Cutaneous leishmaniasis in a hyperendemic metropolitan area in Iran: spatial probability modeling by machine-learning.

Abstract

Cutaneous leishmaniasis (CL), a parasitic, vector-borne infection with a broad, global distribution is prevalent in Mashhad, a city in north-eastern Iran, which is known as a significant hyperendemic area both for anthroponotic and zoonotic CL. This study evaluates the spatial distribution probability of CL prevalence by examining various factors such as socio-demography, built environment, geology, and climate within Mashhad. Data from 3,033 CL patients diagnosed between 2013 and 2020 were analyzed. Sociological, environmental, and climatic variables were assessed using the generalized linear regression model (GLM) and the maximum entropy (MaxEnt) model. The results indicate socio-demography and built environment and geology as significant influencers of the CL distribution in Mashhad. The MaxEnt model identified 42.6% (138.5 km2) of the study area as high-risk for CL, including both urban and rural areas characterized by specific geological and geographical conditions, high urbanization rates, and poor environmental quality. As confirmed by previous studies, we found that ~0.04% of the city's population were infected, young individuals with low literacy levels and those living in densely populated areas. In addition to the known environmental variables, such as temperatures between 20 and 40 °C and humidity, we also found risk areas associated with built environment and a certain type of exposed rocks that is attractive for the vector. The findings presented provide valuable insights for urban planners and health managers to target CL control programs and allocate resources effectively.