Gastric cancer ranks among the leading malignant tumors worldwide in both incidence and mortality. Its early symptoms are often atypical, and diagnosis frequently occurs at advanced stages, resulting in low surgical cure rates and poor prognosis. Therefore, exploring and identifying early warning signals of gastric cancer development is of great significance for achieving early diagnosis and treatment. In recent years, advances in systems biology and statistical physics have offered new approaches for modeling and predicting complex diseases. In particular, gene regulatory networks, as core mechanisms for maintaining and transitioning cellular phenotypic steady states, contain rich dynamic information. Traditional studies have largely focused on static differences in gene expression, which are insufficient to capture critical dynamic changes during disease progression. This study, grounded in nonequilibrium statistical physics, constructs a regulatory network of key gastric cancer genes and applies theoretical tools such as potential landscape and steady-state probability flux to quantify the system’s dynamic stability and energy dissipation processes. By simulating state transitions under various mutations or regulatory strengths, early warning indicators such as entropy production rate, mean flux, critical slowing down, and flickering are extracted, providing theoretical support for identifying critical transition points in gastric cancer onset. This research offers a novel perspective to analyze the mechanisms of gastric cancer development and its early diagnostic signals from a dynamical viewpoint, and also provides methodological support for establishing clinical early warning systems.

Recently, Dr. Jin Wang from University of Chinese Academy of Sciences/ State University of New York at Stony Brook and Dr. Xiaona Fang from Northeast Normal University published an article titled “Identifying early warning signals of cancer formation” in the journal Quantitative Biology. Using theories of nonequilibrium potential landscape and steady-state probability flux, and indicators such as entropy production rate and mean flux derived from time-series gene expression data, they identified early warning signals during gastric cancer development, providing theoretical basis for early diagnosis.

The simulation results demonstrate that variations in the regulatory strength of TGF-β trigger transitions in the system’s steady-state structure. By quantifying changes in a series of nonequilibrium dynamical indicators, we accurately capture the key dynamical features preceding the steady-state transition. These indicators systematically elucidate the physical mechanisms underlying cancer state formation and provide a robust theoretical foundation for early diagnostic warnings (Figure 1).

Although mean flux and entropy production rate (EPR) can quantify early warning signals in the diffuse gastric cancer (DGC) system, they are difficult to measure in clinical practice. To address this, authors calculated the difference between forward and backward two-gene cross-correlation functions from simulated stochastic gene expression trajectories to quantify time irreversibility, thereby characterizing the system’s nonequilibrium dynamics. This approach enables effective quantification of nonequilibrium behavior in cancer systems without the need for direct measurement of mean flux or EPR, thus overcoming the practical limitations associated with detecting these indicators.
DOI: 10.1002/qub2.81