Medical Physics Guidelines for Clinical Software, Scripts and Spreadsheets
Abstract
Purpose
Over the past two decades Radiation Therapy has become increasingly reliant on complex software systems, locally developed scripts and spreadsheets. Recent advances in AI technology suggest that this reliance will only increase in the future. While software engineers may take the lead in the development of commercial software, this is rarely the case for in-house scripts. Regardless of who wrote the software, Therapy Medical Physicists have the ultimate responsibility for the safe use of the software in the treatment of patients. With their combined knowledge of the physics embedded in the software and the processes involved in the use of that software, medical physicists are in the best position to develop guidelines for the management and QA of clinical radiation therapy software.
Methods
A series of literature review sessions was conducted which covered 59 articles including government documents, reports from professions organizations, journal articles, conference proceedings, book chapters, newsletter opinion pieces, web sites and unpublished works. In combination with the expert opinion within the group, 57 initial recommendations were distilled. A modified Delphi process was then used to achieve consensus-based quality guidelines for clinical software.
Results
The final recommendations were divided into four categories: “Planning”, “Development”, “Deployment”, and “Maintenance”. Within each category, recommendations were also ranked based on the risk and complexity of the software. Supporting guidance for risk assessment was also provided. In addition, the guidelines included a discussion on practical considerations such as incorporating external code sources and the use of AI during development.
Conclusion
The Canadian Organization of Medical Physicists working group has drafted guidelines to ensure the quality and safety of in-house developed and commercial software, its smooth deployment into the clinic, and its reliability in a continuously changing clinical environment.