Multiplexed fluorescence tomography with spectral and temporal data: demixing with intrinsic regularization.

Area of Science:

• Biomedical Optics
• Molecular Imaging
• Computational Biology

Background:

• Multiplexed fluorescence molecular tomography faces challenges due to limited near-infrared windows and signal distortion during tissue propagation.
• Accurate imaging of multiple fluorescent targets requires effective demixing of overlapping spectral and temporal signals.
• Current methods struggle with the complexity and data volume inherent in high-throughput tissue imaging.

Purpose of the Study:

• To develop and evaluate a novel method for simultaneous demixing of multiple fluorescent targets using combined spectral and temporal data.
• To improve the performance of fluorescence molecular tomography for high-throughput applications in bulk tissue.
• To address the limitations of existing modalities by enhancing signal demixing capabilities.

Main Methods:

• A two-stage algorithm was proposed, decoupling demixing and tomographic reconstruction for efficient data handling.
• The demixing stage incorporates principles from sparse subspace clustering and compressed sensing.
• The method was evaluated using simulations without requiring a regularization parameter.

Main Results:

• Simulations demonstrated successful simultaneous demixing of four spectrally and temporally overlapping fluorophores.
• Achieved a root-mean-square error of less than 3% per fluorophore in a 25 mm cross-sectional area.
• Analyzed the method's sensitivity to model mismatch, confirming its robustness.

Conclusions:

• Joint spectral-temporal data analysis offers a significant performance improvement for multiplexed fluorescence molecular tomography.
• The proposed demixing approach is effective for complex biological samples with closely overlapping fluorophore signals.
• This technique holds promise for advancing high-throughput molecular imaging in biomedical research.