Reconstruction based on adaptive group least angle regression for fluorescence molecular tomography

Area of Science:

• Biomedical Imaging
• Medical Physics
• Computational Biology

Background:

• Traditional fluorescence molecular tomography struggles with clustered tumor cells, limiting 3D reconstruction accuracy.
• Existing methods using tumor sparsity priors do not account for the spatial structure of tumor cell clusters.

Purpose of the Study:

• To develop and validate an improved reconstruction method for fluorescence molecular tomography.
• To enhance 3D tumor imaging by integrating local spatial structure and group sparsity.

Main Methods:

• Developed the adaptive group least angle regression elastic net (AGLEN) method.
• Integrated elastic net regularization with least angle regression, incorporating local spatial correlation and group sparsity.
• Employed iterative residual vector and median smoothing for robust optimization.

Main Results:

• AGLEN demonstrated superior performance over state-of-the-art methods in numerical simulations and in vivo mouse models (liver and melanoma tumors).
• The method showed improved accuracy across various light source configurations and distances, and robustness to Gaussian noise (5-25%).
• AGLEN accurately imaged tumor expression of cell death ligand-1, a potential biomarker for immunotherapy response.

Conclusions:

• The AGLEN method offers a significant advancement in 3D fluorescence molecular tomography reconstruction.
• This technique provides more accurate tumor imaging, crucial for guiding personalized cancer therapies like immunotherapy.