Exact parameter identification in PET pharmacokinetic modeling using the irreversible two tissue compartment model

Area of Science:

• Nuclear medicine
• Pharmacokinetics
• Mathematical modeling

Background:

• Quantitative dynamic positron emission tomography (PET) relies on tissue concentration and arterial input function for kinetic parameter estimation.
• Arterial input function is typically derived from blood sampling, a complex and invasive procedure.
• Mathematical analysis is lacking regarding the necessity of specific arterial blood measurements for kinetic parameter identification.

Purpose of the Study:

• To mathematically analyze the necessity of arterial blood measurements for kinetic parameter identification in dynamic PET.
• To determine if kinetic parameters can be identified without arterial input function measurements.
• To investigate the impact of noise on parameter identification.

Main Methods:

• Analytical approach using the irreversible two-tissue compartment model for dynamic PET data.
• Application of Tikhonov regularization to address noisy measurements.
• Numerical simulations to illustrate analytical findings in a synthetic example.

Main Results:

• Mathematical proofs demonstrate unique identification of all metabolic tissue parameters without arterial blood sampling.
• A consistency result shows stable reconstruction of ground-truth parameters in the vanishing noise limit.
• Numerical experiments suggest approximate kinetic parameter reconstruction is feasible with moderate noise.

Conclusions:

• Fully quantitative dynamic PET imaging is theoretically possible without arterial blood sampling for irreversible tracers.
• This analytical result simplifies the process of kinetic parameter estimation in dynamic PET.
• Eliminating blood sampling could reduce costs and complexity in clinical PET applications.