This study investigates the difference in cancer mortality rates between migrant groups and the native Dutch population, and determines the extent of convergence of cancer mortality rates according to migrants' generation, age at migration and duration of residence. Data were obtained from the national cause of death and population registries in the period 1995-2000. We used Poisson regression to compare the cancer mortality rates of migrants originating from Turkey, Morocco, Surinam, Netherlands Antilles and Aruba to the rates for the native Dutch. All-cancer mortality among all migrant groups combined was significantly lower when compared to that of the native Dutch population (RR 5 0.55, CI: 0.52-0.58). For a large number of cancers, migrants had more than 50% lower risk of death, while elevated risks were found for stomach and liver cancers. Mortality rates for all cancers combined were higher among second generation migrants, among those with younger age at migration, and those with longer duration of residence. This effect was particularly pronounced in lung cancer and colorectal cancer. For most cancers, mortality among second generation migrants remained lower compared to the native Dutch population. Surinamese migrants showed the most consistent pattern of convergence of cancer mortality. The generally low cancer mortality rates among migrants showed some degree of convergence but did not yet reach the levels of the native Dutch population. This convergence implies that current levels of cancer mortality among migrants will gradually increase in future years if no specific preventive measurements are taken. ' 2006 Wiley-Liss, Inc.Key words: cancer; generation; convergence; migrant population; mortality; The Netherlands While molecular epidemiology has identified several examples of genetically determined differences between races, classical epidemiology has shown that the environment and lifestyle predominates in determining cancer incidence.1,2 The role of the environment and behavior is particularly visible in the changing incidence and mortality rates of cancer among migrant populations. Many migrant studies on cancer have shown that the initially different levels of cancer incidence and mortality of migrant groups gradually converge toward the levels of the new host population.3-16 At present, it is still not known how quickly the convergence develops, and how the pace of convergence may differ according to migrant group and type of cancer. This information would better position the role of environmental factors as well as provide knowledge for more rational planning of specific preventive and curative health services for migrant populations.About 10% of the population of the Netherlands is currently of nonwestern foreign origin. 17 The largest migrant groups originate from Turkey, Morocco, and the former Dutch colonies in South America and the Caribbean (Surinam and Netherlands Antilles/ Aruba). Turkish and Moroccans are mostly labor migrants followed by their immediate family and descendants for ...

No abstract

Background Deep learning-based delineation of organs-at-risk for radiotherapy purposes has been investigated to reduce the time-intensiveness and inter-/intra-observer variability associated with manual delineation. We systematically evaluated ways to improve the performance and reliability of deep learning for organ-at-risk segmentation, with the salivary glands as the paradigm. Improving deep learning performance is clinically relevant with applications ranging from the initial contouring process, to on-line adaptive radiotherapy. Methods Various experiments were designed: increasing the amount of training data (1) with original images, (2) with traditional data augmentation and (3) with domain-specific data augmentation; (4) the influence of data quality was tested by comparing training/testing on clinical versus curated contours, (5) the effect of using several custom cost functions was explored, and (6) patient-specific Hounsfield unit windowing was applied during inference; lastly, (7) the effect of model ensembles was analyzed. Model performance was measured with geometric parameters and model reliability with those parameters’ variance. Results A positive effect was observed from increasing the (1) training set size, (2/3) data augmentation, (6) patient-specific Hounsfield unit windowing and (7) model ensembles. The effects of the strategies on performance diminished when the base model performance was already ‘high’. The effect of combining all beneficial strategies was an increase in average Sørensen–Dice coefficient of about 4% and 3% and a decrease in standard deviation of about 1% and 1% for the submandibular and parotid gland, respectively. Conclusions A subset of the strategies that were investigated provided a positive effect on model performance and reliability. The clinical impact of such strategies would be an expected reduction in post-segmentation editing, which facilitates the adoption of deep learning for autonomous automated salivary gland segmentation.