Approaches to Generating Virtual Patient Cohorts with Applications in Oncology

Surendran, Anudeep; Le Sauteur-Robitaille, Justin; Kleimeier, Dana; Gevertz, Jana; Wilkie, Kathleen; Jenner, Adrianne L.; Craig, Morgan

doi:10.1007/978-3-031-32614-1_8

Cited by 5 publications

(7 citation statements)

References 56 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Though not particularly surprising, we also observe that for this treatment protocol, whether a storm occurs is much more sensitive to the rate of cytokine stimulation bxy than to the rate of tumor killing kT. If we think of each point in this parameter space as representing a plausible "virtual patient" (38,39), we can immediately see a problem with treating everyone with a single dose of 10 units. While some individuals will be effectively treated with limited toxicity (no cytokine storm), other individuals run the risk of a cytokine storm accompanying the effective treatment, and others may not be able to be effectively treated by this dose, toxicity considerations aside.…”

Section: All Four Regimes Can Occur For the Same Dose Depending On In...mentioning

confidence: 78%

Cytokine storm mitigation for exogenous immune agonists

Kareva

Gevertz

2023

Preprint

View full text Add to dashboard Cite

Cytokine storm is a life-threatening inflammatory response characterized by hyperactivation of the immune system. It can be caused by various therapies, auto-immune conditions, or pathogens, such as respiratory syndrome coronavirus 2 (SARS-CoV-2) which causes coronavirus disease COVID-19. Here we propose a conceptual mathematical model describing the phenomenology of cytokine-immune interactions when a tumor is treated by an exogenous immune cell agonist which has the potential to cause a cytokine storm, such as CAR T cell therapy. Numerical simulations reveal that as a function of just two model parameters, the same drug dose and regimen could result in one of four outcomes: treatment success without a storm, treatment success with a storm, treatment failure without a storm, and treatment failure with a storm. We then explore a scenario in which tumor control is accompanied by a storm and ask if it is possible to modulate the duration and frequency of drug administration (without changing the cumulative dose) in order to preserve efficacy while preventing the storm. Simulations reveal existence of a sweet spot in protocol space (number versus spacing of doses) for which tumor control is achieved without inducing a cytokine storm. This theoretical model, which contains a number of parameters that can be estimated experimentally, contributes to our understanding of what triggers a cytokine storm, and how the likelihood of its occurrence can be mitigated.

show abstract

Section: All Four Regimes Can Occur For the Same Dose Depending On In...mentioning

confidence: 78%

Cytokine storm mitigation for exogenous immune agonists

Kareva

Gevertz

2023

Preprint

View full text Add to dashboard Cite

show abstract

“…Yet another complication is that while a PPop can help explore the broad range of responses the model can produce, the outcome may not necessarily reflect the distribution of population-level data; that is, the probability of observing each outcome [8]. In this work, and in other scenarios where such population-level data is not available, one can follow the lead of Surendran et al [7] and treat the plausible population as the virtual population. Even when population-level data is available, a downside of "matching" the VPop to population-data is that the VPop will then recreate the biases inherent in that dataset [3].…”

Section: Discussionmentioning

confidence: 99%

“…In this method, a model parametrization is determined by randomly sampling (a subset of the) model parameters from a uniform distribution with userspecified lower and upper bounds. Each parametrization p = (r, d) is then optimized using simulated annealing (simannealbnd in MATLAB) to ensure that the volume trajectory corresponding to the virtual patient parametrization falls within the feasible region F. The objective function to be minimized using simulated annealing is [7,26]:…”

Section: Methods For Generating Virtual Patientsmentioning

confidence: 99%

“…In the first method, that we call the "accept-or-reject" method, a model parametrization is determined by randomly sampling (a subset of) parameters from a normal distribution, while rejecting any negative parameters [6,7]. As the normal distribution is used in all implementations of this method that we are familiar with [6,7,21], we will refer to this as the "standard prior" for the accept-orreject method.…”

Section: Methods For Generating Virtual Patientsmentioning

confidence: 99%

“…Virtual clinical trials (VCTs) are growing in popularity as a tool for predicting and quantifying the uncertainty of the effects of therapy on disease progression [5,6]. VCTs combine data-validated mathematical models with computational techniques to enhance the efficiency, increase the success rates, and decrease the costs associated with the drug development process [7]. In this manuscript we adopt the VCT terminology used by Rieger and colleagues [8].…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Assessing the Role of Patient Generation Techniques in Virtual Clinical Trial Outcomes

Gevertz,

Wares

2024

Preprint

View full text Add to dashboard Cite

Virtual clinical trials (VCTs) are growing in popularity as a tool for quantitatively predicting heterogeneous treatment responses across a population. In the context of a VCT, a plausible patient is an instance of a mathematical model with parameter (or attribute) values chosen to reflect features of the disease and response to treatment for that particular patient. A number of techniques have been introduced to determine the set of model parametrizations to include in a virtual patient cohort. These methodologies generally start with a prior distribution for each model parameter and utilize some criteria to determine whether a parameter set sampled from the priors should be included or excluded from the plausible population. No standard technique exists, however, for generating these prior distributions and choosing the inclusion/exclusion criteria. In this work, we rigorously quantify the impact that VCT design choices have on VCT predictions. Rather than use real data and a complex mathematical model, a spatial model of radiotherapy is used to generate simulated patient data and the mathematical model used to describe the patient data is a two-parameter ordinary differential equations model. This controlled setup allows us to isolate the impact of both the prior distribution and the inclusion/exclusion criteria on both the heterogeneity of plausible populations and on predicted treatment response. We find that the prior distribution, rather than the inclusion/exclusion criteria, has a larger impact on the heterogeneity of the plausible population. Yet, the percent of treatment responders in the plausible population was more sensitive to the inclusion/exclusion criteria utilized. This foundational understanding of the role of virtual clinical trial design should help inform the development of future VCTs that use more complex models and real data.

show abstract