264136 Diet and lifestyle factors and risk of subtypes of esophageal and gastric cancer: Classification tree analysis

Tuesday, October 30, 2012 : 12:45 PM - 1:00 PM

Stephanie Silvera, PhD, CPH , Department of Health and Nutrition Sciences, Montclair State University, Montclair, NJ
Susan Mayne, PhD , Chronic Disease Epidemiology, Yale School of Public Health, New Haven, CT
Harvey Risch, MD, PhD , Chronic Disease Epidemiology, Yale School of Public Health, New Haven, CT
Marilie Gammon, MSPH, PhD , Epidemiology, University of North Carolina Gillings School of Global Public Health, Chapel Hill, NC
Thomas Vaughan, MD , Program in Epidemiology, Fred Hutchinson Cancer Research Center, Seattle, WA
Wong-Ho Chow, PhD , Cancer Epidemiology & Genetics, National Cancer Institute, Bethesda, MD
Joel Dubin, PhD , Department of Statistics & Actuarial Science, and School of Public Health & Health Systems, University of Waterloo, Waterloo, ON, Canada
Robert Dubrow, MD, PhD , Chronic Disease Epidemiology, Yale School of Public Health, New Haven, CT
Janet Stanford, PhD , Program in Epidemiology, Fred Hutchinson Cancer Research Center and University of Washington, School of Public Health and Commu, Seattle, WA
A. Brian West, MD , Pathology, Yale - New Haven Hospital, New Haven, CT
Heidrun Rotterdam , Department of Pathology, Columbia University, New York, NY
William Blot, PhD , Cancer prevenetion, control and population-based research, Vanderbilt-Ingram Cancer Center, Rockville, MD
While a number of risk factors for squamous cell carcinoma of the esophagus (SCC) and adenocarcinomas of the esophagus (EA), gastric cardia (GC) and other (non-cardia) gastric sites (OG) have been identified, little is known about how these factors interact. Our objective is to examine the interaction of diet, other lifestyle, and medical factors and risk of subtypes of esophageal and gastric cancer. We used classification tree analysis to analyze data from a multi-center, population-based case-control study (including 1095 cases and 687 controls) conducted in Connecticut, New Jersey, and western Washington state. We found that frequency of gastroesophageal reflux (GERD) symptoms was the most important risk factor for EA, associated with an increased risk, and that several dietary factors (red meat, non-citrus fruits, dark green vegetables, and raw vegetables) appeared to modify that risk. GERD also emerged as a key risk factor for both GC and OG; several dietary factors further impacted risk of OG. Cigarette smoking, low income, and African American race were important risk factors for SCC, with an additional contribution of diet (non-citrus fruit intake). Our results suggest that different combinations of risk factors interact to determine risk of each of these cancers, with diet playing at least some role in all 4 cancer sites. The findings suggest that classification tree analysis may be useful in partitioning risk and that mapping the complex interactions between risk variables therefore can be important both in future research and in targeted prevention interventions.

Learning Areas:
Biostatistics, economics
Chronic disease management and prevention
Epidemiology
Public health or related research

Learning Objectives:
By the end of the presentation, the learner will be able to: 1. Identify how classification tree analysis may be useful in partitioning risk in epidemiological studies. 2. Discuss how mapping complex interactions between risk variables can be used to guide future research and targeted prevention interventions.

Keywords: Cancer, Epidemiology

Presenting author's disclosure statement:

Qualified on the content I am responsible for because: I have advanced degrees in both nutritional sciences and epidemiology with a focus on cancer outcomes. I have published over 20 manuscripts in the area of cancer epidemiology, including the use of advanced statistical methods for the analysis of population-level data.
Any relevant financial relationships? No

I agree to comply with the American Public Health Association Conflict of Interest and Commercial Support Guidelines, and to disclose to the participants any off-label or experimental uses of a commercial product or service discussed in my presentation.

Back to: 4213.0: Cancer Epidemiology 1