Using image segmentation models to analyse high-resolution earth observation data: new tools to monitor disease risks in changing environments.

BACKGROUND: In the near future, the incidence of mosquito-borne diseases may expand to new sites due to changes in temperature and rainfall patterns caused by climate change. Therefore, there is a need to use recent technological advances to improve vector surveillance methodologies. Unoccupied Aerial Vehicles (UAVs), often called drones, have been used to collect high-resolution imagery to map detailed information on mosquito habitats and direct control measures to specific areas. Supervised classification approaches have been largely used to automatically detect vector habitats. However, manual data labelling for model training limits their use for rapid responses. Open-source foundation models such as the Meta AI Segment Anything Model (SAM) can facilitate the manual digitalization of high-resolution images. This pre-trained model can assist in extracting features of interest in a diverse range of images. Here, we evaluated the performance of SAM through the Samgeo package, a Python-based wrapper for geospatial data, as it has not been applied to analyse remote sensing images for epidemiological studies. RESULTS: We tested the identification of two land cover classes of interest: water bodies and human settlements, using different UAV acquired imagery across five malaria-endemic areas in Africa, South America, and Southeast Asia. We employed manually placed point prompts and text prompts associated with specific classes of interest to guide the image segmentation and assessed the performance in the different geographic contexts. An average Dice coefficient value of 0.67 was obtained for buildings segmentation and 0.73 for water bodies using point prompts. Regarding the use of text prompts, the highest Dice coefficient value reached 0.72 for buildings and 0.70 for water bodies. Nevertheless, the performance was closely dependent on each object, landscape characteristics and selected words, resulting in varying performance. CONCLUSIONS: Recent models such as SAM can potentially assist manual digitalization of imagery by vector control programs, quickly identifying key features when surveying an area of interest. However, accurate segmentation still requires user-provided manual prompts and corrections to obtain precise segmentation. Further evaluations are necessary, especially for applications in rural areas.

The incidence, and spatial trends of cholera in Sabah over 15 years: Repeated outbreaks in coastal areas.

Vibrio cholerae remains a notable public health challenge across Malaysia. Although the Malaysian state of Sabah is considered a cholera-affected area, gaps remain in understanding the epidemiological trends and spatial distribution of outbreaks. Therefore, to determine longitudinal and spatial trends in cholera cases data were obtained from the Sabah State Health Department for all notified cases of cholera between 2005-2020. A cholera outbreak is defined as one or more confirmed cases in a single locality with the evidence of local transmission. All records were geolocated to village level. Satellite-derived data and generalised linearized models were used to assess potential risk factors, including population density, elevation, and distance to the sea. Spatiotemporal clustering of reported cholera cases and zones of increased cholera risk were evaluated using the tau statistic (τ) at 550m, 5km and 10km distances. Over a 15-year period between 2005-2020, 2865 cholera cases were recorded in Sabah, with a mean incidence rate of 5.6 cases per 100,000 (95% CI: 3.4-7.9). From 2015-2020, 705 symptomatic cases and 727 asymptomatic cases were reported. Symptomatic cases primarily occurred in local Malaysian populations (62.6%, 441/705) and in children and adolescents under 15-years old (49.4%, 348/705). On average, cases were reported in areas with low population density (19.45 persons/km2), low elevations (19.45m) and near coastal areas. Spatiotemporal clustering of cholera cases was identified up to 3.5km, with increased village-level cholera risk within 500m and 5 days of initial case presentation to a health facility (Risk Ratio = 9.7, 95% CI: 7.5-12.4). Cholera incidence has high spatial and temporal heterogeneity within Sabah, with some districts experiencing repeated outbreaks. Cholera cases clustered across space and time, with village-level risk of cholera highest within 5 days and within close proximity to primary case villages, suggesting local transmission.

Landscape drives zoonotic malaria prevalence in non-human primates.

Zoonotic disease dynamics in wildlife hosts are rarely quantified at macroecological scales due to the lack of systematic surveys. Non-human primates (NHPs) host Plasmodium knowlesi, a zoonotic malaria of public health concern and the main barrier to malaria elimination in Southeast Asia. Understanding of regional P. knowlesi infection dynamics in wildlife is limited. Here, we systematically assemble reports of NHP P. knowlesi and investigate geographic determinants of prevalence in reservoir species. Meta-analysis of 6322 NHPs from 148 sites reveals that prevalence is heterogeneous across Southeast Asia, with low overall prevalence and high estimates for Malaysian Borneo. We find that regions exhibiting higher prevalence in NHPs overlap with human infection hotspots. In wildlife and humans, parasite transmission is linked to land conversion and fragmentation. By assembling remote sensing data and fitting statistical models to prevalence at multiple spatial scales, we identify novel relationships between P. knowlesi in NHPs and forest fragmentation. This suggests that higher prevalence may be contingent on habitat complexity, which would begin to explain observed geographic variation in parasite burden. These findings address critical gaps in understanding regional P. knowlesi epidemiology and indicate that prevalence in simian reservoirs may be a key spatial driver of human spillover risk.

Zoonotic diseases are infectious diseases that are transmitted from animals to humans. For example, the malaria-causing parasite Plasmodium knowlesi can be transmitted from monkeys to humans through mosquitos that have previously fed on infected monkeys. In Malaysia, progress towards eliminating malaria is being undermined by the rise of human incidences of 'monkey malaria', which has been declared a public health threat by The World Health Organisation. In humans, cases of monkey malaria are higher in areas of recent deforestation. Changes in habitat may affect how monkeys, insects and humans interact, making it easier for diseases like malaria to pass between them. Deforestation could also change the behaviour of wildlife, which could lead to an increase in infection rates. For example, reduced living space increases contact between monkeys, or it may prevent behaviours that help animals to avoid parasites. Johnson et al. wanted to investigate how the prevalence of malaria in monkeys varies across Southeast Asia to see whether an increase of Plasmodium knowlesi in primates is linked to changes in the landscape. They merged the results of 23 existing studies, including data from 148 sites and 6322 monkeys to see how environmental factors like deforestation influenced the amount of disease in different places. Many previous studies have assumed that disease prevalence is high across all macaques, monkey species that are considered pests, and in all places. But Johnson et al. found that disease rates vary widely across different regions. Overall disease rates in monkeys are lower than expected (only 12%), but in regions with less forest or more 'fragmented' forest areas, malaria rates are higher. Areas with a high disease rate in monkeys tend to further coincide with infection hotspots for humans. This suggests that deforestation may be driving malaria infection in monkeys, which could be part of the reason for increased human infection rates. Johnsons et al.'s study has provided an important step towards better understanding the link between deforestation and the levels of monkey malaria in humans living nearby. Their study provides important insights into how we might find ways of managing the landscape better to reduce health risks from wildlife infection.

Serological evidence of zoonotic filovirus exposure among bushmeat hunters in Guinea.

Human Ebola virus (EBOV) outbreaks caused by persistent EBOV infection raises questions on the role of zoonotic spillover in filovirus epidemiology. To characterise filovirus zoonotic exposure, we collected cross-sectional serum samples from bushmeat hunters (n = 498) in Macenta Prefecture Guinea, adjacent to the index site of the 2013 EBOV-Makona spillover event. We identified distinct immune signatures (20/498, 4.0%) to multiple EBOV antigens (GP, NP, VP40) using stepwise ELISA and Western blot analysis and, live EBOV neutralisation (5/20; 25%). Using comparative serological data from PCR-confirmed survivors of the 2013-2016 EBOV outbreak, we demonstrated that most signatures (15/20) were not plausibly explained by prior EBOV-Makona exposure. Subsequent data-driven modelling of EBOV immunological outcomes to remote-sensing environmental data also revealed consistent associations with intact closed canopy forest. Together our findings suggest exposure to other closely related filoviruses prior to the 2013-2016 West Africa epidemic and highlight future surveillance priorities.

Quantifying human-animal contact rates in Malaysian Borneo: Influence of agricultural landscapes on contact with potential zoonotic disease reservoirs.

Changing landscapes across the globe, but particularly in Southeast Asia, are pushing humans and animals closer together and may increase the likelihood of zoonotic spillover events. Malaysian Borneo is hypothesized to be at high risk of spillover events due to proximity between reservoir species and humans caused by recent deforestation in the region. However, the relationship between landscape and human-animal contact rates has yet to be quantified. An environmentally stratified cross-sectional survey was conducted in Sabah, Malaysia in 2015, collecting geolocated questionnaire data on potential risk factors for contact with animals for 10,100 individuals. 51% of individuals reported contact with poultry, 46% with NHPs, 30% with bats, and 2% with swine. Generalised linear mixed models identified occupational and demographic factors associated with increased contact with these species, which varied when comparing wildlife to domesticated animals. Reported contact rates with each animal group were integrated with remote sensing-derived environmental data within a Bayesian framework to identify regions with high probabilities of contact with animal reservoirs. We have identified high spatial heterogeneity of contact with animals and clear associations between agricultural practices and high animal rates. This approach will help inform public health campaigns in at-risk populations and can improve pathogen surveillance efforts on Malaysian Borneo. This method can additionally serve as a framework for researchers looking to identify targets for future pathogen detection in a chosen region of study.