Alternative approach to estimate lumped constant in the deoxyglucose model: simulation and validation

Area of Science:

• Neuroscience
• Radiochemistry
• Medical Imaging

Background:

• The lumped constant (LC) is crucial for quantitative analysis in positron emission tomography (PET) using deoxyglucose tracers.
• Accurate LC estimation is vital for reliable measurements of brain glucose metabolism.

Purpose of the Study:

• To develop and validate an alternative method for estimating the lumped constant (LC) in the deoxyglucose model.
• To compare the proposed method with the conventional approach for LC estimation.

Main Methods:

• A nonlinear least-squares (NLSQ) method was employed, utilizing data from the initial 10 minutes post-tracer injection.
• The method was evaluated through computer simulations with varying noise levels and input functions.
• The new technique was applied to measure whole brain LC and rate constants in cats using 2-[18F]fluoro-2-deoxy-D-glucose (2-[18F]FDG).

Main Results:

• The proposed NLSQ method provides more reliable LC estimates compared to the conventional method.
• The new technique allows for shorter experimental periods and accommodates various input function shapes.
• Measured whole brain LC in cats was 0.443 ± 0.012, with k2* = 0.124 ± 0.009 and k3* = 0.024 ± 0.001.

Conclusions:

• The developed NLSQ method offers a more robust and efficient approach to estimating the lumped constant (LC).
• This advancement can improve the accuracy and reduce the time required for PET studies of brain metabolism.
• The findings support the utility of this new technique for quantitative PET imaging in neuroscience research.