An Inversion Scheme for Hybrid Fluorescence Molecular Tomography Using a Fuzzy Inference System

Area of Science:

• Medical Imaging
• Biomedical Engineering
• Optical Imaging

Background:

• Fluorescence molecular tomography (FMT) imaging performance is enhanced by incorporating anatomical information as priors.
• Hybrid FMT systems coupled with X-ray CT or MRI are emerging to leverage anatomical priors.
• Methodological approaches for integrating anatomical priors into FMT inversion are critical.

Purpose of the Study:

• To develop a new method for utilizing anatomical prior information in FMT.
• To preferentially minimize residual errors in FMT reconstruction based on anatomical segment relevance.
• To improve the accuracy and reduce bias in hybrid FMT reconstructions.

Main Methods:

• A weighted least square (WLS) approach was employed for preferential error minimization.
• A Mamdani-type fuzzy inference system was used to optimize weights in the WLS approach.
• The method was implemented as a two-step structured regularization approach and validated experimentally.

Main Results:

• The proposed method demonstrated accurate performance in phantom, ex vivo, and in vivo animal studies.
• Reconstruction bias was significantly minimized.
• The approach eliminated the need for user input in setting regularization parameters.

Conclusions:

• The novel method effectively incorporates anatomical priors into FMT.
• This advancement facilitates the realization of the full potential of hybrid FMT systems.
• The technique offers improved accuracy and reduced bias in FMT imaging without manual parameter tuning.