Estimation of organ transport function: model-free deconvolution by recursive quadratic programming optimization

Area of Science:

• Physiology
• Biomedical Engineering
• Mathematical Modeling

Background:

• Accurate assessment of organ transit times is crucial for understanding physiological processes.
• Existing deconvolution methods often require assumptions about the transport function's form or initial values, limiting their applicability.

Purpose of the Study:

• To develop and validate a model-free deconvolution method for evaluating organ transit time frequency distribution functions.
• To assess the method's performance on both simulated and biological data.

Main Methods:

• A nonlinear constrained optimization problem formulation for deconvolution.
• Utilized a modified constrained variable metric approach with non-negativity constraints.
• Tested on mathematical systems with noisy input/output functions and applied to rabbit lung perfusion data.

Main Results:

• The method demonstrated stable estimates that closely matched known analytical transport functions in simulations.
• Reconvolution error and deviation from analytical functions were key performance metrics.
• Successful application to biological data from a perfused rabbit lung preparation.

Conclusions:

• The proposed model-free deconvolution method provides stable and accurate estimates of organ transit time distributions.
• It eliminates the need for pre-defined functional forms or initial guesses for transport functions.
• This approach enhances the analysis of physiological transport processes using biological data.