L(p) regularization for early gate fluorescence molecular tomography.

Area of Science:

• Biomedical Optics
• Medical Imaging
• Computational Biology

Background:

• Time domain fluorescence molecular tomography (TD-FMT) offers high resolution and quantification capabilities.
• Early time gates in TD-FMT data are crucial for spatial resolution but suffer from low signal-to-noise ratios.
• Image reconstruction in TD-FMT is an ill-posed inverse problem, sensitive to noise.

Purpose of the Study:

• To develop and evaluate an L(p) regularization-based algorithm for improving TD-FMT image reconstruction using early time gates.
• To compare the performance of different L(p) regularization parameters (p∈{1/16,1/8,1/4,1/3,1/2,1,2}) for early time-gate reconstruction.
• To assess the algorithm's efficacy in both simulated and experimental preclinical data.

Main Methods:

• Implementation of an L(p) regularization algorithm within a wide-field mesh-based Monte Carlo simulation framework.
• Systematic comparison of reconstruction performance using various L(p) norms (p values) on a synthetic murine model and preclinical data.
• Validation using a 3D mouse atlas and data from a mouse experiment.

Main Results:

• The L(1/4) regularization method demonstrated superior performance for early time-gate reconstructions compared to other tested p values.
• The developed algorithm effectively addressed the ill-posed nature of the TD-FMT inverse problem, particularly for low signal data.
• Consistent results were observed across both simulated data and experimental preclinical data.

Conclusions:

• L(1/4) regularization is an effective strategy for enhancing spatial resolution in time domain fluorescence molecular tomography using early time gates.
• The proposed method offers a robust solution for overcoming noise and instability issues in TD-FMT image reconstruction.
• This advancement has significant implications for accurate preclinical imaging and molecular quantification.