Multi-target reconstruction of fluorescence molecular tomography based on blind source separation

Abstract

Fluorescence molecular tomography (FMT) is an important molecular imaging technique for tumor detection in early stage. In realistic research, due to the spread and metastasis of tumor cells, it is necessary to comprehensively analyze multiple tumor regions for cancer staging studies. Therefore, high-precision multi-light source reconstruction results are required for quantitative analysis in FMT research. However, the existing methods perform well in the reconstruction of single fluorescent source but may fail in reconstructing multiple targets, which is an obstacle for FMT practical application. In this paper, we proposed a multi-target reconstruction strategy for Fluorescence Molecular Tomography based on Blind Source Separation (BSS) by converting multi-target reconstructions into multiple single-target reconstructions. It is a breakthrough work in multi-target reconstruction for cancer staging using optical molecular technique. Numerical simulation experiments proved that it had the ability of multi-source resolution for FMT in accurate location and morphology recovery. The encouraging results demonstrate significant effectiveness and potential of our method for preclinical FMT applications.