Model-based temporal unmixing towards quantitative photo-switching optoacoustic tomography

Area of Science:

• Biomedical Imaging
• Molecular Imaging
• Preclinical Research

Background:

• Optoacoustic (OA) imaging with photoswitchable proteins offers high-sensitivity, multiplexed imaging in live tissues.
• Existing methods face challenges in accurately unmixing signals from multiple reporter species and background noise.

Purpose of the Study:

• To develop a model-based variational framework for computationally unmixing and imaging photo-switching reporters using optoacoustic tomography.
• To improve the robustness and unmixing quality of optoacoustic imaging for complex biological samples.

Main Methods:

• A detailed mathematical model of the photo-switching mechanism, incorporating kinetic constants and light fluence.
• An iterative image-based inversion algorithm utilizing ℓ1 and total-variation regularization.
• Application to controlled phantoms and in vivo mice experiments.

Main Results:

• Successful disentanglement of multiple, spatially overlapping fluorescent labels.
• Recovery of continuous quantitative maps for different reporter species.
• Demonstrated robustness to noise and improved unmixing accuracy.

Conclusions:

• The proposed framework provides a powerful computational tool for advanced optoacoustic imaging.
• This method enables more precise multiplexed imaging of cells in preclinical research.
• The approach enhances the utility of photoswitchable proteins for in vivo molecular imaging.