The aim of this project is to develop new mathematical theory to support decision making in spatial stochastic settings. Integrating principles from Markov decision theory, spatial statistics and statistical mechanics, we introduce and analyze new types of decision processes characterized by both spatial and temporal interaction structures. Also, we provide theoretical insights in the existence and the shape of optimal policies in such processes and develop algorithms to compute them. Through this work, we aim to pioneer a new chapter in the already rich theory on decision processes and to provide a firm framework that finds application in a variety of fields, such as ecology, economics, logistics and healthcare. One that sparks our interest in particular is the application to cancer research. Previous attempts to model tumor evolution failed to sufficiently account for the spatial interaction structures between cells in the human body. We hope that our research will provide the tools to resolve this issue and open new doors on the path towards a cure for cancer.