Cholangiocarcinoma (CCA), a deadly cancer originating from the bile duct, poses an alarming public health crisis in Thailand, particularly in its northeastern region, which has the world's highest incidence rates of 85 infections per 100,000 person-years. This stark contrast is heavily influenced by the Opisthorchis viverrini infection, a parasitic disease endemic to the area, alongside other risk factors such as liver conditions and hepatitis.

In the battle against CCA, spatial epidemiology emerges as a crucial tool in deciphering the intricate patterns and causes underlying this disease. By implementing advanced spatial prediction models, researchers can glean invaluable insights into how environmental and biological factors interplay in driving disease distribution. These models are game-changers, allowing for better resource allocation, targeted public health interventions, and a thorough understanding of the risks involved.

Traditional approaches to spatial data often falter in grasping complex relationships among various factors influencing health. Enter Machine Learning—a powerful ally in analyzing vast datasets with non-linear connections that conventional methods struggle to interpret. Models like Random Forest and Neural Networks stand out by integrating diverse data sources, including satellite imagery and census statistics, paving the way for heightened predictive accuracy.

Aligned with Thailand’s National Artificial Intelligence Strategy, which advocates for AI applications to drive national progress, this study seeks to fill a significant gap left by earlier research, which mainly focused on limited geographical regions and utilized outdated statistical methods. By leveraging Machine Learning to assess CCA incidence across the entire country, we aim to pinpoint high-risk areas and inform public health tactics.

Study Design and Data Collection