Advanced Software for MRI, PET, SPECT and CT Image Analysis

项目摘要 这个U24项目将有助于传播FireVoxel，一个免费提供的，用户友好的软件包处理 并从动态MRI、CT和PET图像中提取定量信息。该软件将帮助 放射科医生、肿瘤科医生和生物医学研究人员更好地利用所获得的数据。我们的重点是 动态成像和模型拟合。该软件是由一组科学家在过去15年中开发的。 在纽约大学医学院从事忙碌工作的几年中，它是一个成熟的产品， 功能工作流程，远远超出了原理验证阶段。它包含用于图像统计的广泛工具 （放射组学）、分割、共配准、运动校正和建模。FireVoxel已经在 近300篇论文发表。目前，地球仪约有1000名用户正在帮助 验证和改进程序。需要三年时间来编写一份全面的文件， 建立一个自动生产测试系统，以确保广泛传播所需的工业质量水平。 我们将建立一个全面的临床代表性病例数据库和相应的指标， 消除软件缺陷，并确保每次新软件版本后输出一致。库的地方 将以相关教程、基于云的文档和用户指南补充样本案例。这些 增加的内容将使该软件对广大用户具有吸引力。我们还将扩展FireVoxel的功能， 包括三个新的，全自动的措施：（a）一个可靠的动脉输入功能（AIF），用于动态- 对比MRI建模;（B）无AIF单肾功能（肾小球滤过率）;和（c）T1 rho参数 映射.

PROJECT SUMMARY This U24 project will help disseminate FireVoxel, a freely available, user-friendly software package to process and extract quantitative information form dynamic MRI, CT and PET images. The software will help radiologists, oncologist and biomedical researchers to make better use of acquired data. Our emphasis is on dynamic imaging and model fitting. The software has been developed by a team of scientists over the past 15 years in the context of a busy radiology practice at NYU School of Medicine. It is a mature product, with fully functional workflows, well beyond the proof-of-principle stage. It contains extensive tools for image statistics (radiomics), segmentation, coregistration, motion correction and modeling. FireVoxel has been already used in approximately 300 published studies. Approximately 1000 users across the globe are currently helping to validate and improve the program. Three years are needed to develop a comprehensive documentation and put together an automatic build-test system to assure industrial quality level needed for a wide dissemination. We will implement a comprehensive database of clinically representative cases and corresponding metrics to eliminate software defects and assure consistent output after each new software version. The depository of sample cases will be complemented by relevant tutorials, cloud-based documentation and user guides. These additions will make the software attractive to a broad user base. We will also extend FireVoxel's capabilities to include three novel, fully automatic measures: (a) a reliable arterial input function (AIF) for the use in dynamic- contrast MRI modeling; (b) AIF-free single-kidney function (glomerulal filtraction rate); and (c) T1rho parametric mapping.