Cancer of the pancreas, mostly adenocarcinoma of the exocrine pancreas, is one of the most rapidly fatal of the cancers. This cancer is the fifth leading cause of all cancer deaths, totaling an estimated 27,700 new cases and 25,000 deaths in 1992. No safe and effective early screening methods are available. Several environmental determinants and molecular genetic pathogenetic mechanisms, as well as familial inheritance have been suggested. However, no clear, single, unifying etiologic hypothesis has been established. The study proposed will elucidate the etiology of adenocarcinoma of the exocrine pancreas (""""""""pancreatic cancer"""""""") through a rigorous application of established epidemiologic methods to data on approximately 100 families ascertained through individuals diagnosed with pancreatic cancer. The families are part of the American Cancer Society funded initiative """"""""The Ecogenetics of Pancreatic Cancer: A National Registry of Families."""""""" We will use three approaches to investigate the genetic epidemiology of pancreatic cancer. First, we will apply logistic regression models to examine the effects of observable covariates on risk for pancreatic cancer. To account for the natural correlation of the observations within families, quasi-likelihood methods will be used to incorporate an overall correlation in risk among family members; an approach usually termed Generalized Estimating Equations or GEE1. Second, we will assess the risk in these families to various phenotypic traits, such as pancreatic cancer or cancers generally, which is in excess of that expected for a similar group of individuals drawn at random from the general population; a test of familial aggregation. Third, we will use complex segregation analyses to identify patterns of inheritance of pancreatic cancer in these families. We will also examine the evidence for genetic heterogeneity. As part of this third analytic method, we will use computer simulations to establish an appropriate correction for the method in which these families were ascertained. The results of this analysis will be used in the complex segregation analyses. In brief, this is an application to the Small Grants Program for Epidemiology for funding to analyze existing data and to resolve a methodological problem.
