Department of Radiology, Stony Brook University, Stony Brook, New York, USA.
Department of Radiology, Boston Children's Hospital, Harvard Medical School, Boston, Massachusetts, USA.
J Appl Clin Med Phys. 2024 Apr;25(4):e14316. doi: 10.1002/acm2.14316. Epub 2024 Mar 11.
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.
CT 协议管理是一项艰巨的任务，需要来自各种放射学专业人员的专业知识，包括技术员、放射科医生、放射科 IT 专业人员和医学物理学家。每个 CT 供应商都有独特的、专有的协议文件结构，其中一些可能因扫描仪型号而异，因此很难为提取技术参数开发一个通用框架，以将其转换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工作流程中引入一种对所有 CT 扫描仪通用的数据格式，最大限度地减少提取参数所需的工作。在本文中，我们报告了一种 CT 协议管理框架，该框架在将原始协议文件转换为中间格式之前，将其输出为各种实际临床应用程序（包括常规协议审查、协议版本跟踪和跨协议比较）的人类可读格式。该框架是在 Python 3 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员之间的合作确定的。从我们医院范围内的 CT 机队中的各种扫描仪中提取和分析了协议文件，包括来自西门子和通用电气的各种系统。根据相关技术参数，将协议分为常规、双能和心脏 CT 协议。基于每个系统从原始协议文件中提取技术参数并将其转换为 JavaScript 对象表示法（JSON）格式的后端代码。使用在这项工作中开发和提出的通用框架，以相同的方式将 JSON 转换为可读的输出格式（MS Excel）。展示了来自西门子和通用电气扫描仪的示例结果，包括具有相似临床指征的协议的并排比较。总之，我们的 CT 协议管理框架可以部署在任何 CT 系统上，以提高协议审查和维护的临床效率。
