This study presents an extension of the S-MESH (Satellite and ML-based Estimation of Surface air quality at High resolution) framework that integrates validated low-cost sensor (LCS) data into a machine learning model to estimate surface PM2.5 across Central Europe at 1 km resolution. Two integration strategies are tested using LCS data as a target variable and as a spatial input feature. Results show that incorporating LCS data as a spatial feature significantly improves model accuracy, particularly in urban areas, demonstrating the potential of large-scale sensor networks to enhance high-resolution air quality monitoring and complement traditional monitoring stations.