arXiv subjects

Chenghang Li

Publications and source records attributed to Chenghang Li.

9 recordsLinked to original sources

arXiv2026

Modeling tumor cell heterogeneity and plasticity in adaptive therapy

Adaptive therapy (AT) is designed to postpone the emergence of drug resistance by exploiting evolutionary competition among tumor subclones. Most mathematical models of AT assume a binary population structure of drug-sensitive and drug-resistant cells, which neglects the continuous nature of phenotypic plasticity. In this study, we propose a mathematical model that integrates a continuous drug susceptibility index with a probabilistic inheritance function to describe clonal dynamics under therapy. The resulting integro-differential system generalizes traditional two-type competition models and captures both heterogeneity and plasticity of tumor cells. Analytical and numerical studies show that (i) continuous therapy drives rapid expansion of resistant clones, (ii) adaptive therapy maintains long-term tumor control by dynamically regulating sensitive populations, and (iii) high phenotypic plasticity accelerates phenotype switching, leading to earlier tumor relapse following continuous therapy. These results identify critical parameter regimes where adaptive therapy outperforms fixed regimens and highlight the essential role of plasticity in shaping treatment outcomes. The proposed framework provides a more realistic mathematical foundation for the design of clinically relevant adaptive therapy strategies.

q-bio.PE↗

arXiv2026

Mathematical model of tumor-macrophage interactions: Elucidating the tumor-driven macrophage phenotype reprogramming

The interplay between tumor cells and macrophages plays a central regulatory role in cancer progression. In this study, we developed a mathematical model that incorporates tumor cells, M1 type macrophages, M2 type macrophages and an M3 type macrophage population characterized by dual phenotypic features. First, we analyzed the fundamental mathematical properties of the model and derived the conditions under which the system attains a tumor free stable state or a coexistence state of tumor and immune cells. Second, global sensitivity analysis revealed that key parameters governing macrophage polarization and intercellular communication vary dynamically during tumor development. Bifurcation analysis further identified the polarization rate of M1 type macrophages $\kappa$ and the baseline level of resting macrophages $M_0$ as critical determinants of the system's dynamical behavior. Notably, using approximate Bayesian computation for parameter inference and dynamic simulations, the model successfully recapitulated the evolutionary trajectories of eight tumor samples. The results demonstrate that lower tumor burden is significantly associated with higher M1 type macrophage infiltration and delayed peak time of M3 type macrophage activation. Moreover, survival analysis indicated that both enhanced M1 type macrophage infiltration and delayed peak time of M3 type macrophage activation are correlated with longer survival time. In summary, this study not only provides a theoretical framework for understanding the dynamic mechanisms underlying tumor macrophage interactions but also proposes two potential clinical prognostic markers: the level of M1 type macrophage infiltration and the peak time of M3 type macrophage activation.

q-bio.PE↗

arXiv2026

Combination therapy for colorectal cancer with anti-PD-L1 and cancer vaccine: A multiscale mathematical model of tumor-immune interactions

The tumor-immune system plays a critical role in colorectal cancer progression. Recent preclinical and clinical studies showed that combination therapy with anti-PD-L1 and cancer vaccines improved treatment response. In this study, we developed a multiscale mathematical model of interactions among tumors, immune cells, and cytokines to investigate tumor evolutionary dynamics under different therapeutic strategies. Additionally, we established a computational framework based on approximate Bayesian computation to generate virtual tumor samples and capture inter-individual heterogeneity in treatment response. The results demonstrated that a multiple low-dose regimen significantly reduced advanced tumor burden compared to baseline treatment in anti-PD-L1 therapy. In contrast, the maximum dose therapy yielded superior tumor growth control in cancer vaccine therapy. Furthermore, cytotoxic T cells were identified as a consistent predictive biomarker both before and after treatment initiation. Notably, the cytotoxic T cells-to-regulatory T cells ratio specifically served as a robust pre-treatment predictive biomarker, offering potential clinical utility for patient stratification and therapy personalization.

q-bio.PE↗

arXiv2026

Quantitative cancer-immunity cycle modeling to optimize bevacizumab and atezolizumab combination therapy for advanced renal cell carcinoma

The incidence of advanced renal cell carcinoma(RCC) has been rising, presenting significant challenges due to the limited efficacy and severe side effects of traditional radiotherapy and chemotherapy. While combination immunotherapies show promise, optimizing treatment strategies remains difficult due to individual heterogeneity. To address this, we developed a Quantitative Cancer-Immunity Cycle (QCIC) model that integrates ordinary differential equations with stochastic modelling to quantitatively characterize and predict tumor evolution in patients with advanced RCC. By systematically integrating quantitative systems pharmacology principles with biological mechanistic knowledge, we constructed a virtual patient cohort and calibrated the model parameters using clinical immunohistochemistry data to ensure biological validity. To enhance predictive performance, we coupled the model with pharmacokinetic equations and defined the Tumor Response Index (TRI) as a quantitative metric of efficacy. Systematic analysis of the QCIC model allowed us to determine an optimal treatment regimen for the combination of bevacizumab and atezolizumab and identify tumor biomarkers with clinical predictive value. This study provides a theoretical framework and methodological support for precision medicine in the treatment of advanced RCC.

q-bio.QM↗

arXiv2025

Mathematical modeling of tumor-immune interactions: methods, applications, and future perspectives

Mathematical oncology is a rapidly evolving interdisciplinary field that uses mathematical models to enhance our understanding of cancer dynamics, including tumor growth, metastasis, and treatment response. Tumor-immune interactions play a crucial role in cancer biology, influencing tumor progression and the effectiveness of immunotherapy and targeted treatments. However, studying tumor dynamics in isolation often fails to capture the complex interplay between cancer cells and the immune system, which is critical to disease progression and therapeutic efficacy. Mathematical models that incorporate tumor-immune interactions offer valuable insights into these processes, providing a framework for analyzing immune escape, treatment response, and resistance mechanisms. In this review, we provide an overview of mathematical models that describe tumor-immune dynamics, highlighting their applications in understanding tumor growth, evaluating treatment strategies, and predicting immune responses. We also discuss the strengths and limitations of current modeling approaches and propose future directions for the development of more comprehensive and predictive models of tumor-immune interactions. We aim to offer a comprehensive guide to the state of mathematical modeling in tumor immunology, emphasizing its potential to inform clinical decision-making and improve cancer therapies.

q-bio.QM↗

arXiv2025

Face-MakeUpV2: Facial Consistency Learning for Controllable Text-to-Image Generation

In facial image generation, current text-to-image models often suffer from facial attribute leakage and insufficient physical consistency when responding to local semantic instructions. In this study, we propose Face-MakeUpV2, a facial image generation model that aims to maintain the consistency of face ID and physical characteristics with the reference image. First, we constructed a large-scale dataset FaceCaptionMask-1M comprising approximately one million image-text-masks pairs that provide precise spatial supervision for the local semantic instructions. Second, we employed a general text-to-image pretrained model as the backbone and introduced two complementary facial information injection channels: a 3D facial rendering channel to incorporate the physical characteristics of the image and a global facial feature channel. Third, we formulated two optimization objectives for the supervised learning of our model: semantic alignment in the model's embedding space to mitigate the attribute leakage problem and perceptual loss on facial images to preserve ID consistency. Extensive experiments demonstrated that our Face-MakeUpV2 achieves best overall performance in terms of preserving face ID and maintaining physical consistency of the reference images. These results highlight the practical potential of Face-MakeUpV2 for reliable and controllable facial editing in diverse applications.

cs.CV↗

arXiv2025

A stochastic agent-based model for simulating tumor-immune dynamics and evaluating therapeutic strategies

Tumor-immune interactions are central to cancer progression and treatment outcomes. In this study, we present a stochastic agent-based model that integrates cellular heterogeneity, spatial cell-cell interactions, and drug resistance evolution to simulate tumor growth and immune response in a two-dimensional microenvironment. The model captures dynamic behaviors of four major cell types--tumor cells, cytotoxic T lymphocytes, helper T cells, and regulatory T cells--and incorporates key biological processes such as proliferation, apoptosis, migration, and immune regulation. Using this framework, we simulate tumor progression under different therapeutic interventions, including radiotherapy, targeted therapy, and immune checkpoint blockade. Our simulations reproduce emergent phenomena such as immune privilege and spatial immune exclusion. Quantitative analyses show that all therapies suppress tumor growth to varying degrees and reshape the tumor microenvironment. Notably, combination therapies--especially targeted therapy with immunotherapy--achieve the most effective tumor control and delay the emergence of resistance. Additionally, sensitivity analyses reveal a nonlinear relationship between treatment intensity and therapeutic efficacy, highlighting the existence of optimal dosing thresholds. This work demonstrates the utility of agent-based modeling in capturing complex tumor-immune dynamics and provides a computational platform for optimizing cancer treatment strategies. The model is extensible, biologically interpretable, and well-suited for future integration with experimental or clinical data.

q-bio.TO↗

arXiv2025

Face-MakeUp: Multimodal Facial Prompts for Text-to-Image Generation

Facial images have extensive practical applications. Although the current large-scale text-image diffusion models exhibit strong generation capabilities, it is challenging to generate the desired facial images using only text prompt. Image prompts are a logical choice. However, current methods of this type generally focus on general domain. In this paper, we aim to optimize image makeup techniques to generate the desired facial images. Specifically, (1) we built a dataset of 4 million high-quality face image-text pairs (FaceCaptionHQ-4M) based on LAION-Face to train our Face-MakeUp model; (2) to maintain consistency with the reference facial image, we extract/learn multi-scale content features and pose features for the facial image, integrating these into the diffusion model to enhance the preservation of facial identity features for diffusion models. Validation on two face-related test datasets demonstrates that our Face-MakeUp can achieve the best comprehensive performance.All codes are available at:https://github.com/ddw2AIGROUP2CQUPT/Face-MakeUp

cs.CV↗

arXiv2023

Structure Embedded Nucleus Classification for Histopathology Images

Nuclei classification provides valuable information for histopathology image analysis. However, the large variations in the appearance of different nuclei types cause difficulties in identifying nuclei. Most neural network based methods are affected by the local receptive field of convolutions, and pay less attention to the spatial distribution of nuclei or the irregular contour shape of a nucleus. In this paper, we first propose a novel polygon-structure feature learning mechanism that transforms a nucleus contour into a sequence of points sampled in order, and employ a recurrent neural network that aggregates the sequential change in distance between key points to obtain learnable shape features. Next, we convert a histopathology image into a graph structure with nuclei as nodes, and build a graph neural network to embed the spatial distribution of nuclei into their representations. To capture the correlations between the categories of nuclei and their surrounding tissue patterns, we further introduce edge features that are defined as the background textures between adjacent nuclei. Lastly, we integrate both polygon and graph structure learning mechanisms into a whole framework that can extract intra and inter-nucleus structural characteristics for nuclei classification. Experimental results show that the proposed framework achieves significant improvements compared to the state-of-the-art methods.

cs.CV↗