Cloud quantitative imaging for whole-body tumor burden in neurofibromatoses

项目摘要/摘要 神经纤维瘤病是一组常染色体显性遗传病，包括NF1、NF2和神经鞘瘤病 神经遗传性疾病的特点是几乎100%的患者倾向于发展多条神经 鞘肿瘤。肿瘤负荷的测定在磁共振成像(MRI)中是必不可少的 对核因子患者的纵向管理，包括对肿瘤监测的终身随访 进展和治疗反应的评估。然而，核因子肿瘤的体积量化不是一种 临床常规，因为在准确和可重复的分割方面存在技术挑战 MRI表现为高度不规则和浸润性的神经纤维瘤，尤其是丛状神经纤维瘤。 该STTR项目的目标是为NF软件开发云定量成像(CQI)，表示为 CQI-NF，它将为肿瘤的体积量化提供技术和临床服务 通过“虚拟”(云计算)技术实现全身和区域核磁共振。在此开发的产品 STTR将为全国和世界各地的NF临床社区提供对这项技术的访问，而不需要 现场维护先进的体积成像分析软件和硬件的成本过高。这个项目 将建立在软件平台上开发的现有体积成像分析技术的基础上 马萨诸塞州综合医院(MGH)3D成像实验室的3DQI，将使用200进行评估 从世界各地的神经纤维界收集的纵向全身和区域性核磁共振病例。 本二期工程的具体目标是：(1)CQI-NF系统的开发：我们将继续开发 在我们的第一阶段项目中原型的CQI-NF系统通过以下方式提高分割的准确性和效率 多服务器平台并行动态阈值水平集(DT Level Set)，结合DT深度学习 用于分割肿瘤的水平集，并将CQI-NF系统从3DQI/NF转换为私有 云平台(如TeraRecon的直觉云)，用于提供使用 NF社区中的软件即服务(SaaS)模式。(2)CQI-NF系统评估：我们将进行 CQI-NF系统纵向定位准确性和重复性的回顾性临床研究 对MGH癌症中心收集的200名NF患者和我们在该中心的全球临床合作者进行监测 社区。(3)准备FDA 510(K)许可文件：我们将建立质量管理 CQI-NF符合FDA法规的体系，并准备FDA 510(K)批准所需的文件 提交CQI-NF临床翻译的长期项目目标。 建议的CQI-NF系统的成功开发和验证将对 核因子社区通过为核磁共振上的核因子体积成像分析提供云计算基础设施， 这在当前的临床常规中是不可用的，从而导致纵向上的实质性进展 神经营养不良患者的管理。

Project Summary / Abstract The neurofibromatoses (NFs), including NF1, NF2, and schwannomatosis, are a group of autosomal-dominant neurogenetic disorders characterized by a predisposition in virtually 100% of patients to develop multiple nerve sheath tumors. The determination of tumor burden on magnetic resonance imaging (MRI) is indispensable for the longitudinal management of NF patients, which includes life-long follow-up for the monitoring of tumor progression and the assessment of treatment response. However, volumetric quantification of NF tumors is not a clinical routine because of the technical challenges in the accurate and reproducible segmentation of highly-irregular and infiltrating NF tumors, in particular plexiform neurofibromas on MRI. The goal of this STTR project is to develop cloud quantitative imaging (CQI) for NF software, denoted as CQI-NF, which will provide the technical and clinical service for volumetric quantification of NF tumors on whole-body and regional MRI via "virtualization" (cloud computing) technology. The product developed in this STTR will provide access to this technology for the NF clinical community nationwide and worldwide without the excessive cost to maintain on-site advanced volumetric imaging analysis software and hardware. This project will be built upon existing technologies for volumetric imaging analysis developed on the software platform “3DQI” in the 3D Imaging Lab at Massachusetts General Hospital (MGH), and will be evaluated using 200 longitudinal whole-body and regional MRI cases collected from the NF community worldwide. The specific aims of this Phase II project are: (1) Development of CQI-NF system: We will continue to develop the CQI-NF system prototyped in our Phase I project to improve the accuracy and efficiency of segmentation by paralleling dynamic-threshold level set (DT level set) in multi-server platform, combining deep-learning in DT level set for the segmentation of NF tumors, and to translate the CQI-NF system from 3DQI/NF into a private cloud platform (such as TeraRecon's iNtuition CLOUD) for the provision of volumetric quantification of NF using a software-as-a-service (SaaS) model in the NF community. (2) Evaluation of CQI-NF system: We will conduct a retrospective clinical study to evaluate the accuracy and reproducibility of the CQI-NF system in the longitudinal monitoring of 200 NF patients collected at MGH Cancer Center and our clinical collaborators worldwide in the NF community. (3) Preparation of FDA 510(k) clearance submission: We will establish the quality management system for CQI-NF to meet FDA regulation, and prepare the required documentation for FDA 510(k) clearance submission for the long-term project goal of clinical translation of CQI-NF. The successful development and validation of the proposed CQI-NF system will have a high clinical impact in the NF community by providing a cloud-computing infrastructure for the volumetric imaging analysis of NF on MRI, which is not available in current clinical routine, thereby leading to a substantial advance in the longitudinal management of NF patients.