PDF(Pubmed)
Abstract：
CT protocol management is an arduous task that requires expertise from a variety of radiology professionals, including technologists, radiologists, radiology IT professionals, and medical physicists. Each CT vendor has unique, proprietary protocol file structures, some of which may vary by scanner model, making it difficult to develop a universal framework for distilling technical parameters to a human-readable file format. An ideal solution for CT protocol management is to minimize the work required for parameter extraction by introducing a data format into the workflow that is universal to all CT scanners. In this paper, we report a framework for CT protocol management that converts raw protocol files to an intermediary format before outputting them in a human-readable format for a variety of practical clinical applications, including routine protocol review, protocol version tracking, and cross-protocol comparisons. The framework was developed in Python 3. Technical parameters of interest were determined via collaborative effort between medical physicists and lead technologists. Protocol files were extracted and analyzed from a variety of scanners across our hospital-wide CT fleet, including various systems from Siemens and GE. Protocols were subcategorized based on relevant technical parameters into regular, dual-energy, and cardiac CT protocols. Backend code for technical parameter extraction from raw protocol files to a JavaScript Object Notation (JSON) format was performed on a per-system basis. Conversion from JSON to a readable output format (MS Excel) was performed identically for all scanners using the universal framework developed and presented in this work. Example results for Siemens and GE scanners are shown, including side-by-side comparisons for protocols with similar clinical indications. In conclusion, our CT protocol management framework may be deployed on any CT system to improve clinical efficiency in protocol review and upkeep.
摘要：
CT协议管理是一项艰巨的任务，需要各种放射学专业人员的专业知识，包括技术人员,放射科医生,放射学IT专业人员，和医学物理学家。每个CT供应商都有独特的，专有协议文件结构，其中一些可能因扫描仪型号而异，这使得很难开发一个通用框架来将技术参数提取为人类可读的文件格式。CT协议管理的理想解决方案是通过在所有CT扫描仪通用的工作流程中引入数据格式来最大程度地减少参数提取所需的工作。在本文中,我们报告了一个用于CT协议管理的框架，该框架将原始协议文件转换为中间格式，然后以人类可读的格式输出它们以用于各种实际临床应用。包括常规方案审查，协议版本跟踪，和跨协议比较。该框架是在Python3中开发的。感兴趣的技术参数是通过医学物理学家和首席技术人员之间的协作努力确定的。协议文件从我们医院范围内的CT车队的各种扫描仪中提取和分析，包括西门子和GE的各种系统。协议根据相关技术参数被细分为常规，双能量,和心脏CT协议。从原始协议文件到JavaScript对象表示法(JSON)格式的技术参数提取的后端代码是在每个系统的基础上执行的。使用本工作中开发和呈现的通用框架，对所有扫描仪执行从JSON到可读输出格式（MSExcel）的转换。显示了西门子和GE扫描仪的示例结果，包括具有相似临床适应症的方案的并排比较。总之,我们的CT方案管理框架可以部署在任何CT系统上,以提高方案审查和维护的临床效率.
