Medical Image Algorithms for Radiotherapy

The challenge

An international medical technology company intended to expand its software for radiotherapy with a radiation planning feature. The goal was to provide the user with the ability to automatically define the radiation area based on algorithms. UL Solutions Software Intensive Systems (SIS) advised the company on the selection of an appropriate development approach and supported it throughout the introduction of necessary methods and techniques, and the implementation of the software. The concrete implementation took place within a larger project setup to which a total of four SIS engineers contributed.

Incorrectly focused or excessive X-ray radiation can cause serious harm to the patient, such as burn injuries, organ failure or cancer. During the planning phase, it is important that software users can precisely pinpoint the areas for radiation. In addition to the automated treatment planning, more algorithms were required for visualization. The treatment planning is based on a segmentation of the tumor from the organs, which have to be protected from the radiation. The user can also manually correct or reorient the radiation area if no organ segmentation is available or if the available segmentation is suboptimal. In most cases, tumor patients receive multiple radiation treatments. To prevent excessive radiation for the respective tissue beyond its limits, previously applied radiations needed to be included in the visualization.

During the treatment, the shape of the radiated area is controlled by metal plates that partially absorb radiation. One of the biggest technical challenges in the project was the numerous individual physical collimator configurations and their limitations, forming a huge variety of beam shapes. Due to very close cooperation and daily alignment with many stakeholders on customer’s premises, the project was successfully brought to market.

Subproject duration and resourcing

• One year
• Two software engineers (C++, C++/CLI, C#)

Our approach

Based on previous positive outcomes with agile methodology recommended by SIS engineers, the customer decided to intensify their activities in agile development methods in this project. The project team developed all functional and nonfunctional requirements for the needed algorithms and their implementation with Kanban. This enabled the team to include short-term change requests into design and development. According to their functionalities, algorithms are divided into independent components. These individual components can be easily reused in future.

To avoid potential pitfalls in C++ development, the project team applied modern techniques, such as the Resource Acquisition Is Initialization (RAII) pattern, which were complemented with Test-Driven Development (TDD) and detailed quality assurance measures, such as static code analysis, profiling and extensive testing.

Highlights of our approach

• Agile development measures
• TDD
• Modern, tested and reusable C++ code

Benefit

SIS experts' understanding of modern software engineering allowed us to combine agile approaches with awareness of medical regulations to provide feature components within the customers’ deadline. Due to the modular structure of the delivery, future changes can be accomplished quickly and easily with limited additional effort.