面向乳腺癌早期检测的高灵敏度荧光分子成像方法研究

乳腺内微小肿瘤的精准检测和定量评估，仍是当前乳腺癌临床诊断亟待解决的挑战性难题。本项目拟将荧光分子断层成像(Fluorescence Molecular Tomography, FMT)引入到乳腺癌早期检测和筛查中，开展基于光电倍增管(Photomultiplier Tube, PMT)的高灵敏度的荧光分子数据采集系统与高精度荧光靶点目标重建方法两个方面的研究，研制频率调制和空间编码稀疏采样的荧光分子数据采集系统和数据处理方案，实现对微小肿瘤的高灵敏检测，达到数据采集量与采集效率的平衡；研究基于深度学习的器官智能分割方法，为后续基于数据驱动的重建中训练样本的获取提供基础；研究基于多层感知机网络的三维重建方法，实现快速准确的荧光分子断层成像。本项目的研究成果有望扩展荧光分子断层成像在预临床研究中的应用，为肿瘤早期检测等提供新的影像学方法和技术手段。

Accurate detection and quantitative evaluation of intramamary micro-tumors is still challenging in clinical diagnosis of breast cancer. In this study we will apply fluorescence molecular tomography (FMT) technique for early detection and screening of breast cancer. We aim to develop both highly-sensitive fluorescent signal acquisition hardware system based on photomultiplier tube(PMT) and highly-precise fluorescent target reconstruction methods. To enable sensitive and fast detection of micro-tumors, we will construct the fluorescence acquisition system featuring frequency modulation and spatially-coded sampling scheme. Furthermore, we also aim to develop a novel data-driven reconstruction method, with the intelligent organ segmentation based on deep learning from training dataset. Utilizing a multi-layered perception network, we can achieve fast and accurate three-dimensional FMT reconstruction. We expect the project will facilitate application of FMT in preclinical studies, especially provide new imaging tools for early detection of breast tumors.