Demonstration of PBK model evaluations: Case studies in fish species and farm animals

Abstract

tions that do not require in vivo data for model calibration and evaluation. Traditionally, the evaluation of a model's predictive capability relies heavily on the concordance between model simulations and in vivo data from laboratory animals. However, the availability of in vivo data is limited to a few well-studied chemicals, and the traditional approach for model assessment cannot meet the new demands of using more non-animal data to assess health risks posed by chemical exposures. The new OECD guidance presents a model assessment framework to address the need to evaluate the confidence in using a PBK model to support a specific regulatory application when in vivo data are unavailable. This framework includes two categories of considerations , "context and implementations" and "model validity". Context and implementations consider the regulatory application and context of use, which together determine the degree of acceptable uncertainty. Model validity examines confidence and reliability in biological basis, theoretical basis of model equations, relevance of input parameters, sensitivity of output to parameters, and, when data are available, goodness of fit. Under this framework, two complementary tools are provided to modelers and users. The first tool is a model reporting template , which prescribes elements that should be documented to enable more efficient review. The second tool is an evaluation checklist for users to identify specific data or information needed before a model can be accepted, or reasons for rejecting a model submission. This presentation will give an overview of the model assessment framework and provide examples to demonstrate the use of both tools. Animal and environmental risk assessment (RA) of chemicals aims to protect species from adverse health effects as a result of chemical exposure and is a key regulatory and scientific research field. Over the last few years, physiologically-based kinetic models (PBK) for animal species have shown to play a crucial role in the development of next generation RA (NGRA) approaches such as the prediction of plasma or target organ concentrations in these species. This presentation focuses on a demonstration of the evaluation of PBK models developed for fish species (rainbow trout, fathead minnow, zebra fish and three-spined stickleback) and farm animals (cattle, swine, sheep and chicken) in the open-source freeware R. These case studies have been developed under collaborative research projects between the European food safety Authority (EFSA), academia and national authorities. Here, the principles laid out in the OECD guidance on the use of PBK-models in risk assessment are applied systematically to the evaluation of the models. Applications of the PBK models focuses specifically on prediction of blood and organ concentrations and predictions are compared with experimental data and visualised via gg plots. Model evaluations includes a number of steps. The first step is problem formulation providing the scope and purpose of the model. Then other steps include model conceptualisation (structure and mathemat-CEC02-03 PBK modelling workflow and in vitro methods for model parameterisation A. Paini This lecture will introduce the six-step PBK modelling workflow with an emphasis on tools/methods available to parameterise PBK models in the absence of sufficient in vivo data for calibrating the model or assessing its predictive capacity. The workflow allows refining the model based on amendments from new information (new in vitro data, new biological knowledge, etc). After, the introduction to the OECD PBK model workflow, the talk will highlight the battery of in vitro methods to estimate model parameters that describe kinetic processes, such as absorption across external and internal physiological barriers, distribution by partitioning between tissues and blood, active transportation , and systemic clearance (such as metabolite formation and hepat-ic/renal excretion). Since the stability of the test chemicals in the culture medium may affect the cellular response, this lecture will introduce several modelling approaches for extrapolating among the in vitro nominal concentration, free concentration, and effective intracel-lular concentration (OECD GD, Chapter 2, step 3). Giving a flavour of the available in vitro technologies and limitations and uncertainties. CEC02-04 PBK modelling assessment-uncertainty and sensitivity analysis characterisation Physiologically based pharmacokinetic (PBK) models have a potentially significant role in the development of a reliable predictive toxicity testing strategy. The structure of PBK models are ideal frameworks into which disparate in vitro and in vivo data can be integrated and utilized to translate information generated, using alternative to animal measures of toxicity, known as New Approach Methods (NAMs) and human biological monitoring data, into plausible corresponding exposures. However, these models invariably include a description of well-known non-linear biological processes such as, enzyme saturation and interactions between parameters such as, organ mass and body mass and cardiac output, work rate and organ and tissue blood flow. Therefore, an appropriate sensitivity analysis (SA) technique is required which can quantify the influences and uncertainties associated with individual parameters, interactions between parameters and any non-linear processes. The elements of a workflow for SA of PBK models that is computationally feasible, accounts for interactions between parameters with the results displayed in an intuitive form appropriate for toxicol-ogists, risk assessors, and regulators will be presented.