面向小动物模型的功能PAT多生理参数定量成像方法

This project intends to develop the functional photoacoustic tomography (f-PAT) to achieve quantitative imaging of multi-physiological parameters of the small animal model with high spatial resolution, focusing on three important aspects in f-PAT technology: acoustic inversion, optical inversion and physiological parameters calculation. To improve the reconstruction quality of the initial pressure image under the limited-view measurement condition, the deep neural network driven by the PAT reconstruction model is constructed. To realize the quantitative reconstruction of the optical absorption coefficient distribution in the target region, the high-order approximation of radiative transfer equation, which is more suitable for describing photon propagation process in small animals, is introduced. Meanwhile, a more accurate nonlinear iteration strategy for calculating the optical absorption perturbations is proposed. Furthermore, to achieve effective estimation of multiple physiological function parameters in vivo, the multispectral unmixing algorithm based on physiological prior information constraint and an advanced method for kinetic parameters calculation are introduced. In summary, this research can provide the theoretical and methodological support for the study of physiological and pathological mechanisms of biological tissues and the efficacy assessment of antineoplastic drugs.

本项目拟发展体系化的功能光声层析成像（f-PAT）技术，以实现小动物模型的多生理功能参数分布在体高空间分辨率定量成像为目标，围绕f-PAT技术中紧密连接的声学求逆问题、光学求逆问题以及生理参数求逆问题三个重要环节展开研究。构建PAT重建模型驱动的深度神经网络架构，以提高有限角度测量条件下目标区域初始声压图像的重建质量；选择更适合小动物模型的光子输运高阶建模，提出更准确的微扰量非线性迭代求解策略，实现对目标区域光学吸收系数分布的定量重建；研究基于生理先验信息约束的多光谱组分离析算法和动力学参数估计方法，最终实现多生理功能参数定量求解，为生物组织生理与病理学机理研究以及抗肿瘤药物疗效评估提供理论和方法支持。

为有效利用光声断层成像（PAT）实现生物组织生理与病理学机理、抗肿瘤药物疗效的精准可视化评估，本项目发展了体系化的功能光声断层成像（f-PAT）技术。面向f-PAT技术中“声学初始声压重建问题”，构建了“Transformer-卷积”混合深度重建网络，实现了稀疏角度测量条件下初始声压的精准重建；发展了三维PAT加速重建方法，实现了重建图像轴向分辨率的提升以及三维重建效率的提高。接着针对“光学吸收参数定量问题”，发展了结合光子输运高阶模型与非线性迭代更新的定量PAT（q-PAT）方法，对具有中小尺寸复杂结构且呈现大吸收或小散射的生物组织实现了区域光学参数分布的准确定量。在前两项研究基础上，进而瞄准“光谱学功能生色团解混问题”，提出了一种结合空间正则化与半监督光谱解混的多光谱PAT（MSPAT）重建方法，最终实现了对外源性探针光谱和各生色团浓度图的同步优化重建。本项目实现了上述三个研究成果的连通应用，生物学在体成像实验表明上述f-PAT方法的有效性，团队正在积极推动该项技术的临床前应用。

内容获取失败，请点击重试