model. The majority of the model assumptions were based on preliminary data from the screening trials, the recommendations as stated in the consensus paper of the international Cancer of the Pancreas Screening (CAPS)-consortium, and  expert opinions. We varied dwell times of the different health states and test-characteristics of the screening test. We considered different screening ages and intervals and varied follow-up strategies after a positive screening.

Result: Mortality reduction (MR) was 35% and 58% (436 and 722 cases per 10,000 persons) for 5-yearly and annual screening at ages 50 to 75, respectively. For 5 yearly screening, the number needed to screen (NNS) was 117.1 and number needed to treat (NNT) was 2.5 to prevent one cancer death (see Figure (the order (from left to right) of the variations in the factors (between brackets) corresponds to the order (from left to right) in the figure. The dashed lines represent the results of the base case analyses)). The NNT was lowest in case all screen positives with preinvasive stage 3 or cancer are treated (2.4, MR 32%). If only persons are treated who are already in an invasive stage of disease, the NNT was 5.3 (MR 10%). Results were sensitive for pancreatic cancer risk (risk doubled: NNS 64.3, NNT 2.7, MR 38%) and duration of the preclinical stage of the disease (increased to 30 years: NNS 92.6, NNT 3.2, MR 46%). Results were less sensitive for test characteristics.

Conclusion: Modeling shows that there is potential for pancreatic screening to be cost-effective in high-risk individuals. Follow-up strategy of screen positives and duration of the preclinical stage have the highest impact on the outcome of pancreatic cancer screening, as is inclusion of patient populations that are exposed to a certain risk to develop pancreatic cancer.

See more of: COMPARATIVE EFFECTIVENESS RESEARCH

See more of: The 36th Annual Meeting of the Society for Medical Decision Making