Tumor-infiltrating lymphocytes (TILs) are to be subject to clinical applications by cultured TIL infusion in vivo for adoptive cell therapy (ACT) and by ex vivo TIL analysis for determining immune characteristics to kill autologous tumor cells so that TIL has been administrated tumor’s patients to immune-cell therapy and analyze patients’ immune characteristics for tumor diseases. To study the TIL features, we have established quantitative network modeling by TIL’s TCR signaling pathway, IL2 pathway, and TGF-β pathway for personalized immunotherapy for more than fifteen years. However, machine-learning analysis still has some challenges under the traditional quantitative pathway for network configurations to apply for patient treatment. For example, multiple protein complexes competing for downstream DNA binding-site or protein-protein complex will generate different effects. To address this question, we report here a temporal-spatial quantification network, termed a spatial-timely quantification network, to address the spatial-timely competition of complex proteins binding to downstream proteins or DNA in network analysis. After studying spatial-timely quantitative network modeling by TGF-β pathway activity in spatial-timely order, we discover that multiple protein complexes using spatial-timely quantitative networks are much better than traditional quantitative networks. Once the new system modeling is established, we can further analyze all pathways, such as the TCR signaling pathway and IL2 pathway from TIL, for different immunotherapy.