Overview

Abstract (English)

Rare cancers affect few people individually, but collectively account for approximately 22% of new cancer cases in Canada. However, prevention and population-level research on rare cancers have been limited as low case numbers make them difficult to study. The emergence of large, collaborative cohorts may offer the opportunity to study rare cancers with a sufficiently large sample size. The nested case-control design allows all cases to be utilized while maintaining the temporal advantage of a cohort design. This thesis evaluated the feasibility of large cohort studies as a tool for rare cancer research using data from Alberta’s Tomorrow Project (ATP). Two questions were addressed: 1) Are self-reported diagnoses of rare cancer valid to use as an outcome in research when cancer registry linkage is unavailable? and 2) Is a nested case-control design a feasible option to study rare cancers in large cohort studies? The validity of self-reported cancer diagnoses was explored through ATP linkage to the Alberta Cancer Registry (ACR). The first instance of self-reported cancer was compared to the first cancer diagnosis in the ACR after enrollment. Sensitivity and positive predictive value (PPV) were estimated for the reporting of overall cancer status, the reporting of common or rare cancer, and the reporting of site-specific cancer. Logistic regression analysis explored factors associated with false positive, false negative, and incorrect site reporting. Overall, rare cancers had a lower sensitivity and PPV than common cancers. Participants with a rare cancer were more likely to report an incorrect site than those with a common cancer. Rare cancers were also less likely to be captured by active follow-up than common cancers. Therefore, registry linkage is necessary to capture rare cancer diagnoses completely and accurately in large cohort studies. A pilot etiologic study on pancreatic cancer assessed the feasibility of the nested case-control design for rare cancer research in the ATP cohort. Incidence density sampling was used to match controls to cases on follow-up time and other factors. Conditional logistic regression was used to investigate the association of pancreatic cancer with well-established risk factors and with less established dietary risk factors. The analysis was adequately powered to find estimated effects of established risk factors that were consistent with other literature. However, the dietary risk factor analysis was not adequately powered to detect low to moderate effects. Attrition limited the eligible pool of controls and introduced the possibility of healthy volunteer bias. Using a larger, national-scale cohort that would produce more cases and linking to vital statistics for passive follow-up of controls can mitigate these issues. This analysis found that rare cancer research may be feasible in large cohort studies at a national scale if linkage to cancer registry and vital statistics is available. Removing barriers that currently prevent the sharing of linked data cross-provincially would allow for these opportunities in rare cancer research to be explored.

Abstract (French)

Please note that abstracts only appear in the language of the publication and might not have a translation.