Toward Uncertainty-Aware Hemolysis Modeling: A Universal Approach to Address Experimental Variance.

Abstract

The purpose of this study is to address the lack of uncertainty quantification in numerical hemolysis models, which are critical for medical device evaluations. Specifically, we aim to develop a probabilistic hemolysis model, which incorporates experimental variability using the Markov Chain Monte Carlo (MCMC) method to enhance predictive accuracy and robustness. Initially, we examined the objective function landscape for fitting a Power Law hemolysis model, whose parameters are derived from inherently uncertain experimental data, by employing a grid search approach. Building on this, we applied MCMC to derive detailed stochastic distributions for the hemolysis Power Law model parameters C, α, and β. These distributions were then propagated through a reduced order model of the FDA benchmark pump to quantify the experimental uncertainty in hemolysis measurements with respect to the predicted pump hemolysis. Our analysis revealed a global flat minimum in the objective function landscape of the multi-parameter power law model, a phenomenon attributable to fundamental mathematical limitations in the fitting process. The probabilistic hemolysis model converged to a constant optimal C = 3.515 × 10^-5 and log normal distributions of α and β with means of 0.614 and 1.795, respectively. This probabilistic approach successfully captured both the mean and variance observed in the experimental FDA benchmark pump data. In comparison, conventional deterministic models are not able to describe experimental variation. Incorporating uncertainty quantification through MCMC enhances the robustness and predictive accuracy of hemolysis models. This method allows for better comparison of simulated hemolysis outcomes with in vitro experiments and can integrate additional datasets, potentially setting a new standard in hemolysis modeling.