A Plugin-Based Architecture for Integrating AI Services in an Open-Source PACS.

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Health Informatics

Background:

• Automated medical image analysis using AI offers workflow advantages like reduced screening time and observer variability.
• Current AI applications in medical imaging operate in a non-standardized environment, hindering algorithm reuse and data sharing.
• Developing AI algorithms requires significant time and difficult-to-acquire labeled datasets, limiting their widespread application.

Purpose of the Study:

• To present a framework for standardizing medical image analysis.
• To facilitate the integration and development of new AI algorithms within a production-ready imaging archive.
• To address the limitations posed by non-standardized environments in AI-driven medical imaging.

Main Methods:

• Development of a novel open-source interface.
• Integration of the interface into the Dicoogle Picture Archiving and Communication System (PACS).
• Adherence to standard industry protocols for seamless integration and interoperability.

Main Results:

• The proposed framework enables easier integration of AI algorithms into medical imaging workflows.
• Standardization facilitates the sharing and reutilization of AI algorithms and their outputs.
• The open-source interface promotes a more collaborative and efficient development ecosystem.

Conclusions:

• The framework successfully addresses the standardization challenges in medical image analysis.
• The open-source interface enhances the applicability and reach of AI algorithms in healthcare.
• This approach supports improved diagnostic accuracy and efficiency in medical imaging practices.