Abstract: The tumor microenvironment (TME) is a complex ecosystem surrounding cancer cells, encompassing stromal, immune, and extracellular elements. Understanding the intricate multilinear interactions among different cell types/states, and TME regions is crucial for gaining deeper insights into disease progression, immune evasion mechanisms, and immunotherapy responses. In our research, we leverage the power of tensors to represent and analyze these multilinear interactions within the TME. To account for the underlying distribution and expected sparsity inherent in genomics data, we have incorporated the Zero-Inflated Poisson Model. To address the stochasticity of tensor factorization, we employ a consensus meta-analysis approach, enhancing the robustness of our results. In summary, we present a comprehensive framework, "Robust Bayesian Tensor Factorization with Zero-Inflated Poisson Model and Consensus Aggregation" to characterize the TME of Hodgkin Lymphoma. Our methodology, complemented by a rigorous statistical pipeline, promises to deepen our understanding of the TME by unveiling cellular heterogeneity and intrinsic pathways that drive tumor growth and immune evasion. By elucidating these critical aspects, we aim to contribute significantly to the field, ultimately facilitating the development of more effective therapeutic strategies.