Organized by: Guillermo Lorenzo (University of A Coruna (Spain)), Chengyue Wu (The University of Texas MD Anderson Cancer Center, US), Ernesto A. B. F. Lima (The University of Texas at Austin, US)

Note: this minisymposia has multiple sessions. The other session is: and Part-2.

1. Thomas E. Yankeelov The University of Texas at Austin
   "A practical computational framework for systematically investigating alternative treatment strategies for cancer"
Over the last decade our team has developed a set of partial differential equations that capture key features of tumors growth and treatment response related to tumor cell migration, proliferation, treatment response, and tissue mechanical properties. These models can be calibrated with widely-available medical imaging data to accurately predict the spatio-temporal changes of solid tumors in response to both radiation and systemic therapies. We will briefly summarize those results for cancers of the brain and breast, and then show how these models power digital twins to identify alternative therapeutic regimens that are hypothesized to outperform the standard-of-care interventions. In the case of high-grade gliomas, we will show how our digital twins can be used to identify personalized treatment plans predicted to reduce tumor burden 24% more than the standard-of-care approach one month after treatment (n = 15), while maintaining toxicity in the organs-at-risk within acceptable clinical limits. In the case of breast cancer, we have used digital twins to personalize neoadjuvant chemotherapy schedules (n = 105) to yield therapeutic strategies that are hypothesized to increase the pathological complete response rate by at least 20%. Furthermore, we have used our digital twin formalism to virtually recapitulate the results of three key clinical trials that led to the current backbone for neoadjuvant therapy. While the results we will present will focus only on cancers of the brain and breast, we emphasize that since our digital twins are based on key underlying biology and physics features of cancer, they are applicable to any solid tumor for which the requisite imaging data is available.
2. Maximilian Strobl Cleveland Clinic
   "What pre-clinical experiments can teach us about digital twins for personalized cancer treatment scheduling"
Cancers are complex and evolving diseases. To tackle this complexity there has been growing interest in developing “digital twins” – personalized computational tumor models – to better inform when and how to treat to reduce toxicity and maximize tumor control. As this idea finds traction, the crucial question is how do we ensure efficacy and safety as we translate from bench to bedside? In this study, we test the digital twin approach to treatment scheduling in vitro, in the context of EGFR+ non-small cell lung cancer. Using fluorescent, time-lapse microscopy we characterize the evolutionary dynamics of co-cultures of Gefitinib-sensitive and paired resistant cell lines (PC9) across four different treatment schedules: i) continuous therapy, ii) intermittent therapy (on/off), iii) intermittent therapy (off/on), iv) continuous therapy at half the full dose. Our results demonstrate that both the dose and the frequency of treatment influence evolutionary dynamics. Intermittent therapy minimizes final resistant cell and total cell count after six treatment changes (18 days total), across four dose levels examined (2uM, 200nM, 100nM, 20nM Gefitinib). Moreover, the off/on intermittent schedule outperforms the on/off schedule, suggesting a role for spatial competition in suppressing resistant cells. Next, we test how well three commonly used mathematical models of sensitive-resistant dynamics can predict the observed dynamics: 1) A simple exponential model, 2) A logistic model which accounts for spatial competition, and 3) A 3-population model which includes an additional subpopulation of drug-tolerant cells in the “sensitive” population. While Models 1 and 2 can capture the dynamics under continuous treatment, the more complex Model 3 is required to predict the outcomes of intermittent treatment. Our work illustrates how in vitro experiments can support the development of digital twins, and how this process can uncover new insights into drug resistance evolution in cancer.
3. Renee Brady-Nicholls H. Lee Moffitt Cancer Center & Research Institute
   "Investigating Response Differences between African American and European American Prostate Cancer Patients Through an In Silico Study"
African American (AA) men have the highest incidence and mortality rates of prostate cancer (PCa) compared to any other racial group. The increased incidence as well as mortality are likely due to socioeconomic factors, environmental exposure, access to care, and biologic variations. Deciphering the specific drivers of increased incidence and mortality is difficult due to a scarcity in available data from AA patients. Mathematical modeling offers the opportunity to run in silico studies to investigate treatment responses in a larger cohort of virtual patients. Here, we investigate response differences between AA and European American (EA) prostate cancer patients receiving hormone therapy. We simulate prostate-specific antigen (PSA) dynamics, using a mathematical model of interactions between PCa stem cells and differentiated cells. We use propensity score matching to identify 15 EA patients that most closely matched the 10 AA patients. Bayesian calibration is used to determine plausible parameter sets that accurately describe longitudinal PSA dynamics on a per-patient basis. Model parameters are compared between AA and EA patients to determine potential drivers of resistance. Our findings show that the initial PSA levels, stem-cell self-renewal, and PSA production rates significantly differ between AA and EA patients. Using the plausible parameter sets drawn from the calibration, we simulate adaptive therapy as a potential strategy to maximally delay progression. Our findings show that both patient groups receive benefit from adaptive therapy when compared to continuous, with AA patients receiving a significantly greater advantage. Our study presents an important step in identifying race-specific, patient-specific treatment options that can be used to maximally delay time to progression.
4. Fatemeh Beigmohammadi Université de Montréal
   "Efficient methods for generating virtual patient cohorts using trajectory-matching ABC-MCMC"
Virtual patient cohorts (VPCs) are computer-generated representations of patients that mirror real-world populations. VPCs use mechanistic mathematical models to establish the effects of inter-patient variability on disease and treatment outcomes, thereby allowing for the comprehensive exploration of disease mechanisms and therapeutic strategies at low-cost and without burden to patients. Further, they may aid the development and validation of mechanistic mathematical models that aim to capture the underlying mechanisms of biological systems and their responses to drug treatments. Thus, the effective generation of virtual populations is critical. While several methods exist to create VPCs, there is a growing need for more computationally efficient techniques. Previous work combined Approximate Bayesian Computation (ABC) with Markov Chain Monte Carlo (MCMC) to generate heterogeneity in model parameters and predicted outcomes. Unfortunately, because the generated samples must simultaneously meet the ABC and MCMC acceptance criteria, this approach has a very high rejection rate. To reduce this computational burden, we developed a model-based technique we call Trajectory Matching ABC-MCMC (TM-ABC-MCMC). TM-ABC-MCMC modifies the acceptance criteria in ABC-MCMC to only require that model trajectories fall within observed bounds. In this talk, I will discuss the application of TM-ABC-MCMC to three existing mechanistic models of varying complexity and will show that it has a lower rejection rate compared to ABC-MCMC and maintains high fidelity with previous results. Thus, our approach significantly decreases computational costs, boosting the efficiency of virtual patient cohort generation.

Organized by: Mohit Kumar Jolly (Indian Institute of Science), Sarthak Sahoo (Indian Institute of Science)

Note: this minisymposia has multiple sessions. The other session is: and Part-2.

1. Rebecca A Bekker University of Southern California
   "Modeling Cell-State Dynamics to Unravel and Counteract Immune Suppression in Breast Cancer Immunotherapy"
HER2+ breast cancer is aggressive and has historically had poor outcomes. Despite therapeutic advances, such as the development and approval of HER2 targeted therapies and the associated improved outcomes, resistance invariably develops. However, recent preclinical and clinical evidence suggest patients may benefit from combination therapies that include immunotherapy. The PANACEA trial, which investigated the efficacy of the combination of the HER2 monoclonal antibody trastuzumab and the PD-1 inhibitor pembrolizumab in HER2+ breast cancer, reported a 15% response rate in patients with PD-L1+ tumors. This unexpectedly low response rate may be the result of a highly tumor-engineered immune-suppressive niche containing MDSCs, Tregs, and TAMs. Understanding the interplay between pro- and anti-tumor immune subsets and HER2+ breast cancer cells is crucial for improving responses to immunotherapy and identifying novel therapeutic strategies. To this end, we developed an agent-based model (ABM) of tumor-immune interactions, which is initialized with digitized multiplex immunohistochemistry slides of untreated spontaneous lung metastases from the NT2.5LM HER2+ murine model. The ABM includes tumor evolution via mutations that affect PD-L1 expression, chemotherapy sensitivity, and proliferation speed. We explore how treatment regimens, combining chemotherapy, anti-PD-L1 and anti-CTLA-4, impact immune composition and suppression, and tumor evolution. The model provides a framework to explore how immune plasticity and tumor adaptation may co-evolve under therapy, and to generate hypotheses about potential mechanisms of resistance and strategies to counteract immunosuppression.
2. James Greene Clarkson University
   "Understanding therapeutic tolerance through a mathematical model of drug-induced resistance"
Resistance to chemotherapy is a major impediment to successful cancer treatment that has been studied over the past three decades. Classically, resistance is thought to arise primarily through random genetic mutations, after which mutated cells expand via Darwinian selection. However, recent experimental evidence suggests this evolution to resistance need not occur randomly, but instead may be induced by the application of the drug. In this work, we present a mathematical model that describes induced resistance. We utilize our mathematical model to study control-theoretic questions with respect to different clinical treatment protocols, and study the effect of different therapies across parameter regimes (e.g. we investigate patient-specific responses). An extended model is then fit to time-resolved in vitro experimental data. From observational data of total numbers of cells, the model unravels the relative proportions of sensitive and resistance subpopulations and quantifies their dynamics as a function of drug dose; the predictions are then validated using data on drug doses that were not used when fitting parameters. Optimal control techniques are then utilized to discover dosing strategies that could lead to better outcomes as quantified by lower total cell volume. These results are reported in J.L Gevertz, J.M Greene, S. Prosperi, N. Comandante-Lou, and E.D. Sontag. Understanding therapeutic tolerance through a mathematical model of drug-induced resistance. npj Systems Biology and Applications, 11, 2025
3. Sara Hamis Uppsala University
   "Growth rate-driven modelling elucidates phenotypic adaptation in BRAFV600E-mutant melanoma"
Phenotypic adaptation, the ability of cells to change phenotype in response to external pressures, has been identified as a driver of drug resistance in cancer. To quantify phenotypic adaptation in BRAFV600E-mutant melanoma, we develop a theoretical model that emerges from data analysis of WM239A-BRAFV600E cell growth rates in response to drug challenge with the BRAF-inhibitor encorafenib. Our model constitutes a cell population model in which each cell is individually described by one of multiple discrete and plastic phenotype states that are directly linked to drug-dependent net growth rates and, by extension, drug resistance. Data-matched simulations reveal that phenotypic adaptation in the cells is directed towards states of high net growth rates, which enables evasion of drug-effects. The model subsequently provides an explanation for when and why intermittent treatments outperform continuous treatments in vitro, and demonstrates the benefits of not only targeting, but also leveraging, phenotypic adaptation in treatment protocols.
4. Russell C Rockne Beckman Research Institute, City of Hope
   "State-transitions at the single cell and system levels in chronic and acute myeloid leukemia"
In this presentation, I will discuss experimental data and mathematical models used to study state transitions in chronic and acute myeloid leukemia (CML and AML). Our experimental approach involves inducible and constitutively activated mouse models of CML and AML, which are monitored longitudinally through blood sampling and RNA sequencing. The mathematical models employed are stochastic differential equations and their corresponding probability density functions. By integrating experimental data with these mathematical models, and iteratively validating the models while generating new hypotheses, we have demonstrated that state transitions can be detected at very early stages of disease initiation. Furthermore, these transitions can be used to predict responses to chemotherapy and tyrosine kinase inhibitor (TKI) therapies. We explore how state-transitions can be used to characterize and quantify resistance to therapy through analysis of gene programs within cell types over time.

Organized by: Ashlee N. Ford Versypt (University at Buffalo, State University of New York), John Metzcar, University of Minnesota

1. Ashlee N. Ford Versypt University at Buffalo, State University of New York
   "Agent-Based Modeling of the Transwell Migration Assay to Inform Tumor-Immune Microenvironment Simulations"
Cell migration in tumor microenvironments is crucial in disease progression and treatment efficacy. Recent experimental studies reported variations in chemotactic migration of cancer and immune cells. A popular method to study chemotaxis is the transwell migration assay. To complement the in vitro experiments to characterize tumor and immune cells in this assay, we developed a 3D agent-based model with Compucell3D to simulate the effects of random and directed cell migration in response to chemokines. To accommodate various cell lines, we categorized targeted cell lines into 6 groups based on their size and adhesion to the membrane. The model shows a 3D column space of the transwell device with 400 moving agents and periodic boundary conditions applied to vertical surfaces of the domain to simulate the dynamics of the in vitro transwell experiments where cells have realistic biomechanics of neighboring cells and tissue-mimic biomaterials. A solid plane contains randomly distributed pores that mimic the structure of the collagen-coated membrane with the same level of pore density. Chemokines are initiated from the bottom half of the assay below the membrane and can diffuse upwards to generate a concentration gradient. Several parameters, including chemical concentrations, diffusion coefficients, chemotactic potential coefficient, an external potential energy term, and a contact energy term are included with a direct connection to published data. The randomized external potential energy simulates the intrinsic Brownian motion of cells and drives cells to move through the membrane in the negative control group without chemokines. Smaller contact energy between cells and the membrane mimics stronger cell-collagen adhesion. The chemotactic energy term and heterogeneous chemical field regulate directional chemotaxis. We observed that larger external potential energy can induce more cells to migrate through the membrane. Thus, we calibrate this energy term with negative control group data from different cell lines (e.g., Panc1, MiaPaCa2, HCT116, U937, and THP-1). Our simulated results also predicted variations in cell migration with cell density and pore density of the membrane in the negative control groups. We employed various methods to generate Brownian motion and analyzed the resulting trajectories with their velocity profiles and mean squared displacements (MSD), finding these methods affected cell persistence, average velocity, and diffusivity. Next, we are extending the model to investigate the effects of chemokine concentrations and diffusion. In the future, we will implement these validated mechanisms and physiological properties into larger systems of agent-based models to simulate cancer pathology and therapy inside the body, considering cells, chemokines, and tumor immune microenvironments in various tissues. Acknowledgments: This work was supported by the National Institutes of Health grant R35GM133763 and the University at Buffalo. Co-author MBD is supported in part by R01 CA226279.
2. Aaron Meyer University of California, Los Angeles
   "Bridging single cell features to the tissue and patient scale with tensor modeling"
High-dimensional single-cell measurements have revolutionized our ability to study the variation within and between heterogeneous cell populations. As single-cell RNA-sequencing (scRNAseq) and similar technologies have become increasingly accessible, these technologies have extended to studies including multiple experimental conditions, samples, or subjects. Multi-condition single-cell experiments can evaluate how heterogeneous cell populations behave across patients according to disease pathology. Analyzing multi-condition single-cell datasets to link cellular heterogeneity to patient-level features like disease state is crucial, but challenged by the limited sample sizes and the inherent misalignment of cell states across individuals. To overcome these hurdles, we developed ULTRA (Unaligned Low-rank Tensor Regression with Attention). ULTRA leverages a tensor framework to model the multi-way data structure and employs an attention mechanism to derive interpretable, cell-specific gene signatures associated with patient features, crucially without requiring prior cell population alignment across samples. In ULTRA, a fit gene signature scores each cell within a sample; an attention mechanism weights cells based on their expression of this signature. The aggregated cells are then regressed against the patient feature of interest. Importantly, despite using attention, a well-known mechanism enabling transformer models, ULTRA is otherwise a linear model, maximizing data use and interpretability. We applied ULTRA across several scRNAseq datasets, demonstrating its consistently superior ability to identify associative signatures with patient-level features. As one example, I will show how ULTRA identifies features of the tumor microenvironment that are predictive of immunotherapy response across several cancers. A key challenge in deriving associations between single cell measurements and patient-level characteristics is that, while there is an abundance of data, these datasets typically include only a few samples due to cost. Therefore, we used the ULTRA model to optimize future single cell profiling experiments. I will cover several lessons regarding optimal cell numbers and read depths for improving the insights from single cell studies at reduced cost.
3. Erzsébet Ravasz Regan The College of Wooster
   "Cell Interrupted — Modular Boolean Modeling of the Coordination between Mitochondrial Dysfunction-Associated Senescence, Cell Cycle Control and the Epithelial to Mesenchymal Transition"
The steady accumulation of senescent cells with aging creates tissue environments that aid cancer evolution. Aging cell states are highly heterogeneous. 'Deep senescent' cells rely on healthy mitochondria to fuel a strong proinflammatory secretome, including cytokines, growth and transforming signals. Yet, the physiological triggers of senescence, such as reactive oxygen species (ROS), can also trigger mitochondrial dysfunction. This energy deficit alters the secretome of these cells and causes chronic oxidative stress – a state termed Mitochondrial Dysfunction-Associated Senescence (MiDAS). Here, we offer a mechanistic molecular model for MiDAS in the form of  a Boolean regulatory network that qualitatively captures key aspects of mitochondrial dynamics during cell cycle progression (hyper-fusion at the G1/S boundary, fission in mitosis), apoptosis (fission and dysfunction) and glucose starvation (reversible hyperfusion), as well as MiDAS in response to SIRT3 knockdown or oxidative stress. We offer testable predictions about the growth factor- and glucose-dependence of MiDAS and its reversibility at different stages of reactive oxygen species (ROS)-induced senescence. We then link this model to our large modular model of mechano-sensitive Epithelial to Myesenhynal Transition, and show that EMT is incompatible with MiDAS. Our models provide mechanistic insights into the distinct stages of DNA-damage induced senescence, the relationship between senescence and epithelial-to- mesenchymal transition in cancer and offers a foundation for building multiscale models of tissue aging.
4. Stacey D. Finley University of Southern California
   "Systems biology modeling and analyses of metabolic phenotypes in the tumor microenvironment"
Colorectal cancer (CRC) remains one of the most prevalent and lethal malignancies worldwide, ranking third in cancer incidence and second in mortality. Despite advances in targeted therapies and immunotherapy, clinical outcomes often remain poor, largely due to the complex nature of the disease that extends beyond just the malignant cells themselves. Like all solid tumors, CRC develops as an ecosystem rather than a simple mass of cancer cells. As tumors progress, metabolic interactions between tumor and stromal cells in the tumor microenvironment (TME) lead to altered cellular growth and contribute to invasion, metastasis, and drug resistance. The complexity of metabolic interactions in the TME requires computational approaches that can capture system-level behaviors. We have developed genome-scale metabolic models of cancer cells, cancer-associated fibroblasts, and macrophages. We applied the models to predict the flux distributions and compare the cells’ metabolic phenotypes in the TME. We also apply graph theoretical methods to analyze the structure and organization of the cells’ metabolic networks. The integration of genome-scale metabolic modeling and graph theory produces new insights into metabolism in CRC and identifies strategies to exploit metabolic crosstalk to inhibit tumor progression.

Organized by: Yangjin Kim (Brown University and Konkuk University), Magdalena Stolarska at University of St. Thomas

Note: this minisymposia has multiple sessions. The other session is: and Part-2.

1. Magda Stolarska University of St. Thomas
   "A mathematical model of active cortical stress generation and its effect on cell movement"
When moving through a confined, fibrous extracellular environment, many cells use an amoeboid mode of cellular motility. In particular, it is well known that cancer cells can undergo a mesenchymal-amoeboid transition under certain conditions. Amoeboid motility is characterized by weak adhesions to the extracellular environment, a rounded morphology, and flow of the actin cortex. A related, yet simpler, mode of cell motility is cellular swimming. In eukaryotic cell swimming, it is known that active deformation of the cortex induces attachment-free movement through a fluid, but much of the details of this process are not well understood. In this talk, I will present a mathematical model that aims to begin investigating how variability in actomyosin activity in the cortex, properties of the cortex-membrane (ERM) attachment proteins, and the mechanical properties of the microenvironment affect cell movement through a fluid. The hybrid model presented models the intra- and extra-cellular fluid as a continuum and treats the membrane and cortex and a discrete system of connected segments and nodes. By using the finite element method to solve the model equations, we are able to analyze how varying various properties of this system affects cellular swimming velocities.
2. Dumitru Trucu University of Dundee
   "Advancements in multiscale modelling for glioblastoma: emergence of 'on-the-fly' non-local isotropic-to-anisotropic transition in cell population transport"
Despite all recent in vivo, in vitro, and in silico advances, the understanding of the genuine biologically multiscale process of solid tumour invasion remains one of the greatest open questions for scientific community. In this talk we present recent mathematical multiscale moving boundary modelling advancements for solid tumour invasion, with special focus on glioblastoma progression. We focus on enhancing the mathematical modelling for key aspects of the dynamic interactions that the migratory cancer cells population and the accompanying matrix degrading enzymes (MDEs) have with the extracellular matrix (ECM) components, and, in particular, with the ECM fibres. These are complex interactions enabled by a complicated series of integrated multiscale systems, which are at least two-scale in nature and share (and contribute to) the same tumour macro-dynamics (i.e., tissue-scale dynamics) but have independent-in-nature micro-dynamics (i.e., cell-scale dynamics), and despite previous modelling progress, these deserve significant renewed research efforts. Specifically, this talk we seek to address: (1) the enhancing effect of the interfacial presence of ECM fibres on the macro-scale tumour boundary movement; (2) a new non–local “go-or-grow” perspective on the motility of cancer cel population; and (3) the emerging “on-the-fly” non–local isotropic – to – anisotropic transition in the diffusive cell population transport. Mathematical formulations for all these aspects are proposed analytically and then explored computationally and discussed in the context of glioblastoma progression.
3. Padmini Rangamani University of California San Diego
   "Modeling collagen fibril degradation as a function of matrix microarchitecture"
Collagenolytic degradation is a process fundamental to tissue remodeling. The microarchitecture of collagen fibril networks changes during development, aging, and disease. Such changes to microarchitecture are often accompanied by changes in matrix degradability. In a matrix, the pore size and fibril characteristics such as length, diameter, number, orientation, and curvature are the major variables that define the microarchitecture. In vitro, collagen matrices of the same concentration but different microarchitectures also vary in degradation rate. How do different microarchitectures affect matrix degradation? To answer this question, we developed a computational model of collagen degradation. We first developed a lattice model that describes collagen degradation at the scale of a single fibril. We then extended this model to investigate the role of microarchitecture using Brownian dynamics simulation of enzymes in a multi-fibril three dimensional matrix to predict its degradability. Our simulations predict that the distribution of enzymes around the fibrils is non-uniform and depends on the microarchitecture of the matrix. This non-uniformity in enzyme distribution can lead to different extents of degradability for matrices of different microarchitectures. Our simulations predict that for the same enzyme concentration and collagen concentration, a matrix with thicker fibrils degrades more than that with thinner fibrils. Our model predictions were tested using in vitro experiments with synthetic collagen gels of different microarchitectures. Experiments showed that indeed degradation of collagen depends on the matrix architecture and fibril thickness. In summary, our study shows that the microarchitecture of the collagen matrix is an important determinant of its degradability.
4. Noe Mercado Warren Alpert Medical School, Brown University
   "Impact of Cytomegalovirus on Glioblastoma progression"
Background: Glioblastoma (GBM) is the most common primary malignant brain tumor and has no effective treatments. Human Cytomegalovirus (HCMV) has been implicated in GBM progression and antiviral drugs like Cidofovir (CDV) have promising activity in GBM. Previously we reported that in our established syngeneic GBM mouse model perinatally infected with murine cytomegalovirus significant reduction in overall survival compared to uninfected controls. Treatment with CDV improved survival in infected mice and inhibited MCMV reactivation as well as tumor angiogenesis. However, the molecular mechanisms of antiviral drug treatment on GBM have not been studied. Results: Here we show that GBM patient-derived glioma stem cells (GSCs) are highly permissive to HCMV infection compared to established GBM lines commonly used in in vitro (U-373). Neuronal cells that are found in the tumor microenvironment also have high permissiveness to infection although viability significantly decreased post infection. Treatment with antiviral drug Brincidofovir (BCV), a lipid prodrug of cidofovir, significantly reduced viral infection but did not directly induce GBM cell killing. When cells were treated with standard of care (SOC) therapy comprising irradiation (6Gy) and temozolomide (TMZ), resistance to cell death was observed in infected GSCs. This phenotype was reversed by treatment with BCV in a dose-dependent manner. These observations suggest that HCMV induces resistance to SOC in GSCs which may promote GBM progression, and this may be a target of antiviral therapy. Proteomic analysis of infected GSCs revealed upregulation of several pro-tumorigenic proteins including Structural Maintenance of Chromosome 4 (SMC4), WD repeat domain 5 (WDR5) and thymocyte selection associated high mobility group (TOX). Interestingly upon antiviral drug treatment these proteins were no longer upregulated and instead several were significantly downregulated after BCV treatment. Conclusions: Together these data suggest that HCMV may promote tumorigenesis in part due to the glioma stem cell niche. After infection these glioma stem cells are more resistant to chemoradiotherapy which can be overcome by antiviral drug (BCV) treatment. These data provide mechanistic evidence for the role of HCMV in GBM and support ongoing research into antiviral drug approaches in the clinic.

Organized by: Hannah Anderson (Moffitt Cancer Center), Kasia Rejniak, Moffitt Cancer Center

1. Hannah Anderson Moffitt Cancer Center
   "Evaluating robustness of an optimized regimen in a virtual murine cohort of bladder cancer"
Virtual cohorts can capture the heterogeneity across patient populations and thus different responses to treatment. In this talk, we develop an ODE model of a combination therapy for mice implanted with bladder cancer. Using a murine data set, we develop a virtual cohort using a framework that consists of 1) structural identifiability, 2) modifying data use, 3) estimating parameters, 4) determining practically identifiable parameters, 5) obtaining parameter distributions, and then 6) simulating the virtual cohort alongside data. Using the parameter set that represents an average mouse from data, we perform optimal control to optimize a regimen for adoptive cell therapy in combination with gemcitabine. Then, we evaluate the robustness of this regimen by determining its efficacy when applied to the virtual murine cohort.
2. Christian Parkinson Michigan State University
   "Optimal control of a reaction-diffusion epidemic model with noncompliance"
We consider an optimal distributed control problem for a reaction-diffusion-based SIR epidemic model with human behavioral effects. We develop a model wherein non-pharmaceutical intervention methods are implemented, but a portion of the population does not comply with them, and this noncompliance affects the spread of the disease. Drawing from social contagion theory, our model allows for the spread of noncompliance parallel to the spread of the disease. Control variables affect the infection rate among the compliant population, the rate of spread of noncompliance, and the rate at which non-compliant individuals return to a compliant state. We prove the existence of global-in-time solutions for fixed controls and study the regularity properties of the resulting control-to-state map. We establish the existence of optimal controls for a fairly general class of objective functions and present a first-order stationary system which is necessary for optimality. Finally, we present simulations with various parameters values to demonstrate the behavior of the model.
3. Xinyue Zhao University of Tennessee Knoxville
   "Optimal control of free boundary models for tumor growth"
In this talk, we will investigate the optimal control of treatment in free boundary PDE models for tumor growth. The optimal control strategy is designed to inhibit tumor growth while minimizing side effects. In order to characterize it, the optimality system is derived, and a necessary condition is obtained. Numerical simulations will be presented to illustrate the theoretical findings and assess the impact of the optimal control strategy on tumor growth dynamics.
4. Luis Maria Lopes da Fonseca University of Florida
   "Surrogate modeling and control of medical digital twins"
The vision of personalized medicine is to identify interventions that maintain or restore a person's health based on their individual biology. Medical digital twins, computational models that integrate a wide range of health-related data about a person and can be dynamically updated, are a key technology that can help guide medical decisions. Such medical digital twin models can be high-dimensional, multi-scale, and stochastic. To be practical for healthcare applications, they often need to be simplified into low-dimensional surrogate models that can be used for the optimal design of interventions. Here, we introduce surrogate modeling algorithms for optimal control applications. As a use case, we focus on agent-based models (ABMs), a common model type in biomedicine for which there are no readily available optimal control algorithms. By deriving surrogate models based on systems of ordinary differential equations, we show how optimal control methods can be employed to compute effective interventions, which can then be lifted back to a given ABM. The relevance of the methods introduced here extends beyond medical digital twins to other complex dynamical systems.

Organized by: Changhan He (University of California, Irvine), Chengyue Wu, University of Texas MD Anderson Cancer Center

1. Lifeng Han Tulane University
   "Calibrate a phenotype-structured population model with cell viability data to study drug resistance in cancer treatment"
We fit a phenotype-structured population mode to cell viability data on a drug used for ovarian cancer, olaparib. This approach reveals the effects of a drug on fitness landscape and the evolution of a population of cancer cells structured with a spectrum of drug resistance. We will show that maximizing variation in plasma drug concentration over a dosing interval could be important in reducing drug resistance. We will discuss the potential use of this model to design drug treatment regimens to improve cancer treatment.
2. Wenjun Zhao Wake Forest University
   "Dynamical GRN inference via optimal transport"
In this talk, we present a framework for inferring gene regulatory networks from time-stamped single-cell gene expression data. Our algorithm first reconstructs the correspondence between single cells across time points through optimal transport, and then infers the underlying dynamical model governing gene expression dynamics, which can be interpreted in terms of differential equations. We will also discuss extensions of this framework to infer context-specific regulatory mechanisms.
3. Qixuan Wang University of California, Riverside
   "Hair follicle cell fate regulations and the effect on the follicle growth control"
In tissues, proper regulations of cell fate decisions are important in maintaining the shape and functions of the tissue. In this talk, I will present our recent research in modeling cell fate regulation mechanisms, using mammalian hair follicles as a model system. Hair follicles are mini skin organs, and they are highly dynamic in the way that they can undergo cyclic growth during the lifespan of the organism. To maintain tissue homeostasis and functions of a hair follicle, the follicle transient amplifying epithelial cells need to make correct decisions among cell division, differentiation and apoptosis, instructed by various signals produced by the follicle itself as well as by the surrounding skin environment. We develop a probabilistic Boolean model based on both literature and published single-cell RNA sequencing data. Using both computational simulations and attractor analysis, we investigate how hair follicle epithelial cells respond to TGF-beta, BMP and TNF, so to make the correct cell fate decision, and how signals cooperatively regulate hair follicle growth dynamics. Next, we develop a hybrid multiscale model on the bottom part of a HF, and use it to investigate how signaling dynamics, cellular kinetics and movement, and gene regulation coordinate to regulate the HF cell fate decisions in the tissue.
4. Axel Almet University of California, Irvine
   "Systems modeling of cellular senescence using single-cell transcriptomics"
Cellular senescence is a stress-induced cell state characterised by irreversible cell cycle arrest and an enhanced secretome where senescent cells secrete an array of inflammatory signals and remodeling factors. An accumulation of senescent cells across biological age is hypothesised to contribute to aging-related declines in tissue. However, there are contexts where senescent cells may be transiently beneficial. To study the impact of cellular senescence requires a systems approach that analyzes both its intrinsic features, driven by changes in gene transcription dynamics, and its extrinsic impacts on the tissue environment driven by cell-cell communication. Single-cell transcriptomics, which profile molecular states with broad gene coverage and cell type coverage, provide an exciting opportunity to generate further insight into the multi-dimensional features of cellular senescence. In this talk, we present our recent work on using single-cell transcriptomics to dissect the multi-dimensional features of cellular senescence. First, we fit stochastic models of gene transcription to ground truth single-cell RNA-sequencing datasets of cellular senescence to dissect how cellular senescence drives intrinsic changes in gene transcription kinetics. We then shift our analysis to the perspective of cell-cell communication, identifying senescence-driven communication modules that can reveal novel senescent cell states with distinct communication patterns and, in turn, show how these communication patterns are associated with distinct transcriptional features. Overall, these analyses reveal new ways that we can integrate sequencing data and mathematical modeling to better understand cellular senescence.

Organized by: Einar Bjarki Gunnarsson (Science Institute, University of Iceland), Maximilian Strobl (Cleveland Clinic, USA, stroblm@ccf.org)

Note: this minisymposia has multiple sessions. The other session is: and Part-2.

1. Einar Bjarki Gunnarsson Science Institute, University of Iceland
   "Decoding drug-induced persistence: Integrating theory with experiments and optimizing dosing protocols"
Drug resistance is a common reason for treatment failure in cancer. While resistance evolution is usually viewed through the Darwinian lens of random mutation and selection, mounting evidence indicates that anti-cancer drugs designed to kill tumor cells can simultaneously induce the adoption of non-genetic drug-persistent cell states. This confounds conventional high-dose treatment strategies aimed at maximal tumor reduction, since high doses can simultaneously accelerate the evolution of resistance. At the same time, the reversible nature of non-genetic mechanisms creates opportunities for delaying resistance evolution through intermittent or adaptive therapy. In this talk, we discuss mathematical modeling of drug-induced persistence and its role in the evolution of stable drug resistance. We also discuss how mathematical modeling can be used to derive optimal drug schedules aimed at maximally delaying tumor recurrence. In doing so, we introduce the central conversation of the mini-symposium involving the integration of mathematical modeling with experimental data, which we believe is crucial to achieve the long-term goal of personalized model-informed optimal drug scheduling.
2. Mattia Corigliano IFOM - The AIRC Institute of Molecular Oncology, Milan, Italy
   "Optimal treatment for drug-induced cancer persisters involves release periods and intermediate drug doses"
Targeted cancer therapies often induce a reversible drug-tolerant state in subpopulations of cells, akin to bacterial persistence. Precise characterization of these 'cancer persisters' is crucial for designing more effective treatment strategies. Our recent work demonstrates that, unlike bacterial systems, the transition to persistence in colorectal cancer cell lines exhibits a distinct dependence on drug presence and concentration. In this talk, I will present a mathematical modeling framework that leverages these findings to explore intermittent treatment protocols aimed at reducing the long-term fitness of the treated population. By adapting a bacterial persistence model to colorectal cancer dynamics, we identify success and failure regions within a clinically accessible parameter space. Strikingly, our analysis suggests that optimal treatment outcomes may be achieved with non-zero recovery periods and drug concentrations lower than those typically administered in clinics. Moreover, incorporating patient drug pharmacokinetics into the model reveals that intermittent dosing strategies, currently explored in clinical trials, can be optimized to potentially rival the efficacy of continuous treatment regimens. These results highlight the power of mathematical modeling in optimizing cancer treatment protocols, offering insights into non-trivial trade-offs that could improve patient outcomes.
3. Irina Kareva Northeastern University
   "Dosing Strategies for Bispecifics with a Bell-Shaped Efficacy Curve: What Looks Like Resistance May Be Corrected Through Schedule Adjustments"
Bispecific T cell engagers (TCEs) can exhibit bell-shaped efficacy curves, where increasing the dose beyond a certain point leads to reduced, not improved, efficacy. This counterintuitive behavior arises when efficacy depends on forming a trimeric complex between drug, tumor target, and T cell receptor, as is the case with teclistamab, a bispecific targeting BCMA and CD3 in multiple myeloma. Using a semi-mechanistic PK/PD model and a virtual patient population, we demonstrate that apparent loss of response may reflect overdosing rather than true resistance. We explore how measurable pre-treatment biomarkers, such as soluble and membrane-bound BCMA, can guide dose optimization and patient stratification. The model supports a shift toward semi-personalized dosing strategies and highlights that lowering the dose may, in some patients, restore efficacy.
4. Tatiana Miti Moffitt Cancer Center & Research Institute
   "ABM studies on the role of the drug-sheltering effects of stroma on the emergence of resistance"
Lung cancer is the leading cause of cancer-related deaths in the U.S., with non-small cell lung cancer (NSCLC) accounting for 84% of cases and a low 5-year survival rate of 6%. About 20% of NSCLC cases involve abnormal activation of receptor tyrosine kinases, which are treated with tyrosine kinase inhibitors (TKIs). Unfortunately, despite the magnificent initial results, TKIs show modest outcomes as tumors evolve, gain resistance, and eventually relapse. Experimental data suggest that drug-resistant tumor subpopulations emerge after TKI treatment through de novo mutations and epigenetic changes. In vitro studies indicate that this adaptation is supported by microenvironmental factors, particularly cancer-associated fibroblasts (CAFs), which protect tumor cells via secreted paracrine factors, hence access to these pro-survival factors depends on the spatial organization of tumors. However, despite the massive body of studies, the exact contribution of stromal protection to the minimal residual disease and ultimate relapse remains unknown. Based on published data from other teams, as well as published and unpublished data from Dr. Marusyk’s lab we use an Agent-Based Model to assess the stromal effects on the remission-relapse dynamics if (1) CAF mediated stromal protection effects only those tumor cells that are located at close proximity to stroma, (2) both total amount and spatial patterns of stroma within tumors are varied and (3) CAF mediated stromal protection reduces the initial drug induced tumor cell elimination yielding the accumulation of epigenetic mutations, thus contributing to the residual disease and preserving intratumor heterogeneity. Our results show that using mathematical models, we can gain a deeper understanding of the ecological mechanisms that lead to NSCLC relapse under TKI treatment and uncover new therapeutic strategies that account for stromal effects and successfully eradicate tumors.

Organized by: Jana Gevertz (The College of New Jersey), Thomas Hillen (University of Alberta), Linh Huynh (Dartmouth College)

Note: this minisymposia has multiple sessions. The other session is: and Part-2.

1. Thomas E. Yankeelov The University of Texas at Austin
   "Integrating mechanism-based and data driven modeling to predict breast cancer response to neoadjuvant chemotherapy"
While neoadjuvant chemotherapy (NAC) is the standard-of-care for treatment of patients with triple-negative breast cancer (TNBC), only about half of patients attain a pathological complete response (pCR). The addition of immunotherapy increases the pCR rate to about two thirds, but also increases toxicity. Thus, to improve patient outcomes, it is essential to develop methods that can predict and optimize patient response early during NAC. Previously, we showed that a biology-based model calibrated to pre- and on-treatment patient specific MRI data from the ARTEMIS trial (NCT02276443) can accurately predict patient response to NAC. This model describes cell movement and drug induced death globally over the tumor and cell proliferation locally on a voxel-by-voxel basis. We explore two approaches to extend this model. First, we train a convolutional neural network to predict the calibrated model parameters using only pre-treatment MRI data as input. Second, we apply k-means clustering to the longitudinal MRI data to segment tumors into distinct regions called habitats. Then, rather than calibrating cell proliferation locally, we calibrate one proliferation per habitat. To evaluate these models, we use the total tumor cellularity after the first course of NAC in a receiver operating characteristic (ROC) curve analysis to predict pCR status at the end of NAC. For the first method, we obtain an area under the ROC curve (AUC) of 0.72 for predicting the eventual response to NAC before initiating NAC. For the second method, we obtain AUC values of 0.79 and 0.77 for the habitat-informed and voxel-by-voxel proliferation rate calibrations, respectively, for 101 patients. Thus, we can make reasonably accurate predictions of pCR before initiating NAC with our CNN approach and attain comparable accuracy to a local calibration using our habitats approach, which requires fewer model parameters.
2. Lena Podina University of Waterloo
   "Universal Physics-Informed Neural Networks and Their Applications"
Differential equations are widely used to model systems such as predator-prey interactions, and the effect of chemotherapy on cancer cells. However, in order to construct these models, assumptions must be made about the behaviour of these systems, which may require significant manual distillation of the literature if the model is large. In this talk, I will discuss Universal Physics-Informed Neural Networks (UPINNs), and show how UPINNs can be used to learn unknown terms in ordinary and partial differential equations from sparse and noisy data. This approach allows one to use machine learning to identify the best way to model a system, rather than relying on prior assumptions. Physics Informed Neural Networks (PINNs) have been very successful in a sparse data regime (reconstructing entire ODE solutions using only a single point or entire PDE solutions with very few measurements of the initial condition). The Universal PINN approach (UPINN) adds a neural network that learns a representation of unknown hidden terms in the differential equation. These hidden term neural networks can then be converted into symbolic equations using symbolic regression techniques like AI Feynman. In our work, we demonstrate strong performance of UPINNs even when provided with very few measurements of noisy data in both the ODE and PDE regime. We apply UPINNs to learning predator-prey interaction in the Lotka-Volterra model, chemotherapy drug action terms in a model of cancer cell growth, and terms in Burgers’ PDE. UPINNs could be instrumental to paving the way to allow machine learning to help applied mathematicians model systems in a more automatic, data-driven way even when observations are sparse.
3. Adam L. MacLean University of Southern California
   "Dynamic rewiring of cell-cell interaction networks in metastatic TMEs to empower checkpoint inhibition"
Tumors grow, evolve, and metastasize as a result of an intricate set of interactions between the numerous cell types that comprise the tumor microenvironment (TME). Many of these interactions remain poorly understood, even in the absence of therapy, and the size of the networks that must be considered necessitates a systems biology approach. Immune checkpoint inhibition combined with entinostat, a histone deacetylase inhibitor, has been shown to promote durable responses in a minority of patients with metastatic, triple negative breast cancer. But the mechanisms of action of entinostat at metastatic sites has not been investigated. We measured the molecular properties of the metastatic TME in high resolution via scRNA-seq, quantifying 39 cell states across six treatment arms. Entinostat treatment led to an increase in stemness in tumor cells and decreased mesenchymal gene expression. To study the wiring of cell-cell interaction networks with and without treatments we inferred small cell circuits that are over-represented in the full networks. Top ranked cell circuits comprised myeloid and T cell subtypes: interactions between which were dramatically affected by combination therapy. Top pathways contributing to these interactions included chemokine, galectin, and ICAM pathways, which we tested via targeting specific ligand-receptor pairs in functional suppression assays, coupled to predictions from mathematical models to simulate therapeutic responses for a given treatment intervention. Beyond the clinical impact of this work, our results offer a framework with which to decompose large, complex TMEs to infer multiscale networks mediating treatment effects and then to infer the tumor response dynamics via mathematical modeling.
4. Venkata Manem Centre de Recherche du CHU de Québec; Université Laval, Canada Université Laval, Canada
   "Beyond the One-Size-Fits-All Paradigm: Leveraging Bioinformatics and AI to Advance Biomarker-Guided Oncology."
In the era of precision oncology, the convergence of high-throughput technologies and artificial intelligence (AI) is transforming how we understand and treat cancer. Traditional 'one-size-fits-all' regimens are being replaced by biomarker-driven strategies that tailor therapies to individual patients. However, the complexity of tumor genomics and microenvironments limits the effectiveness of many biomarkers, underscoring the need for biologically informed approaches. This talk will explore the intersection of bioinformatics and AI in uncovering biological insights for personalized cancer care. The first part will focus on breast cancer, emphasizing the critical role of epithelium-stroma crosstalk utilizing transcriptomics data to uncover the dynamic interactions within the tumor microenvironment. The second part will address the discovery of biomarkers in non-small cell lung cancer patients treated with immunotherapy, leveraging medical imaging data. Together, these efforts demonstrate how AI and bioinformatics bridge the gap between molecular complexity and actionable insights, paving the way for patient-specific decision-making in oncology.

Organized by: Yangjin Kim (Brown University and Konkuk University), Magdalena Stolarska at University of St. Thomas

Note: this minisymposia has multiple sessions. The other session is: and Part-1.

1. Donggu Lee Konkuk University
   "Asthma-mediated control of optic glioma growth via T cell-microglia interactions: Mathematical model"
Optic glioma, a slow-growing tumor, is associated with Neurofibromatosis type 1 (NF1) mutations and increased midkine (MDK) production. To elucidate the role of asthma in regulation of glioma formation, we investigated the role of T cells and the subsequent pathways in regulation of microglia, a key player in tumor microenvironment (TME). While asthma is often linked to chronic inflammation, our mathematical analysis and experimental evidence suggest that it can play a significant role in suppressing proliferation of optic glioma cells via immune reprogramming of T cells and delicate control of signaling network in microglia. The mathematical model unveil the complex interaction between brain tumor and immune cells in the brain. Our results indicate that asthma-induced T cell reprogramming inhibit tumor growth by promoting the release of Decorin, which leads to a chain of suppression of CCR8 and NFkB in microglia and CCL5 production. These findings highlight the potential of leveraging asthma-induced immune regulation as a novel mechanism for glioma suppression and demonstrate the power of mathematical modeling in uncovering complex tumor-immune interactions.
2. Ji Young Yoo University of Texas Health Science Center at Houston
   "Reshaping the Tumor Microenvironment by targeting IGF2-IGF1R signaling: Enhancing Viro-Immunotherapy"
The FDA approval of oncolytic herpes simplex-1 virus (oHSV) therapy for melanoma patients underscores its therapeutic promise as a cancer immunotherapy. However, despite this promise only a small subset of patients respond favorably in the clinic. Oncolytic virotherapies work both through direct oncolysis of infected cancer cells and by inducing of inflammatory response and concurrent activation of anti-tumor immunity through the release of tumor antigens from the lysed cancer cells, a phenomenon referred to as viro-immunotherapy. However, the induction of an immunosuppressive tumor microenvironment (TME) by the tumor both before and shortly after therapy poses the greatest hurdle to lasting efficacy and viro-immunotherapy. Our work centers on understanding the oHSV therapy resistance mechanism and characterizing the impact of viro-immunotherapy to design a better viro-immunotherapy to broaden their applications in the clinic. RNA-Seq analysis of oHSV-infected glioblastoma (GBM) and breast cancer (BC) cells identified Insulin-like growth factor 2 (IGF2) as one of the top 10 secreted proteins following infection. Moreover, IGF2 expression was significantly upregulated in 10 out of 14 recurrent GBM patients treated with oHSV, rQNestin34.5v.2 (71.4%) (p=0.0020) (ClinicalTrials.gov, NCT03152318), highlighting its clinical relevance. IGF2 is upregulated in tumor malignancies and its overexpression is associated with resistance to chemotherapy and radiation therapy, worse prognosis, anti-tumor immune suppression in the TME, and cancer metastasis. In order to mitigate oHSV therapy-induced IGF2 and improve the therapeutic efficacy of oHSV, we designed a novel oHSV construct, oHSV-D11mt, which integrates a secretable modified IGF2R domain 11 into the parental oHSV genome that serves as an IGF2 decoy receptor. The secreted IGF2RD11mt selectively binds to IGF2, effectively blocking oHSV-induced IGF2-IGF1R signaling, which lead to enhanced tumor cell cytotoxicity, reduced oHSV-induced neutrophils/PMN-MDSCs infiltration, reduced secretion of immunosuppressive/proangiogenic cytokines, and increased Cytotoxic T lymphocytes (CTLs) infiltration. These effects resulted in enhanced survival of both GBM and BC brain metastasis (BCBM) tumor-bearing mice, abrogating the resistance conferred by IGF2 secretion. Collectively, our findings suggest that oHSV-induced secreted IGF2 exerts a critical role in resistance to oHSV therapy and our novel viral construct represents a promising therapeutic for enhanced viro-immunotherapy.
3. Alexandra Shyntar University of Alberta
   "Mathematical Modelling of Microtube-Driven Regrowth of Glioma After Local Resection"
In this talk, I will first summarize the results from the paper Weil et al. (2017) “Tumor microtubes convey resistance to surgical lesions and chemotherapy in gliomas.” In this paper, they perform a series of experiments on glioblastoma tumors motivated by the discovery of tumor microtubes (TMs). TMs are thin long protrusions extending from the glioblastoma cell body which help the cancer grow, spread, and facilitate communication between glioma cells. Weil et al. (2017) implanted glioblastoma tumors into mice and tested various treatments (surgery, surgery with targeted therapy, surgery with anti-inflammation treatment, and chemotherapy). They find that TMs help with the faster and denser tumor repopulation in the lesion area. Furthermore, inhibiting the TMs slows down tumor growth significantly. In the second part of the talk, I will show how the experiments outlined in the paper can be modelled with partial differential equations. The effects from the wound healing response and TMs are simplified but are accounted for in the model. The numerical simulations reveal good agreement with the experimental observations and can capture the experimental trends after treatment application. Based on these results, the wound healing mechanisms as well as TM dynamics are key in explaining the experimental observations.
4. Sean Lawler The Warren Alpert Medical School, Brown University
   "Remodeling the Tumor Microenvironment to Facilitate Glioblastoma Therapy"
Glioblastoma(GBM) is the most common malignant brain tumor and is extremely challenging to treat effectively. Standard of care therapy involves surgical resection followed by radiation and alkylating chemotherapy. This results in a median survival of only 15 months, with resistance to therapy rapidly emerging. Immune checkpoint blockade has not yet shown efficacy in GBM. The GBM tumor microenvironment (TME)is complex and is thought to make a major contribution to resistance to both standard of care and immunotherapies. The GBM TME is characterized by infiltration of suppressive myeloid cells, and microglia, the presence of Tregs and lack of CD8+ T cells. In addition, tumor cells interact with neurons, astrocytes and with one another, subverting neurological signaling mechanisms to drive tumor progression and therapeutic resistance mechanisms. Understanding how to effectively modulate the GBM TME by targeting cellular interactions is essential to provide new therapeutic approaches. In this work, I will discuss approaches being developed in my lab to model these interactions, and the effects of immunomodulatory drugs on the TME that may facilitate responses to immune checkpoint blockade and other therapies, for improved outcomes in this devastating tumor type.

Organized by: Raluca EFTIMIE (University of Marie & Louis Pasteur, France), Dumitru TRUCU, University of Dundee, UK

1. Marom Yosef Ariel University
   "Mathematical Models to Improve Bladder Cancer Therapies"
Bladder cancer (BC) represents a significant clinical challenge, affecting 549,000 new patients annually, with over 200,000 deaths per year. Despite initial surgical intervention, approximately 70%, of patients experience tumor recurrence, necessitating additional treatment. The current gold standard, Bacillus Calmette-Guérin (BCG) immunotherapy, demonstrates limited efficacy: only 50% of patients achieve complete response, while 80% experience adverse effects ranging from mild discomfort to severe complications requiring treatment discontinuation. In the first part of my talk, I show the mathematical models to improve BCG therapy and have explored various protocols, including six-week induction therapy and extended maintenance treatments. However, these modifications have shown limited success in improving patient outcomes. Recent models including combining BCG with interleukin-2 (IL-2) or Interferon (IFN) immunomodulator therapy have demonstrated promising results, potentially enabling reduced BCG dosages while maintaining therapeutic efficacy. However, the unpredictable nature of immune responses to this combined treatment has hindered its widespread clinical adoption. I present a significant advance in translating mathematical modeling into clinical practice, enabling more precise and personalized treatment protocols while minimizing adverse effects. The framework's ability to provide stable, analytical solutions for combined immunotherapy treatments offers immediate applications for BC treatment optimization and broader implications for other immunotherapy-based cancer treatments. In the second part of the talk, I explain the mathematical model for optimizing Mitomycin-C (MMC) treatment for BC. Current drug dosing strategies rely on general guidelines without precise quantitative justification. Our model revolutionizes this approach by introducing analysis for drug-tumor interactions. While existing methods cannot predict required drug doses theoretically, our framework enables precise calculation of MMC concentrations needed for tumor elimination based on specific tumor characteristics. This innovation transforms MMC dosing from an empirical process to a mathematically guided procedure. These innovations enable a shift from standardized protocols to personalized treatment strategies. Unlike current approaches that modify treatments through trial and error, our models provide a theoretical foundation for optimizing treatment protocols based on individual tumor characteristics, potentially improving outcomes while minimizing unnecessary drug exposure. Keywords: bladder cancer, immunotherapy, BCG treatment, Mitomycin-C, mathematical modeling, tumor-immune interactions, treatment optimization, personalized medicine
2. Haralampos Hatzikirou: Khalifa University
   "From cell patterns in biopsies to clinical predictions"
Understanding the dynamic role of immune cells in cancer progression is essential for predicting disease outcomes and developing targeted therapies. This talk delves into the transition from cellular patterns observed in biopsies to clinical predictions, with a particular focus on the role of macrophages in the tumor microenvironment (TME). Drawing on advanced computational models and experimental data, we explore how macrophage phenotypic changes, particularly their transition from pro-inflammatory to pro-tumorigenic (M2) states, influence tumor progression and recurrence. By examining macrophage-fibroblast interactions in kidney transplant biopsies, we uncover key insights into macrophage behavior that can be translated to cancer research, particularly in gliomas and other solid tumors. The talk will discuss how macrophage dynamics, observed through transcriptomic profiling and tissue-specific modeling, can be integrated into predictive models of tumor growth and recurrence. This framework has the potential to improve clinical decision-making by enabling earlier interventions and more accurate predictions of treatment outcomes, highlighting the importance of macrophage-driven processes in cancer biology.
3. Ali Daher University of Marie & Louis Pasteur
   "Integrating High-Throughput Genomic Data with Biologically-Informed Models of Spatiotemporal Dynamics of Skin Lesions: A Computational Parameter Extraction Pipeline"
Skin wound healing typically progresses through three chronologically overlapping stages: inflammation, proliferation, and remodelling [1]. However, in some cases, prolonged or excessive inflammatory and proliferative phases can lead to abnormal wound healing; one such example is keloid scarring. Keloids are benign fibroproliferative tumours characterized by excessive collagen production and extracellular matrix (ECM) deposition by fibroblasts following dermal injury or irritation. Known for their aggressive nature and pathological spread beyond the original wound boundaries, keloids often result in disfiguring scars, have high recurrence rates, and show poor response to current treatment approaches. The rapid advancement of single-cell RNA sequencing (scRNA-seq) techniques has enabled detailed characterization of the cellular landscape, heterogeneity, and intercellular interactions in skin samples from both normal and abnormal wound healing. In the case of keloids, studies have revealed high immune cell infiltration, with a positive correlation between immune cell abundance and keloid recurrence [2]. Additional findings have identified close interactions between immune cells and fibroblasts, whereby immune cells release cytokines and growth factors that drive ECM production and fibroblast proliferation, further fuelling keloid progression [2]. As such, keloids are increasingly regarded as an inflammatory skin disease [2]. Given the limited success of current treatments in resolving or preventing keloid formation and recurrence and the growing evidence of the inflammatory component of keloids, immunotherapy has emerged as a promising novel treatment avenue [3-5], particularly by targeting the communication pathways between immune cells and fibroblasts [2,6]. One prominent communication pathway is mediated by TGF-β, a key effector cytokine secreted by inflammatory cells that promotes fibrotic responses in fibroblasts. Several investigational agents targeting the TGF-β pathway are currently underway in clinical trials for fibrotic and cancer-related diseases [2,7]. However, our current understanding of the immunological underpinnings of keloid pathogenesis remains neither specific nor comprehensive, limiting the effective development of targeted immunotherapies. For instance, persistent inhibition of TGF-β can suppress fibrosis but may also eliminate its anti-inflammatory functions, potentially exacerbating inflammation [8]. The integration of high-throughput genomics technologies, such as scRNA-seq and spatial transcriptomics, has advanced our understanding of the spatial cellular architecture and communication networks in skin wounds, including keloids. Nevertheless, the data generated from these high-resolution technologies alone are insufficient to fully elucidate the inflammatory origins and progression of keloids or to reliably identify optimal immunotherapeutic strategies. In this context, mathematical and computational models, especially spatiotemporal ones, provide powerful complementary tools. They enable the integration and interpretation of experimental data, facilitating in-silico experimentation and hypothesis testing of biological mechanisms underlying normal and abnormal wound healing. These models can simulate the spatiotemporal dynamics of wound healing and keloid progression, offering insights not readily obtainable through conventional experiments. Additionally, they serve as safe and cost-effective testbeds for evaluating immunotherapeutic interventions before clinical application. To this end, we first develop a biologically grounded model capturing the interactions between immune cells (primarily macrophages) and fibroblasts during wound healing. This is achieved through the construction of both particle-based and continuum reaction-diffusion models that describe the production, secretion, diffusion, and uptake of growth factors and ligands mediating these interactions. We then analyse high-throughput genomics data from skin samples of both normal and abnormal wound healing, leveraging matched scRNA-seq and spatial transcriptomics (Visium) data. Through spatial deconvolution, we infer cell type densities across the tissue domain, and we perform intercellular communication analyses to estimate the strength of interactions mediated by specific ligands. Subsequently, we develop a parameter learning framework that combines approximate Bayesian computation with machine learning and backpropagation techniques to infer the parameters of the reaction-diffusion model from the experimental data. By integrating a biologically informed mathematical framework with genomics-derived data, we ensure that our model is both mechanistically and data-driven—an essential requirement for clinical and research relevance. References 1. Liu, Z., et al. “538 Spatiotemporal Single-Cell Roadmap of Human Skin Wound Healing.” Journal of Investigative Dermatology, vol. 144, no. 12, Dec. 2024, p. S321. DOI.org, https://doi.org/10.1016/j.jid.2024.10.551. 2. Zhang, Xiya, et al. “The Communication from Immune Cells to the Fibroblasts in Keloids: Implications for Immunotherapy.” International Journal of Molecular Sciences, vol. 24, no. 20, Oct. 2023, p. 15475. DOI.org, https://doi.org/10.3390/ijms242015475. 3. Zhang, Tao, et al. “Current Potential Therapeutic Strategies Targeting the TGF-β/Smad Signaling Pathway to Attenuate Keloid and Hypertrophic Scar Formation.” Biomedicine & Pharmacotherapy, vol. 129, Sept. 2020, p. 110287. DOI.org, https://doi.org/10.1016/j.biopha.2020.110287. 4. Ekstein, Samuel F., et al. “Keloids: A Review of Therapeutic Management.” International Journal of Dermatology, vol. 60, no. 6, June 2021, pp. 661–71. DOI.org, https://doi.org/10.1111/ijd.15159. 5. Huang, Chenyu, et al. “Managing Keloid Scars: From Radiation Therapy to Actual and Potential Drug Deliveries.” International Wound Journal, vol. 16, no. 3, June 2019, pp. 852–59. DOI.org, https://doi.org/10.1111/iwj.13104. 6. Shan, Mengjie, and Youbin Wang. “Viewing Keloids within the Immune Microenvironment.” American Journal of Translational Research, vol. 14, no. 2, Feb. 2022, pp. 718–27. PubMed Central, https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8902558/. 7. Moss, Marcia L., and Dmitry Minond. “Recent Advances in ADAM17 Research: A Promising Target for Cancer and Inflammation.” Mediators of Inflammation, vol. 2017, 2017, pp. 1–21. DOI.org, https://doi.org/10.1155/2017/9673537. 8. Teicher, Beverly A. “TGFβ-Directed Therapeutics: 2020.” Pharmacology & Therapeutics, vol. 217, Jan. 2021, p. 107666. DOI.org, https://doi.org/10.1016/j.pharmthera.2020.107666.
4. Donggu Lee(*,1), Sunju Oh(2), Sean Lawler(3), and Yangjin Kim (1,3) (1) Konkuk University, (2) Konkuk University, (3) Brown University
   "Bistable dynamics of TAN-NK cells in tumor growth and control of radiotherapy-induced neutropenia in lung cancer treatment"
Neutrophils play a crucial role in the innate immune response as a first line of defense in many diseases, including cancer. Tumor-associated neutrophils (TANs) can either promote or inhibit tumor growth in various steps of cancer progression via mutual interactions with cancer cells in a complex tumor microenvironment (TME). In this study, we developed and analyzed mathematical models to investigate the role of natural killer cells (NK cells) and the dynamic transition between N1 and N2 TAN phenotypes in killing cancer cells through key signaling networks and how adjuvant therapy with radiation can be used in combination to increase anti-tumor efficacy. We examined the complex immune-tumor dynamics among N1/N2 TANs, NK cells, and tumor cells, communicating through key extracellular mediators (Transforming growth factor (TGF-beta), Interferon gamma (IFN-gamma)) and intracellular regulation in the apoptosis signaling network. We developed several tumor prevention strategies to eradicate tumors, including combination (IFN-gamma, exogenous NK, TGF-beta inhibitor) therapy and optimally-controlled ionizing radiation in a complex TME. Using this model, we investigated the fundamental mechanism of radiation-induced changes in the TME and the impact of internal and external immune composition on the tumor cell fate and their response to different treatment schedules.

Organized by: Einar Bjarki Gunnarsson (Science Institute, University of Iceland), Maximilian Strobl (Cleveland Clinic, USA, stroblm@ccf.org)

Note: this minisymposia has multiple sessions. The other session is: and Part-1.

1. Jana Gevertz The College of New Jersey
   "Mitigating non-genetic resistance to checkpoint inhibition based on multiple states of exhaustion"
Despite the revolutionary impact of immune checkpoint inhibition (ICI) on cancer therapy, for most indications the majority of patients do not sustain a durable clinical benefit. In this work, we explore the theoretical consequences of the existence of multiple states of immune cell exhaustion on response to ICI therapy. In particular, we consider the emerging understanding that T cells can exist in various states: fully functioning cytotoxic cells, reversibly exhausted cells that are minimally cytotoxic but targetable by ICIs, and terminally exhausted cells that are cytotoxic yet not targetable by ICIs. Under the assumption that tumor-induced inflammation triggers the transition between these T cell phenotypes, we developed a conceptual mathematical model of tumor progression subject to treatment with an ICI that accounts for multi-stage immune cell exhaustion. Simulations of a ‘baseline patient’ without intrinsic resistance to ICI reveal that treatment response (complete responder versus non-responder with non-genetic resistance) sensitively depends on both the dose and frequency of drug administration. A virtual population analysis uncovered that while the standard high-dose, low-frequency protocol is indeed an effective strategy for our baseline patient, it fails a significant fraction of the population. Conversely, a metronomic-like strategy that distributes a fixed amount of drug over many doses given close together is predicted to be effective across the largest proportion of the virtual population. Taken together, our theoretical analyses demonstrate the potential of mitigating resistance to checkpoint inhibitors via dose modulation, and also suggest avenues for selecting combination drug partners.
2. Raymond Ng University of Pennsylvania
   "The role of tumor gene expression variability in evading CD8+ T cells"
Melanoma cells escape CD8+ T cell killing during tumor progression and development of immunotherapy resistance. While genetic alterations affecting antigen presentation and interferon response pathways are well-established mechanisms of immune escape, melanoma cells display substantial gene expression heterogeneity even prior to acquiring these genetic changes, potentially enabling some cells to survive T cell attack. Here, we investigate how this pre-existing gene expression heterogeneity facilitates melanoma cell evasion of T cell recognition and destruction. Using a model system of ovalbumin-expressing melanoma cells cocultured with OT-I CD8+ T cells, we demonstrate that a subset of melanoma cells consistently survives both acute and prolonged T cell selection. By integrating DNA barcoding and single-cell RNA sequencing with computational approaches, we developed a robust framework to identify survivor versus non-survivor cell lineages. Our bootstrap simulation framework generated empirical null distributions of lineage selection frequencies, enabling robust statistical inference to distinguish “survivor” from “non-survivor” lineages with defined confidence levels. These 'survivor' lineages exhibited elevated expression of oxidative stress response and ferroptosis protection pathways, coupled with reduced expression of epithelial-to-mesenchymal transition (EMT) markers. Through long-term coculture experiments, we generated and characterized T cell-resistant melanoma populations, revealing consistent upregulation of interferon-gamma response pathways while maintaining suppressed EMT-like signatures. Our findings uncover previously unrecognized gene expression programs that enable melanoma immune evasion and suggest potential therapeutic vulnerabilities in pathways controlling oxidative stress responses and cellular plasticity.
3. Chenyu Wu University of Minnesota
   "A statistical framework for detecting therapy-induced resistance from drug screens"
Resistance to therapy remains a significant challenge in cancer treatment, often due to the presence of a stem-like cell population that drives tumor recurrence post-treatment. Moreover, many anticancer therapies induce plasticity, converting initially drug-sensitive cells to a more resistant state, e.g. through epigenetic processes and de-differentiation programs. Understanding the balance between therapeutic anti-tumor effects and induced resistance is critical for identifying treatment strategies. In this study, we introduce a robust statistical framework, based on multi-type branching process models of the evolutionary dynamics of tumor cell populations, to detect and quantify therapy-induced resistance phenomena from high throughput drug screening data. Through comprehensive in silico experiments, we show the efficacy of our framework in estimating parameters governing population dynamics and drug responses in a heterogeneous tumor population where cell state transitions are influenced by the drug. Finally, using recent in vitro data from multiple sources, we demonstrate that our framework is effective for analyzing real-world data and generating meaningful predictions.
4. Michael Cotner The University of Texas at Austin
   "Tracking Resistance to Targeted Therapy in TNBC with Cell Barcodes"
Triple negative breast cancer (TBNC) is marked by fewer standard-of-care treatment options and poorer treatment outcomes than other breast cancer subtypes, with approximately 40% of TNBC patients developing treatment resistance. High intratumoral heterogeneity, a characteristic of TNBC, leads to its difficulty in treatment and rapid acquisition of resistance. To investigate how this heterogeneity influences treatment response and resistance in TNBC, we employ ClonMapper, our DNA barcoding technology that utilizes integrated and heritable unique DNA barcodes, to track clonal cell populations across treatment. ClonMapper barcodes are identifiable in scRNA-seq, which enables tracking of clonal subpopulations and their transcriptomic diversity before and after treatment. We use ClonMapper to follow barcoded heterogenous tumor cell populations through their response to treatment with three clinically-relevant targeted inhibitor chemotherapies, revealing the diverse transcriptomic trajectories taken by different cell subpopulations and how these diverse responses arise from a heterogenous transcriptomic landscape prior to treatment.

Organized by: Tyler Simmons (Therapy Modeling and Design Center, University of Minnesota), John Metzcar and Sarah Anderson: Therapy Modeling and Design Center, University of Minnesota

1. Gabriel Côté Sainte-Justine University Hospital Azrieli Research Centre / Université de Montréal
   "The role of chronobiology on immunotherapies to prevent neutrophil infiltration into the tumour microenvironment in lung cancer"
BACKGROUND: Oscillations, particularly circadian rhythms, are ubiquitous in physiology. A sound understanding of these phenomena may have important implications for the administration of treatments targeting oscillatory behavior. For example, neutrophils, the most abundant immune cells, display circadian oscillating properties under the control of CXCR2 and CXCR4 receptors. In lung cancer, neutrophil infiltration in the tumour promotes metastases. CXCR2 inhibitors were suggested to reduce such events. However, experiments in mice showed that their administration must align with circadian rhythms; when timed improperly these inhibitors were found to have little to no effect. Even worse, improper timing could result in the dangerous depletion of neutrophils, leading to worst outcomes. Thus, there is a need for rationalizing CXCR2 inhibitor treatment schedules. METHODS: We developed a mathematical model of neutrophil dynamics, incorporating CXCR2 regulation, and added a PK/PD model of AZD5069, a CXCR2 inhibitor. We adjusted our parameters to murine data and performed global sensitivity analyses to determine main regulation mechanisms. We then simulated therapeutic responses in virtual cohorts to optimize treatment regimens. RESULTS: Our results highlight key circadian mechanisms regulating circulating neutrophil counts. Further, our virtual clinical trial predicted that neutrophil oscillations are determinant for establishing effective yet non-toxic CXCR2 inhibitor treatment schedules. IMPACT: This study underlines the importance of chronobiology to drug and immune responses. Our work may be extended to investigate immunity in shift workers, jet-lagged travelers, and individuals with circadian rhythm sleep disorders.
2. Jason T. George, MD, PhD Texas A&M University
   "Stochastic modeling of immunomodulation in the tumor-immune microenvironment"
The advent of T cell-based immunotherapy has ushered in a new age of cancer treatment. Cancer immunotherapy – despite durable efficacy in several disease contexts – is still limited in many disease subtypes, often resulting from unfavorable microenvironmental features and subsequent cancer immune-specific adaptation and ultimate evasion. Recent modeling and empirical directions have thus focused on enhancing immunotherapy’s existing anti-tumor effects and activating the immune system in cases that currently lack any therapeutic response. This talk will discuss our recent efforts at understanding cancer immune evasion and our model’s predicted role of the microenvironment on escape dynamics. I will first discuss our group’s development of a stochastic model of ‘variable evasion’ with implications for antigen targeting. Lastly, I will describe how immunomodulation of tumor-specific T cells can impact cancer escape dynamics, which we then use to study clinically observed cancer recurrence times in breast and bladder cancer.
3. Riley Manning University of Minnesota
   "Agent-based modeling of glioblastoma immunotherapy strategies"
Glioblastoma is an aggressive, highly infiltrative malignant brain tumor with minimal treatment options for patients. Integrated genomic analysis of patient tumors enabled the classification of three molecular subtypes of glioblastoma: proneural, classical, and mesenchymal. Despite distinct alterations in the expression of migration and immune activation-related genes these subtypes are all treated with the same standard of care clinically. Clinical trials investigating T cell based-immunotherapies have had limited success, with many patients quickly developing resistance to treatment. In this work, we use a three-dimensional agent based model of glioblastoma to model the progression of two subtypes: proneural and mesenchymal. Mesenchymal tumors have faster single cell migration speeds and increased infiltration of several immune cell types, including cytotoxic T cells. In contrast, proneural tumors have slower cell migration speeds and are immunologically cold. We model migration, proliferation, and T cell-cancer cell interactions at the single cell level. Cytotoxic T cells deliver sublethal hits to cancer cells, ultimately leading to cancer cell death as damage accumulates. We simulated anti-migratory and T cell-based immunotherapies to identify subtype-specific treatment strategies. We observed differential efficacy between the two tumor subtypes, highlighting the need to account for patient subtype in glioblastoma therapy development. Cytotoxic T cells struggled to eliminate diffuse tumors and escaping tumor cells at the periphery, even at high effector to target ratios. Simulated treatment efficacy was improved with the addition of cancer cell-targeting anti-migratory therapy. This research enhances our understanding of the mechanisms driving therapy failure in glioblastoma and provides a strategy for predicting effective future treatments.
4. Katherine Owens Fred Hutchinson Cancer Center, Seattle, WA
   "Spatiotemporal dynamics of tumor - CAR T-cell interaction following local administration in solid cancers"
The success of chimeric antigen receptor (CAR) T-cell therapy in treating hematologic malignancies has generated widespread interest in translating this technology to solid cancers. However, issues like tumor infiltration, the immunosuppressive tumor microenvironment, and tumor heterogeneity limit its efficacy in the solid tumor setting. Recent experimental and clinical studies propose local administration directly into the tumor or at the tumor site to increase CAR T-cell infiltration and improve treatment outcomes. Characteristics of the types of solid tumors that may be the most receptive to this treatment approach remain unclear. In this work, we develop a simplified spatiotemporal model for CAR T-cell treatment of solid tumors, and use numerical simulations to compare the effect of introducing CAR T cells via intratumoral injection versus intracavitary administration in diverse cancer types. We demonstrate that the model can reproduce tumor and CAR T-cell data from small imaging studies of local administration of CAR T cells in mouse models. Our results suggest that locally administered CAR T cells will be most successful against slowly proliferating, highly diffusive tumors. In our simulations, assuming equal detectable tumor diameters at the time of treatment, low average tumor cell density is a better predictor of treatment success than total tumor burden or volume doubling time. These findings affirm the clinical observation that CAR T cells will not perform equally across different types of solid tumors, and suggest that measuring tumor density may be helpful when considering the feasibility of CAR T-cell therapy and planning dosages for a particular patient. We additionally find that local delivery of CAR T cells can result in deep tumor responses, provided that the initial CAR T-cell dose does not contain a significant fraction of exhausted cells.

Organized by: Guillermo Lorenzo (University of A Coruna (Spain)), Chengyue Wu (The University of Texas MD Anderson Cancer Center, US), Ernesto A. B. F. Lima (The University of Texas at Austin, US)

Note: this minisymposia has multiple sessions. The other session is: and Part-1.

1. Heber L. Rocha Indiana University
   "Agent-Based Modeling of Cancer Drug Response with PKPD Calibration Challenges and Personalized Modeling"
In this talk, I will present an agent-based model developed in PhysiCell that integrates transmembrane diffusion and pharmacokineticpharmacodynamic (PKPD) processes to simulate drug responses in cancer cell cultures. While the model is designed to handle a range of therapeutic scenarios, I will focus on its application to AU565 breast cancer cells, using a previously published dataset of in vitro area cultures. Before calibration, we perform a local sensitivity analysis to identify key parameters influencing cell behavior and to reduce dimensionality in the inference process. We then apply Approximate Bayesian Computation (ABC) to calibrate the model using experimental data, highlighting the challenges posed by limited temporal resolution and spatial variability in the measurements. I will discuss how these limitations lead to parameter non-identifiability, and share insights into how experimental design and model assumptions interact to shape the reliability of inference. This work not only demonstrates the potential of integrating ABMs with data-driven calibration, but also underscores the need for careful consideration of data quality when advancing toward more personalized modeling frameworks.
2. Marianna Cerasuolo University of Sussex
   "Mathematical and Statistical Insights into Gut Microbiota–Phytocannabinoid–Diet Interplay in Prostate Cancer Progression in Mice"
This study presents a data-driven framework for modelling host-microbiome-tumour interactions in obesity-associated prostate cancer (PCa). We quantitatively investigated PCa progression in the TRAMP mouse model, focusing on the interplay between dietary fat intake, phytocannabinoid therapy, and gut microbiota composition. Mice on regular or high-fat diets (HFD) were treated with enzalutamide, cannabidiol and cannabigerol, alone or in combination. While combination therapy proved most effective across both dietary groups, a high-fat diet alone was associated with tumour acceleration and significant microbiome dysbiosis. To get further insight into the underlying dynamics, we used statistical analyses to characterise microbial community structure and its modulation by treatment. Using Granger causality and generalised linear models, we inferred predictive relationships among microbial taxa over time. These were complemented by support vector regression and network-theoretic measures to characterise microbial ecosystems under perturbation. Statistical modelling revealed that HFD promotes a more clustered, less modular microbiome with altered predictive relationships between bacterial taxa. Under combination therapy, these patterns were partially reversed, including restoration of Bacteroidetes abundance and damping of pro-tumour microbial signals. Building upon these findings, we developed a dynamical framework to simulate the co-evolution of microbial communities and tumour burden under varied interventions. This model is based on established ecological principles to capture the temporal dynamics of microbial interactions within the gut environment. The framework simulates microbiome-mediated tumour progression and therapeutic responses by integrating host, microbial, and therapeutic variables. Our results support using microbiome-informed digital twins in obesity-associated PCa, highlighting the therapeutic value of integrated, data-driven modelling.
3. Guillermo Lorenzo University of A Coruna
   "Patient-specific forecasting of prostate cancer progression to higher-risk disease during active surveillance"
Prostate cancer (PCa) usually exhibits low or intermediate risk at diagnosis, for which active surveillance (AS) is an established clinical option. Patients in AS are monitored via serum Prostate Specific Antigen (PSA), multiparametric magnetic resonance imaging (mpMRI), and biopsies. If these exams indicate tumor progression to higher-risk disease, curative treatment is typically recommended (e.g., surgery, radiotherapy). Hence, AS combats overtreatment of indolent PCa, thereby avoiding unnecessary treatment that can induce side effects reducing quality of life (e.g., incontinence, impotence) but without prolonging longevity. However, monitoring protocols for AS rely on an observational and population-based approach that does not account for the heterogeneous nature of PCa dynamics and cannot provide an early identification of progressing patients. To address these two critical limitations, we propose using personalized predictions of tumor progression based on biomechanistic features that describe the heterogeneous disease dynamics for each patient (e.g., tumor cell density, proliferation activity). To this end, we first calculate these patient-specific features from MRI-informed, organ-scale predictions of a biomechanistic model of prostate cancer growth. Then, a generalized logistic classifier is leveraged to map these features to risk groups. Since our PCa growth predictions are spatiotemporally-resolved, we can calculate the biomechanistic features describing PCa dynamics and the tumor risk over time, thus enabling the calculation of time to progression for each patient. Here, we present a preliminary study of our approach in which we demonstrate the accuracy of our personalized growth and progression predictions in a small patient cohort. Although further improvement and testing in larger cohorts are required, we believe that our predictive technology could be leveraged to inform clinical decision-making and personalize AS protocols for PCa patients.
4. David A. Hormuth, II The University of Texas at Austin Texas
   "Image-based habitat dynamics in patients with head and neck cancer undergoing radiotherapy"
Intratumoral hypoxia in head and neck cancer plays a crucial role in radiotherapy (RT) response. Accurate prediction of the extent of hypoxia could enable personalized RT planning to target resistant lesions. Multiparameteric MRI (mpMRI) methods have been developed that are sensitive to the underlying tumor biology and microenvironment enabling longitudinal characterization of the dynamics of tumor heterogeneity. Integration of mpMRI methods with biology-based mathematical modeling could enable the prediction of treatment outcomes. We developed a RT response model informed from each patient’s (N = 20) own imaging data to forecast their response at week 4 of RT for both primary and nodal tumors in human papillomavirus-associated oropharyngeal head and neck cancer. Dynamic contrast-enhanced MRI and oxygen-enhanced MRI was collected before and during the delivery of RT. These data were then analyzed to derive physiological parameters describing hypoxia status, perfusion, and cellularity allowing for clustering tumor regions into four distinct habitats. The time course of each habitat’s volume was then tracked to assess tumor dynamics. A four-compartment mathematical model was implemented to describe tumor habitat changes and RT response. Model parameters were optimized using cross-validation and tested on unseen data. Predictions for primary tumor volumes showed strong correlation (Pearson correlation coefficient ; PCC > 0.85) and agreement (concordance correlation coefficient; CCC > 0.78) with measured volumes, with an error of only 5.3% in hypoxic volume at week 4. Predictions for nodal tumors exhibited higher error (16.7%), with moderate correlation (PCC > 0.80) and agreement (CCC > 0.66). By leveraging MRI-derived habitat information, the model provides accurate, patient-specific forecasts of tumor response. These findings support the potential of MRI-based modeling in guiding personalized RT, helping to refine treatment strategies for head and neck cancer.

Organized by: Jana Gevertz (The College of New Jersey), Thomas Hillen (University of Alberta), Linh Huynh (Dartmouth College)

Note: this minisymposia has multiple sessions. The other session is: and Part-1.

1. John Metzcar University of Minnesota
   "Evaluation of mechanistic and machine learning modeling approaches for glioblastoma recurrence prediction using white blood cell dynamics"
Glioblastoma (GBM) is the most aggressive primary brain tumor, with median recurrence times of approximately 9–11 months following surgery, despite intensive standard-of-care interventions. Early detection of recurrence is crucial for timely enrollment in clinical trials, potentially improving patient outcomes. The significant impact of GBM and its associated therapies on the immune system suggests clinically obtained white blood cell (WBC) counts with differential as possible biomarkers for recurrence prediction. We explore how mechanistic ODE modeling, capturing tumor-immune interactions and treatment impacts, compares with data-driven techniques (GPR and CPH) in predicting GBM recurrence. We apply methods individually and in hybrid combinations to patient-specific WBC trajectories spanning the perioperative period through recurrence. This comparative analysis evaluates predictive accuracy, interpretability, and clinical relevance across methodologies. Our aim is to share preliminary insights from applying multiple modeling strategies to a common clinical problem. By evaluating how each technique performs in the context of GBM recurrence, we hope to better understand their respective advantages and limitations. This work serves as a step toward assessing whether integrating mechanistic with data-driven models enables improved recurrence prediction through a clinically determined, dynamic biomarker.
2. Paul Macklin Indiana University
   "Integrating high-throughput exploration and learning with agent-based models of cancer"
Agent-based models (ABMs) simulate individual cells as they move and interact in a virtualized tissue microenvironment (TME). In the context of cancer, ABMs are increasingly being used to model interactions of malignant cells with stromal cells and the immune system. Even before determining the parameters of an ABM, a key modeling step is to determine the “rules” of the cell agents: how stimuli in the TME (e.g., diffusible factors and their gradients, mechanical signals, contact with other cell types) model each cell agent’s behaviors. To date, parameter identification techniques are applied to pre-configured ABMs that are generally written by hand, hampering iterative efforts to learn the rules and parameters of cancer, fibroblast, and immune cell agents–whether by human implementation or artificial intelligence techniques. In this talk, we describe two advances to help attack this problem: first, we describe how large-scale model exploration on high performance computing (HPC) resources can allow us to broadly pre-explore parameter spaces, giving new insights on not just parameter sensitivity, but also giving new insights on patient-to-patient variation (even with identical model parameters) and the (un)likelihood of finding predictive biomarkers based solely upon patient data at a single time point. Second, we demonstrate a new modeling grammar that allows us to easily create alternative models without need for hand-writing code. Taken together, these advances open the automated exploration of large model spaces on HPC resources, including machine learning approaches. We discuss possible integrations of ABMs with machine learning techniques in future models that combine human and artificial intelligence.
3. Kit Gallagher University of Oxford, Moffitt Cancer Center
   "Predicting Treatment Outcomes from Adaptive Therapy — A New Mathematical Biomarker"
Adaptive Therapy dynamically adjusts drug treatment to control, rather than minimize, the tumor burden of metastatic cancer, thus suppressing the growth of treatment-resistant cell populations and delaying patient relapse. Promising clinical results in prostate cancer indicate the potential of adaptive treatment protocols, but demonstrate broad heterogeneity in patient response. This naturally leads to the question: why does this heterogeneity occur, and is a ‘one-size-fits-all' protocol best for patients across this spectrum of responses? Using deep reinforcement learning, we obtain personalized and clinically-feasible treatment protocols based on individual patient dynamics, and present a framework to generate these treatment schedules based on the patient's response to the first treatment cycle. From a Lotka–Volterra tumor model, we also obtain a predictive expression for the expected benefit from Adaptive Therapy and propose new mathematical biomarkers that can identify the best responders from a clinical dataset after only the first treatment cycle. Overall, the proposed strategies offer personalized treatment schedules that consistently outperform clinical standard-of-care protocols.

Organized by: Mohit Kumar Jolly (Indian Institute of Science), Sarthak Sahoo (Indian Institute of Science)

Note: this minisymposia has multiple sessions. The other session is: and Part-1.

1. David P Cook Ottawa Hospital Research Institute
   "Phenotypic constraints in ovarian cancer - a new perspective on targeted therapy"
High-grade serous ovarian cancer (HGSC) remains the most lethal gynecological malignancy, with a five-year survival rate below 50%. Despite the adoption of PARP inhibitors for patients with BRCA1/2 mutations (20% of cases), clinical management has remained unchanged for decades. The complex genetic landscape of HGSC has not revealed opportunities for effective targeted therapies as seen in other cancer types. To address this challenge, we conducted a meta-analysis of single-cell RNA sequencing data from 471 tumor samples, coupled with spatial transcriptomics using the 10x Genomics Xenium platform. Our analysis revealed three recurrent malignant epithelial phenotypes ('epitypes') that mirror fallopian tube lineages: SecA, SecB, and Cil. These epitypes emerge through non-genetic plasticity and serve distinct functions in disease progression—SecA cells showing high proliferation while SecB cells predominate in metastatic sites and post-chemotherapy samples. Each exhibits unique regulatory patterns and microenvironment interactions. Our findings suggest that developmental regulatory networks constrain malignant phenotypes, creating opportunities for phenotype-targeted therapeutics independent of genetic alterations. Targeting cellular plasticity could restrict tumours' adaptive capabilities, potentially enhancing treatment response and immune recognition in this challenging cancer.
2. Jill Gallaher Moffitt Cancer Center
   "Dynamic evolvability during tumor growth and treatment"
Drug resistance is an ongoing problem for maintaining a treatment response in advanced cancers, which are often more heterogeneous and evolvable. There are benefits for cells to be evolvable, e.g. to easily respond to large shifts in the microenvironment with large heritable shifts in traits, like allowing metastases to survive a new environment and thrive even during treatment. However, evolvability may also be a detriment. With too much deviation from the parental phenotype, cells lose important functions necessary to survive. So, is there an optimal rate of evolvability for tumors to grow and survive treatment that can be exploited therapeutically? We use an off-lattice agent-based model to investigate how the rate of change through proliferation-resistance phenotype space affects tumor growth and response to treatment. During growth, proliferation is selected for, but more evolvability leads to more heterogeneity and faster recurrence under treatment. When evolvability can evolve without constraints, faster evolvability changes will lead to faster recurrence. When evolvability is costly, tumor survival depends on the rate and jump size of heritable changes to transiently lose proliferation fitness selected for during growth and gain resistance for survival. We consider how to design treatment strategies based on a tumor’s evolvability dynamics.
3. Cordelia McGehee Mayo Clinic
   " Chemotherapy dosing as a driver of population evolution in models of intra-tumoral cell-cell competition in cancer"
Despite ongoing therapeutic advances in the treatment of cancer, many advanced solid tumors recur after initial therapy. Minimizing the emergence of drug resistance is a central problem in cancer pharmacology. Dose and dose schedule of chemotherapy administration has traditionally followed the maximum tolerated dose principle which aims to quickly eradicate the tumor while minimizing drug toxicity for the patient. In a clonal drug-sensitive cell population, using the highest dose of drug and achieving maximum tumor killing is a logical strategy. However, when a pre-existing drug-resistant cell population resides within a cancer cell population, the rapid elimination of drug sensitive cells has been hypothesized to lead to proliferation of the resistant cell population. In such cases, an alternative dosing paradigm coined adaptive therapy has been proposed to maintain the sensitive cell population in a tumor and thus prevent unchecked proliferation of the drug-resistant cells. In this talk, we use a model of cellular competition to mathematically explore two distinct paradigms of adaptive therapy dosing: continuous dose modulation versus intermittent high dose therapy. We compare these regimens to standard dosing schemes to explore how dose and dose schedule can drive cellular population evolution.
4. Sarthak Sahoo Indian Institute of Science
   "Mathematical modelling of multi-axis plasticity in ER+ breast cancer"
Resistance of anti-estrogen therapy is a major clinical challenge in treating estrogen receptor positive (ER+) breast cancer. Recent studies highlight the role of non-genetic adaptations in drug tolerance, yet the mechanisms remain unclear. One of the key processes that underlies enhanced drug tolerance as well as metastatic potential is the process of epithelial-mesenchymal plasticity (EMP). Furthermore, lineage switching, and acquisition of immunosuppressive traits are also commonly observed. We investigate the role of genes involved in EMP to elucidate origins of such multi-axis cellular plasticity by mathematical modelling of underlying gene regulatory networks. Specifically, we show that six co-existing phenotypes are enabled by underlying gene regulatory network, with epithelial-sensitive and mesenchymal-resistant being dominant. Population dynamics analysis demonstrates how phenotypic plasticity promotes survival among sensitive and resistant cells, suggesting that mesenchymal-epithelial transition inducers could enhance anti-estrogen therapy effectiveness. Additionally, hybrid epithelial/mesenchymal (E/M) phenotypes, which exhibit high PD-L1 levels, contribute to immune evasion and enhanced metastatic fitness, obviating the need for a full EMT. Finally, multi-modal transcriptomic data analysis shows associations between EMT and luminal-basal plasticity, linking luminal breast cancer with epithelial states and basal breast cancer with hybrid E/M phenotypes and higher heterogeneity. Our mechanistic modelling of ER+ breast cancer recapitulates observed clinical and pre-clinical findings in spatial transcriptomic datasets, thus offering a predictive framework for characterizing intra-tumor heterogeneity and potential therapeutic interventions given a spatially heterogenous tissue sample. These insights underscore the importance of understanding phenotypic plasticity and non-genetic heterogeneity and its integration with spatial transcriptomics to improve treatment strategies for ER+ breast cancer.

1. Gustav Lindwall Max Planck Institute for Evolutionary Biology
   "A Mathematical Model for Pseudo-Progression in CAR-T therapy of B-cell Lymphomas"
CAR-T cell therapy, where patient T cells are genetically modified to target CD19-presenting B cells, has transformed the treatment landscape for several types of B-cell lymphomas. However, this therapy often triggers a strong inflammatory response, which can cause the tumor to temporarily swell in the days following CAR-T infusion — a phenomenon known as pseudo-progression. In this work, we present a dynamical model of CAR-T cell therapy that explicitly incorporates the role of pro-inflammatory cytokines in shaping treatment outcomes. Our model reproduces a wide spectrum of clinical trajectories, including complete remission, treatment failure, and transient pseudo-progression. Importantly, the model’s parameters correspond to measurable patient-specific factors, allowing us to explore how individual patient characteristics influence long-term treatment success. The parametrization of the model maps on to measurable patient characteristics, and we discuss how these parameters impact the long term behavior of the model.
2. Rafael Bravo University of Texas at Austin
   "Testing the feasibility of estimating the migration to proliferation rate ratio in glioblastoma from single time-point MRI data"
Introduction: Glioblastoma tumors with a high migration to proliferation ratio (D/k ratio) are more drug resistant, suggesting 1) that estimating D/k ratios can help predict patient responses, and 2) investigating the cellular mechanisms behind D/k ratios can help understand the mechanisms of drug resistance. Here we quantify the identifiability of D/k ratios using synthetic tumor data. Materials and Methods: We used the standard reaction-diffusion model (i.e. logistic growth and diffusion) initialized with a single cell. Using this model, we grew synthetic tumors, saving the density field for calibration once the tumor had filled 10% of the domain. We fixed proliferation (k, 1/day) and used grid search followed by Levenberg-Marquardt optimization to calibrate the diffusion rate (D, mm2/day) and growth time (t, days) that produced density distributions matching the synthetic data as closely as possible. We quantified the ability of the algorithm to accurately and precisely identify D/k ratios in both the presence and absence of Gaussian noise. Results: Our algorithm finds D/k ratios with very high accuracy: 0.95 +/- 0.72% difference from correct D/k with k fixed at 0.01/day. We found that with 5% noise added the ability to accurately recover D/k ratios improved as its magnitude increased: 9.6 +/- 13.38% when D/k = 10-2 mm2 versus 2.7 +/- 3.65% when D/k = 1 mm2 Future Directions: Ongoing work will establish the minimal data requirements to accurately estimate D/k ratios within 10% of the correct values, and then apply the technique to the brain tumor image segmentation (BRATS) dataset. All BRATS patients have a single segmented pre-treatment MRI (N = 103) which we will use to estimate their D/k ratios. A subset of the BRATS patients also has RNA microarray data available (N = 91). We plan to correlate the patients’ D/k ratios with gene set scores derived from their microarray data to identify cellular mechanisms that potentially underly the D/k ratios.
3. Aaron Li University of Minnesota
   "Using a pharmacokinetic ctDNA shedding model to develop a biomarker of tumor response to targeted therapy"
Early prediction of response to therapy or lack thereof is essential for efficient treatment planning. Next-gen sequencing (NGS) has made it possible to non-invasively collect and sequence circulating tumor DNA (ctDNA) from longitudinal plasma samples. While ctDNA data continues to be collected on a wide variety of cancers and treatment types, it is still unclear what ctDNA biomarkers are most indicative of treatment success or failure. We present a pharmacokinetic model of ctDNA shedding under targeted therapy. Using this model to simulate a cohort of virtual patients, we demonstrate the predictive potential of a biomarker based on early ctDNA sampling. We compare the performance of the biomarker to that of a neural network classifier as well as existing ctDNA biomarkers. We show that our biomarker is able to match and exceed the performance of alternatives, both in terms of accuracy and earliness of prediction.
4. Pujan Shrestha Texas A&M University
   "An ODE-SDE Model for Ct-DNA dynamics"
Effective cancer therapies, while continuously improving, are often constrained by lower detection limits of disease. Tumor-immune dynamics in this limit present one of the most pressing knowledge gaps as cancer ultimate escape or elimination are often determined following an intervening period of population equilibrium sustained at low population size. Population dynamics in this small-population limit are affected by intrinsic noise in the tumor-immune interaction, as are estimates of population disease burden by extrinsic noise in acquiring such estimates through associated biomarkers. We present a modeling framework that investigates the interactions between tumor cells and the immune system in the small population regime, focusing on how these interactions influence biomarker levels. The framework combines deterministic elements, which describe tumor growth and immune responses, with stochastic components that capture the inherent variability in biomarker release. We use a system of ordinary differential equations (ODEs) to represent the tumor-immune dynamics between an adaptive immune compartment, immunogenic tumor cells, and evasive tumor cells. The immune system’s role in controlling tumor growth is reflected in the tumor-immune interaction terms. Apoptotic death via tumor-immune interactions and necrotic death via the tumor competition under a shared carrying capacity both contribute to the release of a tumor biomarker. We focus on applying our model to ct-DNA, wherein we frame ct-DNA dynamics using a stochastic differential equation (SDE). This SDE framework accounts for the variability in ct-DNA release due to the dynamic tumor-immune interactions, as well as inherent biological noise, such as DNA degradation and clearance. By coupling the ODE system of equations for tumor-immune dynamics with the SDE for ct-DNA release, we can use the model to study the fluctuations in ct-DNA levels driven by tumor-immune dynamics and exogenous sampling noise.

1. Ruby Nixson Mathematical Institute, University of Oxford
   "A structured-PDE approach to targeting a quiescent sub-population under hypoxia and anti-tumour therapies in paediatric glioma."
Paediatric diffuse midline gliomas are highly aggressive, incurable, childhood tumours. Their location in the brainstem limits treatment to radiotherapy, which allows an average survival time of 9-11 months. A sub-population of quiescent tumour cells are thought to be responsible for the poor outcomes of these patient. Quiescence is often viewed as a reversible resting state in which cells temporarily exit the cell cycle, the process controlling DNA replication and cell division. Radiosensitivity varies during the cell cycle, and quiescent cells exhibit a higher relative level of radio-resistance. Hypoxia (physiologically low levels of oxygen) also impacts cell cycle progression and quiescence, as well as response to radiotherapy, contributing to poor patient outcomes. We build on existing mathematical models of cell cycle progression under treatment which account for the radio-resistance of quiescent cells and their ability to re-enter the cell cycle and proliferate. We derive a system of partial differential equations (PDEs), which structures cells by the time spent in each cell cycle phase and allows transitions to and from a quiescent phase. By considering oxygen-dependent cell cycle progression, we use the model to investigate how the proportion of quiescent cells changes when we impose fluctuating oxygen dynamics and treat with radiotherapy. We extend existing studies that optimise treatment schedules using a balance of treatment outcome and toxicity/cost by incorporating a fixed radiotherapy schedule to investigate the impact of a hypothetical drug that alters the transition dynamics to and/or from quiescence. By considering different mechanisms of action for this hypothetical drug, we use our PDE model to identify candidate drugs with the potential to slow tumour progression and improve patient outcomes. This work will inform our clinical collaborators if such an improvement is possible, and what the design of a suitable drug should be.
2. Reshmi Patel The University of Texas at Austin
   "MRI-based mathematical modeling to predict the response of cervical cancer patients to chemoradiation"
Concurrent chemoradiation followed by brachytherapy is the standard-of-care treatment for locally advanced cervical cancer (LACC), but 30% of treated patients experience local recurrence [1], indicating a need for patient-specific, optimized therapeutic regimens to improve outcomes. We aim to predict patient-specific response to chemoradiation by applying our established MRI-based mathematical modeling framework [2]. The study cohort consisted of 10 LACC patients who underwent imaging with T2-weighted MRI, dynamic contrast-enhanced MRI, and diffusion-weighted MRI (DWI) before (V1), after two weeks (V2), and after five weeks (V3) of chemoradiation [3]. We registered all patient-specific MRI data within and between visits, and maps of the number of tumor cells (NTC) were calculated from the DWI-derived apparent diffusion coefficients. Our biology-based reaction-diffusion models characterize the spatiotemporal change in NTC as a function of cell diffusion, proliferation, and therapy-induced death. We consider two model options for cell death: (A) distinct chemotherapy (exponential decay) and radiotherapy (instantaneous decrease according to the linear-quadratic model) terms and (B) a single exponential decay term describing chemoradiation. Proliferation and chemoradiation efficacy rates were calibrated to the V1 and V2 NTC maps, and the calibrated model was run forward to predict the NTC at V3. Using Model (A), the concordance correlation coefficient (CCC) between the observed and predicted V1 to V3 change in total tumor cellularity was 0.87, and the CCC between the observed and predicted change in tumor volume was 0.90; using Model (B), the CCC values were 0.97 and 0.90, respectively. These preliminary findings show the promise of our mathematical modeling framework in predicting LACC response to chemoradiation. References: [1]. CCCMAC. Cochrane Database Syst Rev. 2010. [2]. Jarrett et al. Nat Protoc. 2021. [3]. Bowen et al. J Magn Reson Imaging. 2018.
3. Fabian Spill University of Birmingham
   "Regulation of Intra- and Intercellular Metabolite Transport in Cancer Metabolism"
Metabolite transport is essential for cellular homeostasis, energy production, and metabolic adaptation. In cancer, dysregulated transport sustains tumor growth and alters redox balance. The mitochondrial solute carrier SLC25A10 facilitates succinate, malate, and phosphate exchange, influencing central carbon metabolism. However, its transport kinetics and physiological directionality remain poorly understood. We present a mathematical model of SLC25A10 based on a ping-pong kinetic mechanism, capturing competitive dynamics between malate and succinate. Our simulations reveal that under normal conditions, malate flux dominates due to its higher binding affinity. However, in succinate dehydrogenase (SDH) dysfunction, excess succinate induces a transient efflux shift and phosphate flux reversal. If experimentally validated, this metabolic shift could serve as a biomarker for tumors with SDH mutations. Integrating our kinetic model with genome-scale metabolic networks, we highlight the role of mitochondrial transport in cancer metabolism. Specifically, in multiple myeloma, metabolic crosstalk between plasma cells and bone marrow stromal cells is key to tumor progression. Our findings demonstrate the power of mathematical modeling in uncovering transport-mediated metabolic vulnerabilities, offering potential therapeutic targets for cancer and metabolic diseases.

1. Veronika Hofmann Technical University of Munich
   "Spectral Spatial Analysis of Cancer Biopsies: Validation through in-silico data and extension to logistic growth models"
MD Anderson's Enderling lab recently invented a spectral spatial analysis method for estimating tumor cell diffusivity and proliferation rate from single-point-in-time biopsies of breast cancer. In combination with clinical data from the patients these parameters could help identify a new biomarker for radiotherapy. In their first study, they investigate the relationship between the power spectral density (PSD) of the three-dimensional reaction-diffusion (RD) equation with exponential growth (as model of spreading cancer cells) and the two-point correlation function of the cell distribution in the biopsy (a spatial statistic). Their results make the approach seem promising, and this work aims to validate and extend their findings. Firstly, we develop a model to generate in-silico data to validate the parameter estimation method. This is done by solving the RD equation for different growth terms (exponential and logistic), adding Gaussian noise and 'translating' its continuous results into spatial point patterns which are interpreted as cell nuclei in the 'biopsy', and then applying the method to see if the original parameters can be retrieved. This model contains several features: dimensionality can be switched between 2D and 3D, cell size can be adjusted, cuts can be added to the point pattern, and in the 3D case, biopsy thickness is variable and the plane where the slice through the 'tumor' is made can be freely chosen. And secondly, the spectral analysis method is altered by proposing a numerical solution to the PSD of the RD equation with logistic growth (valid for arbitrary dimensions). Logistic growth is assumed to be the more realistic model, however, it is harder to handle as no analytical solution is available for the equation, and hence neither for the PSD. The validation results from the in-silico data are assessed and their meaning for the application to real patient data is discussed under consideration of the different types of cell growth.
2. Nicholas Lai University of Oxford
   "Mathematical Modelling of Tertiary Lymphoid Structures in Cancer"
Tertiary lymphoid structures (TLSs) are organised aggregates of immune cells that form at sites of inflammation in chronic diseases, such as cancer. It is hypothesised that, in cancer, TLSs act as local hubs for the generation and regulation of a tumour-specific immune response from inside the tumour microenvironment (TME). TLSs initially form as well-mixed aggregates of T- and B-cells and mature into organised structures consisting of an inner B-cell zone surrounded by an outer T-cell zone. The presence of TLSs correlates with positive patient outcomes in several cancer types, but the mechanisms governing their formation, maturation, and role in the antitumour response remain poorly understood. Motivated by analysis of spatial transcriptomics images of TLSs in colorectal cancer, we develop an agent-based model to investigate TLS formation, maturation, and function in cancer. We model T-cells and B-cells as discrete agents which are attracted to diffusible chemokines (CXCL13 and CCL19) produced by resident stromal cells in the TME. These interactions lead to the formation of a well-mixed lymphoid aggregate that later matures into distinct T- and B-cell zones due to the segregated expression of these chemokines. Our results identify key parameters governing TLS development and suggest conditions under which TLSs are able to control tumour growth. This framework provides a qualitative basis for understanding TLS dynamics and their potential role in cancer immunotherapy.
3. Chenxu Zhu Institute for Computational Biomedicine - Disease Modeling
   "Machine learning-assisted mechanistic modeling to predict disease progression in acute myeloid leukemia patients"
Blood cell formation is a complex process which is driven by hematopoietic stem cells (HSCs). HSCs give rise to progenitors and precursors which eventually produce mature blood cells, such as white blood cells, red blood cells, and platelets. Acute myeloid leukemia (AML) is an aggressive blood cancer which originates from leukemic stem cells (LSCs) and is characterized by the accumulation of aberrant immature cells, referred to as leukemic blasts. Due to the impairment of healthy blood cell formation, many AML patients suffer from life-threatening complications, such as bleeding or infection. Although treated with high-dose chemotherapy, many patients relapse and need salvage therapy. To reveal the mechanisms of disease progression and relapse, we proposed a mathematical model that accounts for competition of HSCs and LSCs in the stem cell niche and physiological feedback regulations before, during, and after chemotherapy. We fit the model to data of 7 individual patients and simulate variations of the treatment protocol. Our simulation results can recapitulate the non-monotonic recovery of HSCs observed in relapsing patients. The model suggests using the decline of HSC counts during remission as an indication for salvage therapy in patients lacking minimal residual disease markers. To bring our model closer to clinical applications, we propose a machine learning assisted mechanistic model that ensuring adherence to biological principles while learning from a larger clinical AML dataset. By embedding mechanistic constraints into machine learning, we aim to identify patient-specific predictors of relapse while preserving biological interpretability.

1. Ana Forero Pinto Moffitt Cancer Center/ University of South Florida
   "An agent-based model with ECM to study the mechanics of DCIS microinvasions"
Microinvasions in ductal carcinoma in situ (DCIS) are malignant cells that have broken through the basement membrane (BM) and extend into the stroma with no focus larger than 1 mm. Since microinvasions constitute the first step in the metastatic cascade, identifying the causes of microinvasions will help distinguish between progressors or non-progressors among the DCIS patients, thus improving treatment. The mechanical tumor-stroma interactions play an important role in this process. Studies have shown that elevated collagen stiffening, deposition, and fibril crosslinking are correlated with tumor aggressiveness and invasion in breast cancer. Therefore, here we present SilicoDCIS, a 2D off-lattice center-based agent-based model (ABM) of ductal carcinoma in situ (DCIS) growth and its interaction with the extracellular matrix (ECM) to investigate the mechanical conditions that may lead to tumor microinvasions. SilicoDCIS simulates the division, growth, and migration of tumor cells in DCIS while interacting with other cell types and the ECM. This includes the BM, the myoepithelial and epithelial cell layers, and the collagen in the ECM. The ECM was modeled as a vector field, where the direction of each vector gives the orientation of a collagen bundle, and the vector magnitude is related to the bundle density. The growing DCIS can remodel the ECM (density and orientation), and in turn, the ECM applies a reciprocal force (proportional to the local collagen density) opposite to the tumor growth. With SilicoDCIS, we studied the mechanical effects of cancer cell proliferation and migration on the BM and the ECM. We found that higher cell migration force leads to increased BM stress and ECM density (on the tumor edges where cells migrate) and that the escape of the migrating cells from the duct vs. their intraductal confinement depends on cell speed. SilicoDCIS may provide insights into the mechanics of DCIS microinvasions to guide the design of future experiments.
2. Chay Paterson University of Manchester
   "Wave-like behaviour in cancer evolution"
Compound birth-death processes are widely used to model the age-incidence curves of many cancers [1]. There are efficient schemes for directly computing the relevant probability distributions in the context of linear multi-stage clonal expansion (MSCE) models [2]. However, these schemes have not been generalised to models on arbitrary graphs, forcing the use of either full stochastic simulations or mean-field approximations, which can become inaccurate at late times or old ages [3, 4]. Here, we present a numerical integration scheme for directly computing survival probabilities of a first-order birth-death process on an arbitrary directed graph, without the use of stochastic simulations. As a concrete application, we show that this new numerical method can be used to infer the parameters of an example graphical model from simulated data.
3. Nathan Schofield University of Oxford
   "Mechanistic modelling of cluster formation in metastatic melanoma"
Melanoma is the most aggressive type of skin cancer, yet survival rates are excellent if it is diagnosed early. However, if metastasis occurs, five-year survival rates drop significantly. During the early stages of tumour initiation, melanoma cells form clusters within the primary tumour which promote metastasis. In the absence of biological tools to visualise cluster formation at primary tumour sites, we develop mathematical models to generate mechanistic insight into their formation. For this work we utilise in vitro data for two distinct melanoma cell phenotypes, one more proliferative and the other more invasive. This data consists of experiments for each phenotype individually, resulting in homogeneous clusters, as well as mixtures of the two phenotypes, resulting in heterogeneous clusters. We develop a series of differential-equation-based models using a coagulation-fragmentation-proliferation framework to describe the growth dynamics of homogeneous clusters, incorporating different functional forms for cell proliferation and cluster splitting. We then extend these models to describe the formation of heterogeneous cell clusters by considering both cluster size and phenotypic composition. We fit the models to experimental data, using a Bayesian framework to perform parameter inference and information criteria to perform model selection. In this way, we characterise and quantify differences in the clustering behaviour of two melanoma phenotypes in homogeneous and heterogeneous clusters, particularly the cluster coagulation, proliferation, and splitting rates. We find that the coagulation rate for the invasive phenotype is much larger than that for the proliferative phenotype, and evaluate how well different modelling assumptions fit the data in order to increase our understanding of the mechanisms driving metastasis. In future work, the models will be used to inform further experiments and, in particular, to suggest and test strategies for inhibiting metastasis.
4. Sergio Serrano de Haro Ivanez University of Oxford
   "Topological quantification of colorectal cancer tissue structure"
A hallmark of colorectal cancer is the structural disruption of the colonic tissue, a process correlated with disease progression. Intestinal crypts, glands essential for homeostasis, lose their tubular morphology - and function - due to uncontrolled cell proliferation and tissue invasion. Evaluating this deterioration in biopsied samples is critical for both patient diagnosis and prognosis. Histopathological methods are essential for assessing colorectal cancer status, but their precision and reproducibility can be improved. Spatial biology provides a mathematical framework to analyse the structural properties of biological data; in this work, we apply techniques from topological data analysis and network science to quantify architectural changes in colorectal cancer progression. Using cell point clouds derived from immunohistochemistry imaging, we construct cell networks that encode topological tissue features. We employ these networks to segment large, imaged samples into smaller, biologically meaningful regions of interest that preserve tissue architecture. We compare the performance of our approach to conventional segmentation methods such as quadrat division. Within these segmented regions, we further employ methods from persistent homology to quantify tissue structure, with the long-term goal of identifying novel biomarkers of disease progression.

1. Paulameena Shultes Case Western Reserve University
   "Cell-Cell Fusion in Cancer: Key In Silico Tumor Evolutionary Behaviors"
Cell-cell fusion is a known phenomenon throughout the human body. It characterizes a wide range of physiological and pathological processes, ranging from placentation and embryogenesis to cancer stem cell (CSC) formation. There is increasing evidence that cell-cell fusion can play key roles in the development and progression of cancer, particularly by increasing intratumor heterogeneity and potentiating somatic evolution. There are many unanswered questions surrounding the characteristics that define cancer cell-cell fusion events, their frequency in in vivo tumor conditions, and whether or not cell-cell fusion is a universal phenomenon across cancer. Using a combination of in vitro and in silico approaches, we can begin to answer some of these questions. We have developed a preliminary cellular automata model using HAL to evaluate the effect of variable cell-cell fusion rates and behaviors under a range of tumor microenvironmental conditions. By comparing our spatial model to a suite of ordinary differential equations, we can begin to estimate the effects of cell-cell fusion on the genomic heterogeneity and malignancy potential of cancers in vivo. I demonstrate the importance of improving fusion rate estimates using the simplest iteration of an in silico cellular automata model (coined SimpleFusion). The preliminary SimpleFusion model results illustrate how much the impact of cell fusion, as measured by the percentage of cells that have had a fusion event in their lineage, changes between orders of magnitude of fusion rates. Corresponding ODE models demonstrate similar results despite the lack of encoded spatial information. By studying these two types of models (ABM, ODEs) in combination, we can begin to understand what parameters most directly define the cell-cell fusion population dynamics in our in vitro fusion experiments and, in turn, in vivo conditions as well.
2. Thomas Stiehl Institute for Computational Biomedicine and Disease Modeling, University Hospital RWTH Aachen, Aachen, Germany & Department of Science and Environment, Roskilde University, Roskilde, Denmark
   "Computational Modeling of the Aging Human Bone Marrow and Its Role in Blood Cancer Development"
Blood cancers pose a growing medical and economic challenge in aging societies. Every day, the human bone marrow (BM) generates more than 100 billion blood cells. This process is driven by hematopoietic stem cells (HSCs), which retain their ability to proliferate and self-renew throughout life. However, over time, HSCs accumulate mutations that may lead to malignant transformation, as seen in acute myeloid leukemia (AML), one of the most aggressive cancers. Even in healthy individuals, the BM undergoes age-related changes, including a decline in cell numbers, remodeling of the BM micro-environment, and a bias in HSC differentiation. Emerging evidence suggests that these alterations create a favorable environment for the expansion of mutated cells, thereby promoting blood cancer development and progression. Mathematical and computational models facilitate our understanding of how BM aging contributes to malignant cell growth. We propose nonlinear ordinary differential equation models to describe blood cell formation and clonal competition in the human BM. The models incorporate micro-environmental and systemic feedback loops and are informed by data from both healthy individuals and cancer patients. Our findings suggest that the age-related decline in HSC self-renewal, combined with increased chronic inflammation (inflammaging), makes the BM more susceptible to the expansion of mutated cells and at the same time impairs treatment response. Through mathematical analysis, quantitative simulations, and patient data fitting, we study the following questions: 1. How do HSC proliferation & self-renewal change during physiological aging? 2. How do age-related alterations in healthy BM contribute to blood cancer development? 3. What is the impact of chronic inflammation on HSC function and blood cancer progression? 4. How do age-related BM changes affect treatment responses, e.g., in AML patients? 5. How could treatment protocols be adapted to elderly patients?
3. Aisha Turysnkozha Nazarbayev University
   "Traveling wave speed and profile of a “go or grow” glioblastoma multiforme model"
Glioblastoma multiforme (GBM) is a fast-growing and deadly brain tumor due to its ability to aggressively invade the nearby brain tissue. A host of mathematical models in the form of reaction–diffusion equations have been formulated and studied in order to assist clinical assessment of GBM growth and its treatment prediction. To better understand the speed of GBM growth and form, we propose a two population reaction–diffusion GBM model based on the ‘go or grow’ hypothesis. Our model is validated by in vitro data and assumes that tumor cells are more likely to leave and search for better locations when resources are more limited at their current positions. Our findings indicate that the tumor progresses slower than the simpler Fisher model, which is known to overestimate GBM progression. Moreover, we obtain accurate estimations of the traveling wave solution profiles under several plausible GBM cell switching scenarios by applying the approximation method introduced by Canosa.
4. Brian Johnson UC San Diego
   "Integrating clinical data in mechanistic modeling of colorectal cancer evolution in inflammatory bowel disease"
Patients with inflammatory bowel disease (IBD) face an elevated risk of colorectal cancer (CRC), necessitating lifelong surveillance to find and remove precancers before they become malignant. Current one-size-fits-all approaches are inadequate and tailored strategies that consider cancer evolution are needed. To address this, we developed a mechanistic framework of IBD-CRC progression. Our multi-type branching process model accounts for IBD onset, mutational processes, and both precancerous (adenoma/dysplasia) and malignant clonal expansion. Initial parameter estimation for mutation and growth rates when fitting the multi-stage clonal expansion model to epidemiological IBD-CRC data yielded similar estimates to those found previously in sporadic CRC but suggest higher mutation rates and slightly lower growth rates in IBD. However, this data may not perfectly represent the natural history, as surveillance colonoscopy with lesion removal and colectomy alter the observable progression. Further, fitting to cancer incidence data alone presents parameter identifiability issues, restricting our initial fit to four parameters. To address these limitations, our study draws upon extensive clinical data from the U.S. Veterans Health Administration, employing validated methods using large language models to construct high-quality datasets with detailed information on surveillance colonoscopy timing, colectomies, and intermediate lesions extracted from pathology reports. To integrate these data, we developed a complementary fast simulation model, which will be released as an R package. This simulation model incorporates clinical interventions, such as colonoscopy with size-dependent lesion removal. Our combined analytical and simulation approach captures the complex precancerous evolution in IBD, providing a quantitative foundation for more effective, personalized surveillance guidelines. Further, this approach can be adapted to improve surveillance in the general population.

1. Siti Maghfirotul Ulyah Khalifa University, Abu Dhabi, United Arab Emirates
   "Estimating the Growth Rate of Tumor Cells from Biopsy Samples Using an Extended Mean Field Approximation"
A biopsy is a common procedure used to diagnose diseases like cancer, infections, or inflammatory conditions. In cell population studies, biopsy samples provide valuable data to analyze cellular growth, proliferation rates, and structural abnormalities, which are essential for understanding disease progression. Estimating the growth (proliferation) rate of human cells is a challenging task. To address this, we have developed a method based on the birth-death Markov process to simulate the logistic growth model. We applied an extended Mean Field Approximation (MFA) for birth-death Markov processes, which accounts for fluctuations in the evolution of observables, such as moments. By calculating the theoretical moments from the birth-death process, we solved the inverse problem and estimated the growth rate. Additionally, we performed Markov Chain Monte Carlo (MCMC) simulations for both logistic growth and logistic growth with the Allee effect. The moments of the simulated population were used to predict the growth rate through regression analysis, achieving a high R-squared value. Finally, by applying this approach to biopsy data, one can estimate the proliferation rate of human cells with greater accuracy.
2. Hooman Salavati Ghent University
   "Patient-Specific MRI-Integrated Computational Modeling of Tumor Fluid Dynamics and Drug Transport"
Introduction: Mathematical modeling is a key tool for understanding solid tumor biophysics, progression, and treatment resistance. Biophysical changes, such as elevated interstitial fluid pressure (IFP), are identified as major barriers to effective drug delivery. Incorporating patient-specific data into mathematical models offers the potential for personalized prognosis and treatment strategies for cancer patients. In this study, we explored the integration of patient-specific data from dynamic contrast-enhanced MRI (DCE-MRI) and diffusion-weighted MRI (DW-MRI) into a computational fluid dynamics (CFD) model of solid tumors to estimate the IFP and drug penetration profiles. Methods: As part of a translational study (EC/2019/1330, approved by Ghent University Hospital, Belgium), a patient with peritoneal metastasis underwent multi-sequential MRI, including T1-weighted (T1w) anatomical imaging, DCE-MRI, and DW-MRI. Tumor interstitial fluid pressure (IFP) was directly measured using a pressure transducer-tipped catheter for model validation. The CFD tumor model described interstitial fluid flow using Darcy’s law, the continuity equation, and Starling’s law, while drug penetration was modeled via the convection-diffusion-reaction equation. The 3D tumor geometry was derived from T1w images, vascular permeability from DCE-MRI using the extended-Tofts model, and hydraulic conductivity from DW-MRI. Results: An elevated IFP zone was observed in the central region of the tumor (up to 14 mmHg), while a lower IFP zone appeared at the tumor's edge. The clinically recorded IFP values (12.0 ± 2.5 mmHg) corresponded well with the simulation results. Drug penetration varied across the tumor, with deeper penetration in low-IFP regions. Conclusion: An image-based CFD model captured IFP and drug distribution variability, aligning with clinical data. This approach advances personalized oncology, potentially improving treatment strategies through noninvasive, patient-specific modeling.
3. Rachel Sousa University of California, Irvine
   "Identifying Critical Immunological Features of Tumor Control and Escape Using Mathematical Modeling"
The immune system can eradicate cancer, but various immunosuppressive mechanisms active within a tumor curb this beneficial response. Cytotoxic T cells (CD8s), regulatory T cells (Tregs), and antigen-presenting dendritic cells (DCs) play an important role in the immune response; however, it is very cumbersome to unravel the effects of multimodal interactions between tumor and immune cells and their contributions to tumor control using an experimental approach alone. Thus, to better understand the mechanisms that govern the interactions between immune cells and tumor cells and to identify the critical immunological features associated with tumor control and tumor escape, we built a mechanistic mathematical model of CD8s, Tregs, DCs, and tumor cells. The model accounts for tumor immunogenicity, the effects of IL-2 prolonging T cell lifespan, Treg suppression of antitumor immune response through CTLA-4, recruitment of immune cells into the tumor environment, and interferon-gamma upregulation of PD-L1 on DCs and tumor cells to deactivate T cells. We successfully fit the model to experimental data of tumor and immune cell dynamics. Employing Latin Hypercube Sampling, we generated over 1000 parameter sets that capture the sensitivity of αPD-1 immunotherapy. By comparing the parameter sets, we gain an insight into which mechanisms impinge the success of immunotherapy. We are now utilizing the model to explore combination immunotherapies that enhance the immune response in partial- and non-responders of αPD-1 immunotherapy. In particular, we are investigating what combinations of αPD-1, αCTLA-4, αICOSL, and αLAG-3 will stimulate CD8 activation without promoting Treg activation. After identifying the top combination therapies, we will validate our predictions experimentally. This integrated approach of modeling and experimental validation aims to advance our understanding of tumor-immune interactions and guide the development of more effective immunotherapeutic strategies.
4. Alexis Farman UCL (University College London)
   "Enhancing immunotherapies: Insights from the mathematical modelling of a microfluidic device"
A pivotal aspect of developing effective immunotherapies for solid tumors is the robust testing of product efficacy inside in vitro platforms.Collaborating with an experimental team that developed a novel microfluidic device at Children’s National Hospital (CNH), we developed a mathematical model to investigate immune cell migration and cytotoxicity within the device. Specifically, we study Chimeric Antigen Receptor (CAR) T-cell migration inside the channels, treating the cell as a moving boundary driven by a chemoattractant concentration gradient. The chemoattractant concentration is governed by two partial differential equations (PDEs) that incorporate key geometric elements of the device. We examine the motion of the cell as a function of its occlusion of the channel and find that certain cell shapes allow for multiple cells to travel inside the channel simultaneously. Additionally, we identify parameter regimes under which cells clog the channel, impairing their movement. All our findings are validated against experimental data provided by CNH. We integrate our model results into a broader model of the device, which also examines the cytotoxicity of CAR T-cells. This provides a tool for distinguishing experimental artifacts from genuine CAR T-cell behavior. This collaboration enabled the team at Children’s National Hospital to refine experimental conditions and uncover mechanisms enhancing CAR T-cell efficacy. [1] D Irimia, G Charras, N Agrawal, T Mitchison, M Toner, Polar stimulation and constrained cell migration in microfluidic channels,, Lab on a Chip 7 (12), 1783-1790

1. Simon Syga TUD Dresden University of Technology
   "Evolution of phenotypic plasticity leads to tumor heterogeneity with implications for therapy"
Cancer is a significant global health issue, with treatment challenges arising from intratumor heterogeneity. This study examines the complex relationship between somatic evolution and phenotypic plasticity, explicitly focusing on the interplay between cell migration and proliferation [1]. We propose that evolution does not act directly on phenotypic traits, like the proliferation rate, but on the phenotypic plasticity in response to the microenvironment [2]. We study this hypothesis using a novel, spatially explicit model that tracks individual cells' phenotypic and genetic states. We assume cells change between mobile and growing states controlled by inherited and mutation-driven genotypes and the cells' microenvironment. We observe that cells at the tumor edge evolve to favor migration over proliferation and vice versa in the tumor bulk. However, this phenotypic heterogeneity can be realized by distinct regulations of the phenotypic switch, which depend on the apoptosis rate and the cells' ability to sense their environment. Emerging synthetic tumors display varying levels of heterogeneity, which we show are predictors of the cancer's recurrence time after treatment. Interestingly, higher phenotypic heterogeneity predicts poor treatment outcomes, unlike genetic heterogeneity. [1] Hatzikirou, H. et al. (2010). 'Go or Grow': the key to the emergence of invasion in tumour progression? Math. Med. Biol., 29(1), 49-65. [2] Syga S. et al. (2024) Evolution of phenotypic plasticity leads to tumor heterogeneity with implications for therapy. PLOS Comput. Biol. 20(8): e1012003.
2. Luke Heirene University of Oxford
   "Data Driven Mathematical Modelling Highlights the Impact of Bivalency on the Optimum Affinity for Monoclonal Antibody Therapies"
Monoclonal antibody (mAb)-based therapeutics are pivotal in treating a wide range of diseases, including cancer. One key mechanism by which these antibodies exert anti-tumour effects is through antibody-dependent cellular cytotoxicity (ADCC). In ADCC, mAbs bind to specific antigens on tumour cells and Fc receptors on immune effector cells. This trimeric complex triggers these effector cells to kill the tumour. ADCC is influenced by multiple factors, notably the properties of the mAb and its interactions with Fc receptors and target antigens. However, the optimum conditions for ADCC remain unclear. In this study, we investigate how variations in target antigen and mAb properties, particularly antibody valency, modulate ADCC response to identify parameters that maximize its potency. We developed an ordinary differential equation (ODE) model to simulate mAb binding within the immune synapse and quantify trimeric complex formation. To link the number of trimeric complexes to ADCC response, we validated the model using Bayesian inference on ADCC assay data. The results suggest that lower-affinity mAbs enhance ADCC by increasing the number of target cell-bound antibodies. Our validated model indicates that a “steric penalty” is necessary for bivalently target-bound versus monovalently target-bound antibodies. Due to constraints from dual antigen binding, these antibodies experience limited mobility, reducing Fc receptor engagement. After model validation, we explored variations in target expression, binding affinity, and antibody valency on ADCC potency, quantified by EC50. Our key finding is that the optimal binding affinity for maximizing ADCC potency depends on antibody valency. Monovalent antibodies are most potent at high affinity, while bivalent antibodies peak at lower affinities. Furthermore, the magnitude of this effect varies with target expression levels.