Mathematical Sciences: Massively Parallel Monte Carlo Simulation of Positron Emission Tomography for Scatter Correction

数学科学：用于散射校正的正电子发射断层扫描的大规模并行蒙特卡罗模拟

• 批准号: 9318760
• 负责人: Jill Mesirov
• 金额: $ 4.58万
• 依托单位: Thinking Machines Corporation
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 1994
• 资助国家: 美国
• 起止时间: 1994-09-01 至 1996-08-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=9318760&HistoricalAwards=false
• 关键词: Mathematical; Sciences; Massively; Parallel; Monte

Scattered photons blur positron emission tomography images, degrading image resolution, contrast, and quantitative accuracy. No general correction for scattered photons has been found, because the spatial distribution of scattered photons has yet to be characterized. The investigators study photon scattering based on a state-of-the-art Monte Carlo computer program, developed at Los Alamos National Laboratory, that can provide more information than is readily measurable experimentally. At present, this powerful analysis technique requires enormous computational resources. The investigators evaluate the advantages of using a massively parallel supercomputer to conduct the computer experiments and also examine methods for improving the computational efficiency. Positron emission tomography (PET) is an emerging technology with unique capability for imaging biochemical processes in the human body. The motivation for this research is that by accounting for scattered photons, the diagnostic and research value of PET data is improved. Scattered photons is a significant problem as 10 to 25 percent of detected photons are deflected from their straight path prior to detection. Moreover, the fraction of photons is even greater, 40 to 50 percent, with the latest generation of PET scanners that use three-dimensional imaging capability to scan the entire body for cancer detection. Even as such, PET is more sensitive for cancer detection and can detect cancer earlier than computed tomography (CT) and magnetic resonance imaging (MRI). PET images can show the vast changes in sugar metabolism associated with cancerous tumors, whereas the tumors must grow to a size sufficent to significantly alter the normal tissue structure in order to be visualized by CT and MRI. More generally, a proper scatter correction will benefit PET's ability to detect and quantify the biochemical changes associated with numerous diseases and particularly those of the heart and brain.

散射光子使正电子发射断层扫描图像模糊，降低图像分辨率、对比度和定量准确性。 由于散射光子的空间分布还没有被确定，因此没有发现散射光子的一般校正。 研究人员基于洛斯阿拉莫斯国家实验室开发的最先进的蒙特卡罗计算机程序研究光子散射，该程序可以提供比实验上容易测量的更多的信息。 目前，这种强大的分析技术需要大量的计算资源。 研究人员评估了使用大规模并行超级计算机进行计算机实验的优势，并研究了提高计算效率的方法。 正电子发射断层扫描（PET）是一种新兴的技术，具有独特的成像能力，在人体内的生化过程。 这项研究的动机是通过考虑散射光子，提高PET数据的诊断和研究价值。 散射光子是一个重要的问题，因为探测到的光子中有10%到25%在探测之前偏离了它们的直线路径。 此外，光子的比例甚至更高，达到40%至50%，最新一代的PET扫描仪使用三维成像能力扫描整个身体以检测癌症。 即使如此，PET对癌症检测更敏感，并且可以比计算机断层扫描（CT）和磁共振成像（MRI）更早地检测癌症。 PET图像可以显示与癌性肿瘤相关的糖代谢的巨大变化，而肿瘤必须生长到足以显著改变正常组织结构的大小，以便通过CT和MRI可视化。 更一般地说，适当的散射校正将有利于PET检测和量化与许多疾病相关的生化变化的能力，特别是心脏和大脑的变化。