Noninvasive estimation of the input function for dynamic mouse 18F-FDG microPET studies.

Area of Science:

• Nuclear Medicine
• Biomedical Imaging
• Pharmacokinetics

Background:

• Accurate estimation of the plasma input function (IF) is crucial for quantitative analysis in dynamic positron emission tomography (PET) studies.
• Traditional methods for IF quantification often require invasive procedures or specialized equipment, limiting their application in small animal models.

Purpose of the Study:

• To develop and validate a novel noninvasive method for estimating the plasma time-activity curve (IF) in dynamic (18)F-FDG microPET mouse studies.
• To assess the accuracy of the proposed method by comparing it with the gold standard IF and evaluating its impact on tracer kinetic modeling parameters.

Main Methods:

• A four-step approach involving constraint nonnegative matrix factorization for segmentation, a seven-parameter mathematical model for IF, and a dual-output model accounting for partial-volume and spillover effects.
• Image-derived time-activity curves (TACs) from the left ventricle (Lv) and myocardium (Myo) were fitted to the dual-output model to estimate IF parameters.
• Validation involved comparing the estimated IF with the gold standard IF, considering delay and dispersion effects, using datasets from 20 mice.

Main Results:

• The proposed method demonstrated a low error of 7.237% ± 6.742% (r = 0.969) in the area under the curve between the estimated and gold standard IF.
• The error in the (18)F-FDG influx constant Ki for the myocardium was significantly low at 4.910% ± 6.810% (r = 0.992).

Conclusions:

• The developed noninvasive method is effective for estimating the plasma input function in dynamic (18)F-FDG microPET mouse studies.
• This technique offers a reliable alternative for accurate tracer kinetic modeling, reducing the need for invasive procedures and improving quantitative accuracy.