Dicomflex: A novel framework for efficient deployment of image analysis tools in radiological research

Area of Science:

• Medical image processing
• Radiological research
• Software engineering

Background:

• Medical image analysis tools are often developed ad-hoc, leading to reliability and maintenance issues.
• A need exists for standardized, maintainable software structures in medical image research.

Purpose of the Study:

• To present and evaluate Dicomflex, a novel framework for developing medical image analysis tools.
• To establish a uniform workflow for common medical image analysis tasks.

Main Methods:

• Developed using object-oriented Matlab code.
• Features a standardized workflow: image-slice selection, user interaction, processing, visualization, and progression.
• Comprises three core classes, two configuration files, and a user interface for displaying images and data.

Main Results:

• Three research tools successfully developed using the Dicomflex framework.
• Users reported a shorter learning curve, simpler handling, and a more intuitive interface compared to previous tools.
• Framework-inherent software version management enhances maintenance and data management.
• Facilitates rapid development of new tools for tasks like segmentation and ROI analysis on 2D datasets.

Conclusions:

• Dicomflex offers a reliable and maintainable solution for medical image processing tool development.
• The standardized workflow and interface benefit radiological research requiring quick deployment and dependable results.
• Potential applications include cardiac performance assessment, cerebrovascular disease detection, and cancerous lesion characterization.