Given the 38% lower sensitivity of CR imaging systems compared with SFM, programs should assess the continued use of this technology for breast screening.

BackgroundThyroid cancer has been rapidly rising in incidence in Canada; however, in contrast, lung cancer appears to be decreasing in incidence in Canadian men and stable in women. Moreover, disease-related mortality risk is generally very low in TC but high in LC. We performed a geographic spatial analysis in metropolitan Toronto, Canada to determine if there is regional variability of respective risks of thyroid cancer (TC) and lung cancer (LC), among women. Women were of particular interest for this study, given their known predilection for thyroid cancer.MethodsThe postal codes of all females with TC or LC, residing in metropolitan Toronto from 2004 to 2008, were geocoded to point locations according to 2006 Canadian Census data. The data were analysed using a log-Gaussian Cox Process, where the intensity of age-adjusted cancer cases was modelled as a log-linear combination of the population at risk, explanatory variables (race, immigration, and median household income), and a residual spatially varying random effect. For each respective malignancy, statistical models were fit to make quantify the relationship between cancer incidence and explanatory variables.ResultsWe included 2230 women with TC and 2412 with LC. The distribution of TC and LC cases contrasted inversely among Toronto neighbourhoods with the highest TC incidence in the Northeast and the highest LC incidence in the Southeast. A higher proportion of Asian ethnicity was associated with higher regional risk of TC and lower risk of LC. A higher proportion of recent immigrants was associated with increased LC and lower TC risk, whereas median household income and proportions of African ethnicity were not significantly associated with risk of either cancer, after adjustment for other socio-demographic variables.ConclusionsWe observed contrasting regional distributions of female TC and LC cases in Toronto. The differences were partly attributed to ethnic composition variability and the proportion of recent immigrants, but substantial unexplained residual variation of incidence patterns of these malignancies exists, suggesting that more individual-level research is needed to explain the regional variability of incidence of these malignancies.

A large number of factors can affect the statistical power and bias of analyses of data from large cohort studies, including misclassification, correlated data, follow-up time, prevalence of the risk factor of interest, and prevalence of the outcome. This paper presents a method for simulating cohorts where individual's risk is correlated within communities, recruitment is staggered over time, and outcomes are observed after different follow-up periods. Covariates and outcomes are misclassified, and Cox proportional hazards models are fit with a community-level frailty term. The effect on study power of varying effect sizes, prevalences, correlation, and misclassification are explored, as well as varying the proportion of controls in nested case-control studies.

Most studies that have examined the effects of mammographic density and hormone therapy use on breast cancer detection have included screen-film mammography. This study further examines this association in post-menopausal women screened by digital mammography. Approved by the University of Toronto Research Ethics Board, this study identified 688,418 women of age 50-74 years screened with digital or screen-film mammography from 2008 to 2009 within the Ontario Breast Screening Program. Of 2993 eligible women with invasive breast cancer, 2450 were contacted and 1421 participated (847 screen-film mammography, 574 digital direct radiography). Mammographic density was measured by study radiologists using the standard BI-RADS classification system and by a computer-assisted method. Information on hormone therapy use was collected by a telephone-administered questionnaire. Logistic regression and two-tailed tests for significance evaluated associations between factors and detection method by mammography type. Women with >75 % radiologist-measured mammographic density compared to those with <25 % were more likely to be diagnosed with an interval than screen-detected cancer, with the difference being greater for those screened with screen-film (OR = 6.40, 95 % CI 2.30-17.85) than digital mammography (OR = 2.41, 95 % CI 0.67-8.58) and aged 50-64 years screened with screen-film mammography (OR = 10.86, 95 % CI 2.96-39.57). Recent former hormone therapy users were also at an increased risk of having an interval cancer with the association being significant for women screened with digital mammography (OR = 2.08, 95 % CI 1.17-3.71). Breast screening using digital mammography lowers the risk of having an interval cancer for post-menopausal women aged 50-64 with greater mammographic density.