From 2000 to 2002, male patients at a Canadian cancer treatment center with new-incident head-and-neck or esophageal cancers were invited to participate in a population-based study. The study population included 87 cases and 172 controls. A lifetime-history questionnaire was administered. Odds ratios (ORs) were calculated for occupational groups with a minimum of five cases, adjusted for duration of employment, age, smoking, alcohol, education, and income. A significantly increased risk was shown for construction workers (OR = 2.20; 95% CI 1.25-3.91). This investigation of a set of rare cancers over a limited time period demonstrates the feasibility of this research approach. The increased risk among construction workers supports the need for more comprehensive study of exposures in this occupational group.