Malyarenko Dariya, Fedorov Andriy, Bell Laura, Prah Melissa, Hectors Stefanie, Arlinghaus Lori, Muzi Mark, Solaiyappan Meiyappan, Jacobs Michael, Fung Maggie, Shukla-Dave Amita, McManus Kevin, Boss Michael, Taouli Bachir, Yankeelov Thomas E, Quarles Christopher Chad, Schmainda Kathleen, Chenevert Thomas L, Newitt David C
University of Michigan, Radiology, Ann Arbor, Michigan, United States.
Brigham and Women's Hospital and Harvard Medical School, Boston, Massachusetts, United States.
J Med Imaging (Bellingham). 2018 Jan;5(1):011006. doi: 10.1117/1.JMI.5.1.011006. Epub 2017 Oct 30.
This paper reports on results of a multisite collaborative project launched by the MRI subgroup of Quantitative Imaging Network to assess current capability and provide future guidelines for generating a standard parametric diffusion map Digital Imaging and Communication in Medicine (DICOM) in clinical trials that utilize quantitative diffusion-weighted imaging (DWI). Participating sites used a multivendor DWI DICOM dataset of a single phantom to generate parametric maps (PMs) of the apparent diffusion coefficient (ADC) based on two models. The results were evaluated for numerical consistency among models and true phantom ADC values, as well as for consistency of metadata with attributes required by the DICOM standards. This analysis identified missing metadata descriptive of the sources for detected numerical discrepancies among ADC models. Instead of the DICOM PM object, all sites stored ADC maps as DICOM MR objects, generally lacking designated attributes and coded terms for quantitative DWI modeling. Source-image reference, model parameters, ADC units and scale, deemed important for numerical consistency, were either missing or stored using nonstandard conventions. Guided by the identified limitations, the DICOM PM standard has been amended to include coded terms for the relevant diffusion models. Open-source software has been developed to support conversion of site-specific formats into the standard representation.
本文报告了定量成像网络磁共振成像(MRI)子组发起的一个多中心合作项目的结果,该项目旨在评估当前能力,并为在利用定量扩散加权成像(DWI)的临床试验中生成标准参数扩散图(数字医学成像和通信,DICOM)提供未来指导方针。参与的站点使用单个体模的多厂商DWI DICOM数据集,基于两种模型生成表观扩散系数(ADC)的参数图(PMs)。对结果进行了评估,以检查模型之间以及与体模真实ADC值的数值一致性,以及元数据与DICOM标准要求的属性的一致性。该分析确定了ADC模型之间检测到的数值差异来源的描述性元数据缺失。所有站点都未存储DICOM PM对象,而是将ADC图存储为DICOM MR对象,通常缺少用于定量DWI建模的指定属性和编码术语。对于数值一致性很重要的源图像参考、模型参数、ADC单位和比例,要么缺失,要么使用非标准约定存储。在确定的局限性的指导下,DICOM PM标准已进行修订,以纳入相关扩散模型的编码术语。已开发开源软件以支持将特定于站点的格式转换为标准表示形式。