Purpose: To implement a software-development governance framework including an automated testing platform to ensure patient safety for custom-user scripts for an application programming interface (API) for a commercial treatment planning system (TPS).
Methods: Stakeholders formed a group to disseminate best-practices, development standards, and work proactively with developers to determine testing requirements specific to each project. An automated testing platform was developed to execute the new software against either the clinical TPS or a non-clinical instance of the TPS reserved for testing, and to capture and compare the state of the database before and after the execution. This snapshot allows comparison between all API-accessible data for treatment plans and structure sets for a given test patient.
Results: The testing platform is flexible and allows the user to either pre-specify all required input or execute in batch-mode, or interactively run the script. The batch-run can be configured using a graphical user interface or using an API for more custom configuration. Acquiring the snapshot of the database takes between 1-3 seconds per test case, allowing rapid evaluation of large test sets. Tools are provided to visualize and document the observed changes to the database to simplify evaluation of the results. This framework has been used to validate multiple projects that are now in routine clinical use.
Conclusion: The use of an automated testing platform within a development-governance framework can streamline the process of development and integration of user-built tools for automation of treatment planning, reducing the effort required by developers to ensure the safety and effectiveness of their software.
Treatment Planning, Quality Assurance, Quality Control
TH- External Beam- Photons: Treatment planning using machine learning/Knowledge Based Planning/automation