Modeling the photon propagation for optical molecular imaging

描述（由申请人提供）：基于荧光和生物发光的光学分子断层扫描已成为生物医学成像的主要领域。它有助于定位和量化人类疾病的小动物模型中的分子和细胞特征，并有助于监测病理变化，评估治疗反应，促进药物研究和开发。光学分子成像通过生物组织捕获扩散光子。光子在这种介质中的输运主要表现为吸收和散射。辐射输运方程在理论上是准确的，但在计算上是不切实际的。生物医学光学中最流行的模型是扩散近似，它近似描述了光子与生物组织的相互作用，但它只在某些条件下才能很好地工作，例如在高散射和弱吸收介质中。由于扩散近似与物理现实之间的不匹配，这种限制显著地损害了多光谱光学分子成像。例如，生物发光成像涉及光谱范围[400 nm，600 nm]，在该光谱范围内，固体小鼠器官中存在相对较大的吸收，使得扩散近似相当不准确。此外，扩散近似在荧光分子成像中使用的高频模式中是有问题的。在求解不适定光学层析成像问题时，模型失配对成像质量的影响最大。目前，还没有完善的光子传播模型用于光学分子成像。我们的总体目标是发展一个相位近似模型，以取代光学分子成像的扩散近似模型。新模型是基于广义相函数，使扩散近似的一个特殊情况下，数学等价的辐射传输方程和计算可比的扩散近似模型。我们的初步数据表明，我们的相位近似工程作为准确的辐射输运方程在广泛的生物相关的光学参数。具体目标是（1）分别在稳态、频率和时间分辨模式下系统地建立关于光子注量率和通量矢量的相位近似模型，（2）开发基于相位近似的实用算法来描述光子在小鼠解剖结构中的传输，（3）进行数值模拟和体模实验，以评估和验证具有强吸收体、近光源、跨边界和高频模式的基于相位近似的算法。本项目完成后，在任何一组生物相关光学参数上，从相位近似模型获得的光子注量率和通量矢量的相对误差与蒙特卡罗数据和实验数据相比将被验证为<5%，并且对于λ <10，与扩散近似模型相比，精度提高>30%。相近似模型的计算成本将被评估为比扩散近似模型的计算成本增加<20%。所提出的技术将显着提高光学分子成像的各种临床前成像应用。在这个项目中，我们将开发一种新的光子传输模型-相位近似模型，以取代几十年来在生物光子学领域流行的扩散近似模型，并专门用于光学分子断层扫描。在广泛的生物相关的光学参数，新的模型将产生的结果始终比扩散近似模型更准确的计算成本可比的扩散近似对应。所提出的技术将显着提高光学生物医学成像的各种临床前的应用。

DESCRIPTION (provided by applicant): Optical molecular tomography based on fluorescence and bioluminescence has emerged as a major area of biomedical imaging. It is instrumental for localizing and quantifying molecular and cellular features in small animal models of human diseases, and helps monitor pathological changes, evaluate therapeutic responses, and facilitate drug research and development. Optical molecular imaging captures diffusive photons through the biological tissue. The photon transport in such media is mainly characterized by absorption and scattering. The radiative transport equation is theoretically accurate but computationally impractical. The most popular model for biomedical optics is the diffusion approximation that approximately describes the interaction of photons with the biological tissue but it only works well under certain conditions such as in highly scattering and weakly absorbing media. This limitation significantly compromises multi- spectral optical molecular imaging due to the mismatch between the diffusion approximation and the physical reality. For example, bioluminescence imaging involves a spectral range [400nm, 600nm] over which there is relatively large absorption in solid mouse organs, rendering the diffusion approximation quite inaccurate. Moreover, the diffusion approximation is problematic in the high frequency mode as used in fluorescence molecular imaging. The model mismatch would have the most adverse effect on image quality in solving ill-posed optical tomography problems. Currently, there is no photon propagation model that is well-rounded for optical molecular imaging. Our overall goal of this project is to develop a phase approximation model to replace the diffusion approximation model for optical molecular imaging. The new model is based on a generalized phase function, make the diffusion approximation a special case, mathematically equivalent to the radiative transport equation and computationally comparable to the diffusion approximation model. Our preliminary data show that our phase approximation works as accurately as the radiative transport equation over a broad range of biologically relevant optical parameters. The specific aims are to (1) formulate the phase approximation model systematically with respect to the photon fluence rate and flux vector in the steady-state, frequency and time-resolved modes respectively, (2) develop practical algorithms based on the phase approximation to describe the photon transport in the mouse anatomy, (3) perform numerical simulation and phantom experiments to evaluate and validate the phase approximation based algorithms with strong absorbers, near light sources, across boundaries, and in the high frequency mode. Upon completion of this project, the relative errors of the photon fluence rate and flux vector obtained from the phase approximation model will have been validated as <5% as compared with the Monte Carlo data and experimental data over any set of biologically relevant optical parameters, and >30% accuracy improvement been made against the diffusion approximation model for albedo <10. The computational cost of the phase approximation model will have been evaluated as <20% increment than that of the diffusion approximation model. The proposed techniques will significantly enhance optical molecular imaging for a wide variety of pre-clinical imaging applications. In this project, we will develop a novel photon transport model - the phase approximation model to replace the diffusion approximation model that has been popular in the bio-photonics field for decades and exclusively used for optical molecular tomography. Over a broad range of biologically relevant optical parameters, the new model will produce results consistently more accurate than the diffusion approximation model at a computational cost comparable to the diffusion approximation counterpart. The proposed techniques will significantly improve optical biomedical imaging for a wide variety of pre-clinical applications.