A Universal Balance and Feedback Framework for Cancer-Cell State Transitions A Node-Network Theory of Malignant Attractors, Cellular Reversion, and Regulatory-State Control (Version 1.1)

Cancer is commonly characterized by genetic mutations, epigenetic alterations, abnormal signaling, uncontrolled proliferation, altered metabolism, immune evasion, and disruption of cellular differentiation. Recent systems-biology research provides evidence that malignant cellular phenotypes can also be understood as dynamically maintained states of complex regulatory networks. In particular, recent work from the Korea Advanced Institute of Science and Technology (KAIST) has demonstrated computational and experimental approaches for identifying regulatory interventions capable of driving colorectal cancer cells toward normal-like differentiated states. The BENEIN framework identified MYB, HDAC2, and FOXA2 as regulators whose simultaneous inhibition induced enterocyte differentiation, while the subsequent REVERT framework analyzed attractor landscapes during colorectal tumorigenesis and experimentally investigated candidate reversion mechanisms. This paper proposes a Universal Balance and Feedback Framework (UBFF) formulation of cellular state transitions. A biological system is represented as a weighted, directed network of interacting nodes. Each node possesses a dynamic state, while interactions generate reinforcing and balancing feedback. Cellular phenotypes emerge from the collective state of the network rather than from isolated molecular components. The central hypothesis is that malignant phenotypes can, under appropriate biological conditions, be represented as stable or metastable network configurations maintained by abnormal feedback relationships. Cancer-cell reversion is formulated as a controlled transition between attractors, A_C →┴( U ) A_N, where A_C is a malignant attractor and A_N is a normal or differentiated attractor. A generalized UBFF node equation is proposed: (dx_i)/dt=K_i (x_i )+∑_(j=1)^N▒W_ij F_ij (x_j,x_i )+B_i (X)+U_i (t)-R_i (x_i )+η_i (t), where x_i is the state of node i, K_i is intrinsic dynamics, W_ij is relational coupling, F_ij is interaction, B_i is collective feedback, U_i is external intervention, R_i is relaxation (dissipation), and η_i is stochastic variation. Version 1.1 adds a linearized reversion-threshold relation. Near a malignant fixed point with Jacobian J_C, the sustained input required to hold the system at a differentiated state separated from the malignant state by ΔX is U_clamp=-J_C ΔX, bounded by the extreme singular values of J_C. A node-wise control score derived from the same linearization provides a testable ranking of candidate reversion switches. The proposed model is not presented as a demonstrated cancer treatment. It is a theoretical framework intended to generate experimentally testable predictions.