Neoplasms change over time through cell-level evolutions. Cellular tumor components are shaped by selective pressures of microenvironmental ecology and are driven by genetic and epigenetic alterations. However, modeling the dynamics of tumor evolution and ecology centralizes primarily on tumor cells and largely under-appreciates cell-cell interplays. We leverage four models, prey-predator, competitive, generalized, and random Lotka-Volterra models to study cell interactions in the heterogeneous colorectal cancer microenvironment from cancer patient samples by carrying out wavelet-based transformations of scRNA-seq data. Further, we infer cellular dynamics by cell type pseudotime data from trajectory analysis. We derive a game-theoretic Lotka-Volterra model that attempts to model among and within group interactions more accurately. Our proposed ecological models and game-theoretic approaches might shed light on cancer evolution.