Computational Modeling of Vascular Tumor Growth and Combined Treatment Response

Lampropoulos, Ioannis; Λαμπρόπουλος, Ιωάννης

dc.contributor.author	Lampropoulos, Ioannis
dc.contributor.author	Λαμπρόπουλος, Ιωάννης
dc.date.accessioned	2025-03-07T07:22:15Z
dc.identifier.uri	https://dspace.lib.ntua.gr/xmlui/handle/123456789/61238
dc.identifier.uri	http://dx.doi.org/10.26240/heal.ntua.28934
dc.rights	Default License
dc.subject	finite elements method	en
dc.subject	cancer biology	en
dc.subject	tumor growth modeling	en
dc.subject	multiphase model	en
dc.subject	bayesian optimization	en
dc.subject	βιολογία του καρκίνου	el
dc.subject	μοντέλα ανάπτυξης καρκίνου	el
dc.subject	μέθοδος πεπερασμένων στοιχείων	el
dc.subject	συνδυαστική χημειοθεραπεία	el
dc.subject	βελτιστοποίηση θεραπείας	el
dc.title	Computational Modeling of Vascular Tumor Growth and Combined Treatment Response	en
dc.contributor.department	Τομέας Ανάλυσης, Σχεδιασμού και Ανάπτυξης Διεργασιών και Συστημάτων	el
heal.type	doctoralThesis
heal.classification	chemical engineering	en
heal.classification	bioengineering	el
heal.dateAvailable	2026-03-06T22:00:00Z
heal.language	en
heal.access	embargo
heal.recordProvider	ntua	el
heal.publicationDate	2024-10-18
heal.abstract	Cancer growth is a highly complex, dynamic process driven by various genetic, cellular, and environmental factors. Traditional experimental methods often fall short in fully capturing these intricate interactions. To address this, computational modeling has emerged as a powerful tool for simulating tumor development and progression, providing insights into cancer dynamics at multiple scales. This dissertation presents a comprehensive approach to modeling cancer growth, integrating mathematical frameworks and computational simulations to explore tumor behavior under different biological conditions. The biological areas explored include intra-tumor heterogeneity involving specific secondary mutations expressed by a fraction of cancer cells, vascular heterogeneity with classification based on mechanical robustness, tumor response to cytotoxic therapy, antiangiogenic therapy, radiation therapy and their combinations, and cancer cell-macrophage interactions. In this work, a continuous modeling approach is employed to represent the evolving tumor microenvironment and its interactions with the surrounding tissue. The system is treated as a multiphase fluid consisting of non-mixable, highly viscous phases, each corresponding to a cellular phase, with the interstitial fluid considered a separate phase, though not a cellular structure. For the system's numerical solution, partial differential equations (PDEs) for mass and momentum balance equations were developed with necessary closure relationships. The computational problem was solved through the Finite Elements Method (FEM), utilizing the commercial software Comsol Multiphysics. These models effectively simulate the mechanical, biochemical, and proliferative dynamics of cancerous cells and their neighbouring cellular populations. The first field of study is intra-tumor heterogeneity (ITH), where certain cancer cells within a tumor acquire advantageous properties due to secondary mutations. ITH plays a significant role in cancer diagnosis and is associated with acquired resistance to therapy. In the present thesis, three instances of ITH were simulated: 1) overexpression of epidermal growth factor receptor (EGFR), 2) overexpression of heme oxygenace-1 (HO-1), and 3) acquired resistance to the cytotoxic drug docetaxel, a commonly used cytotoxic drug which belongs in the category of taxanes. For the first two cases, heterogeneity correlates with more aggressive tumors, showingincreased angiogenic and metastatic potential. Computational experiments with docetaxel resistance demonstrated reduced drug efficacy against the mutated cancer sub-population. The study of ITH highlighted the Darwinian nature of the cancer micro-environment. With cytotoxic chemotherapy in place, we modeled the anti-angiogenic agent bevacizumab, a monoclonal antibody that binds the free VEGF molecules inhibiting angiogenesis initiation. This results in suppression of the peripheral capillary network, depriving the tumor of nutrients and limiting its spread. Bevacizumab is often combined with taxane chemotherapy in clinical settings for several types of cancer, such as breast and ovarian cancer. Our findings indicate that while bevacizumab monotherapy is not highly effective, its combination with cytotoxic agents is significant, prolonging the residence time of docetaxel and enhancing its efficacy. Lastly, we performed a series of computational experiments suggesting that optimal therapeutic outcomes occur with the concurrent administration of both drugs, consistent with clinical practice. Building on these anti-cancer therapies, we integrated external radiotherapy. The inclusion of radiation therapy in combination with docetaxel and bevacizumab further enhanced therapeutic efficacy. Radiation therapy significantly reduced the tumor's cancerous population, making subsequent combination treatments more effective. We also explored radiosensitization, in which agents like oxygen and docetaxel increase the efficacy of radiation therapy. This revealed the synergistic effects of the therapeutic agents and allowed us to study the impact of radiation on tumor morphology. Next, we focused on macrophages, key components of the innate immune response, and their interactions with tumor spheroids. Macrophages can adopt divergent phenotypes depending on environmental conditions: an anti-tumor M1 phenotype or a pro-tumor M2 phenotype. We simulated a wide range of phenomena, including macrophage infiltration, chemotaxis, and activation, and modeled tumor-macrophage interactions in multi-seed tumors to study macrophage localization in hypoxic pockets. Finally, immunotherapy was also implemented via the experimental agent vactosertib, designed to safeguard M1 macrophages from transitioning to a pro-tumor state. The final chapter of this thesis focused on optimizing combination therapy scheduling using Bayesian optimization. This probabilistic approach enables efficient exploration of the parameter space to predict tumor growth trajectories under varying therapeutic strategies. We optimized the timing between initial administrations of each therapy to achieve the best long-term tumor suppression, validating that concurrent administration yields optimal effects. We also investigated optimal schedules in triple chemo-radiotherapy schemes; there, the possibility of total tumor elimination was investigated, using Bayesian optimization to determine the minimal time required for this outcome. Our models successfully replicate various reported behaviours in the tumor microenvironment, including VEGF localisation in hypoxic zone, increased cytotoxic therapy residence time in the presence of antiangiogenic therapy, and localisation of pro-tumor macrophages in hypoxic niches of the tumor spheroid. Furthermore, our model's findings align with clinical practice regarding the concurrent use of docetaxel and bevacizumab for optimal long-term outcomes. In summary, this study presents a versatile computational framework suitable for the realistic simulation of a wide range of cancer biology phenomena. The model provides qualitative predictions of behaviours within the cancer microenvironment and can be used to suggest optimal therapeutic strategies.	en
heal.sponsor	The author acknowledges the support from: • the N.T.U.A. ELKE scholarship for doctoral candidates. • the Basic Research Program, NTUA, PEVE (No. 65232000).	en
heal.advisorName	Καβουσανάκης, Μιχαήλ
heal.committeeMemberName	Παπαθανασίου, Αθανάσιος
heal.committeeMemberName	Μπουντουβής, Ανδρέας
heal.committeeMemberName	Παπαδόπουλος, Βησσαρίων
heal.committeeMemberName	Κλάπα, Μαρία
heal.committeeMemberName	Byrne, Helen
heal.committeeMemberName	Kevrekidis, Panayotis
heal.academicPublisher	Σχολή Χημικών Μηχανικών	el
heal.academicPublisherID	ntua
heal.numberOfPages	231
heal.fullTextAvailability	false