Tong Wang

Introduction: China’s digestive cancer landscape presents a dual challenge, the significant geographic disparity and the distorting effect of large-scale internal migration on burden estimation. The household-registered population (HRP)-based surveillance system does not fully capture the impact of large-scale internal migration, and county-level spatial evidence that accounts for migration remains limited. This study aimed to characterize the spatial distribution of the migration-adjusted burden of five digestive cancers in China, assess spatial autocorrelation, and identify county-level high-high and low-low clusters. Methods: This spatial epidemiological study used previously developed migration-adjusted county-level incidence and mortality estimates for five digestive cancers in mainland China, derived using Bayesian INLA-SPDE modeling based on quality-controlled data from 487 cancer registries covering 28.6% of the national population. The spatial distribution was visualized using thematic mapping, and Global and Local Moran’s I analyses were conducted to assess spatial autocorrelation and identify significant spatial clusters. Results: Spatial description revealed geographic patterns in cancer incidence and mortality, closely corresponded to clusters identified by spatial autocorrelation analysis. Global Moran’s I analysis confirmed significant positive spatial autocorrelation in the migration-adjusted burden of all five cancers nationwide ( p < 0.001). Distinct geographic patterns of HH clusters were evident across cancers: liver cancer in the southeastern coast, colorectal cancer in the eastern coast and northeast, gastric cancer in the east and northwest, esophageal cancer in central-eastern regions, and pancreatic cancer in the northeast and eastern coast. LL clusters were mainly in western, central-western, and southwestern China. Conclusion: This study identified significant spatial clustering in the migration-adjusted burden of the five major digestive cancers in China. High-burden areas were mainly concentrated in eastern coastal, northeastern, and parts of central-eastern China, whereas lower-burden areas were more common in western and parts of southwestern China. These findings highlight the importance of incorporating population mobility into digestive cancer burden assessment and provide evidence to support region-specific cancer prevention, control prioritization, and more dynamic surveillance in China.

Background: China’s internal migration poses major challenges to cancer surveillance systems that rely on household-registered populations (HRP). Excluding migrants can lead to biased incidence estimates and misinformed public health planning. Methods: This study estimated the prostate cancer incidence among the resident population (RP) using a Bayesian integrated nested Laplace approximation with stochastic partial differential equation (INLA-SPDE) model, incorporating inter-provincial migrant weights, and explored spatial clustering. Results: The findings revealed a substantial interprovincial migrant population of 73,459,708 individuals, based on data from the 2016 China Migrants Dynamic Survey conducted by the Migrant Population Service Center, National Health Commission of China. With Shanghai and Beijing showing relatively high difference proportions of 40.7 and 37.9%, respectively. Nationally, the differences in estimated incidence between RP and HRP were substantial, ranging from 1.1/100,000 in Guizhou (HRP 5.4/100,000, RP 6.5/100,000) to −9.9/100,000 in Shanghai (HRP 27.6/100,000, RP 17.7/100,000). The analysis estimated that the provinces with the largest differences between incident cases among RP and HRP were Guangdong (469 cases, 9.7% relative to HRP cases) and Jiangsu (305 cases, 6.2% relative to HRP cases). Inflow provinces tended to have their cases underestimated and their incidence overestimated, whereas the opposite pattern was observed in outflow provinces. Incidence exhibits significant spatial clustering, with higher incidence in eastern coastal metropolitan areas and lower incidence in central-western regions and northeastern parts. Conclusion: Migration substantially influences prostate cancer incidence surveillance in China. Incorporating migrant-adjusted estimates provides a more accurate representation of disease burden, supports equitable allocation of healthcare resources, and offers methodological guidance for improving cancer registration systems in regions with high population mobility.

Keshan disease, a condition endemic to China, has been associated with selenium deficiency. However, spatial epidemiological research that integrates Keshan disease prevalence with selenium nutrition remains limited. This study aimed to explore the bivariate spatial autocorrelation between Keshan disease prevalence and serum selenoprotein P levels in Heilongjiang Province, in order to provide evidence for disease prevention and control. Prevalence data for all forms of Keshan disease were obtained through a national surveillance project and analyzed using spatial empirical Bayesian smoothing. Serum selenoprotein P levels were measured using enzyme-linked immunosorbent assay. Bivariate spatial autocorrelation analyses were conducted to assess the relationship between selenoprotein P levels and the prevalence of overall Keshan disease, chronic Keshan disease, and latent Keshan disease. Local spatial autocorrelation analysis of selenoprotein P levels revealed three low–low clusters, one low–high cluster, five high–low clusters, and one high–high cluster. No bivariate global spatial autocorrelation was observed between selenoprotein P levels and the prevalence of overall Keshan disease, chronic Keshan disease, or latent Keshan disease. In total, eight local clusters were identified, including three high–low and five low–high clusters. Nine clusters were identified as key regions for Keshan disease. The risk of Keshan disease among low-selenium nutrition groups was concentrated in three clusters. Residents living in or adjacent to these regions should be prioritized for prevention and control interventions.

Cancer registration in mainland China traditionally focuses on household-registered residents (HRR) and does not include the migrant population among permanent residents (PR), leading to significant selection bias. Estimating incidence among permanent residents provides a less biased and more representative measure of the true incidence. We developed a Bayesian Integrated Nested Laplace Approximation with Stochastic Partial Differential Equation model, incorporating inter-provincial migrant population weights to estimate uterine corpus cancer incidence among permanent residents. The findings revealed a substantial interprovincial migrant population of 67,509,881 individuals, with Shanghai and Beijing showing relatively high difference proportions of 39.6% and 37.0%, respectively. Nationally, the differences in estimated uterine corpus cancer incidence between female PR and HRR were marginal, ranging from 0.2/100,000 in Qinghai to − 0.4/100,000 in Shanghai. The analysis estimated that the provinces with the largest differences between incident cases among female PR and HRR were Henan (− 899 cases, 15.7%) and Guangdong (630 cases, 13.7%). This research holds significant implications for countries relying on HRR-based cancer registration system, particularly those contending with substantial migrant populations. The estimated differences in uterine corpus cancer incidence between PR and HRR provide crucial data support for optimizing prevention strategies and enabling precise allocation of regional healthcare resources.

Bivariate spatial autocorrelation between the prevalence of Keshan disease (KD) and hair selenium levels in residents of KD-endemic areas has not previously been reported. In this study, we investigated the types of spatial clusters between hair selenium levels and the prevalence of KD among residents of KD-endemic areas in Heilongjiang Province, using a bivariate spatial autocorrelation analysis. Population demographic information and hair samples were obtained using questionnaires and sample collection, respectively. Hair selenium levels were determined using hydride generation atomic fluorescence spectroscopy. Bivariate spatial autocorrelation analysis was performed using Geoda 1.22.0.2. The results revealed that the prevalence of individuals with chronic Keshan disease (CKD) with hair selenium levels in the high-low (HL) cluster was greatest in Shangzhi. Furthermore, the prevalences of KD-hair selenium levels, CKD-hair selenium levels, and latent Keshan disease (LKD)-hair selenium levels were classified into low–high (LH) clusters in Lindian, Saertu, Mashan, Dongning, Ningan, and Lanxi. Four high-selenium areas, centered on Lindian, Saertu, Mashan, and Ningan, and one low-selenium area, centered on Shangzhi, were located in Heilongjiang Province. Thus, the residents of KD-endemic areas in Heilongjiang Province remain at risk of selenium deficiency. Additionally, the geographical scope of HL clusters in terms of the number of counties was much smaller than that of LH clusters. This study provides a basis for the identification of KD-endemic areas with low selenium levels and the development of precise prevention and control strategies for KD in Heilongjiang Province.

Selenium deficiency is strongly associated with Keshan disease (KD). It is essential to investigate serum selenium levels from the perspective of etiological surveillance and evaluation of the populations at-risk. This research aims to identify areas at high risk of selenium deficiency using spatial epidemiological methods and to provide a scientific basis for targeted selenium supplementation. This cross-sectional study measured serum selenium levels using an atomic fluorescence spectrometer among 6,643 residents in KD endemic and non-endemic counties across China, applying spatial epidemiological methods, including global autocorrelation and local Moran’s I analysis, to identify spatial clustering. Additionally, ordinary least squares regression was used to assess the correlation between serum selenium levels and per capita disposable income. The median serum selenium level among participants was 66.69 μg/L. Serum selenium levels of residents in KD endemic areas were significantly lower than those in non-endemic areas. A total of 72 counties were classified as selenium-deficient, 638 as selenium-marginal, and 967 as selenium-sufficient. Spatial clustering analysis identified 375 clusters of low selenium levels. Regression analysis revealed a positive correlation between serum selenium levels and per capita disposable income. The geographical distribution of serum selenium levels exhibits significant spatial clustering. The low-low clustering areas in provinces such as Sichuan, Shaanxi, Gansu and Yunnan should be prioritized for selenium nutrition interventions and surveillance. Furthermore, areas with low incomes may need enhanced strategies for selenium supplementation.