Novel platform for development, training, and validation of computer-assisted detection/diagnosis software.

Area of Science:

• Medical Informatics
• Radiology Software Development
• Clinical Decision Support Systems

Background:

• Current systems for developing and validating computer-assisted detection/diagnosis (CAD) software face challenges in development and clinical application.
• The existing Clinical Infrastructure for Radiologic Computation of United Solutions (CIRCUS) system requires an upgrade to meet modern clinical and research needs.

Purpose of the Study:

• To develop a novel, open-source, web-based platform to overcome limitations in the development and clinical deployment of CAD software.
• To replace the existing CIRCUS system with an improved platform for CAD software development and validation.

Main Methods:

• The new system features two primary user-facing applications: CIRCUS DB (web-based image database) and CIRCUS CS (Docker plug-in-based CAD execution platform).
• These applications are supported by a shared API server, a 3D image viewer, and an image repository.
• The system leverages Docker containerization for flexible CAD software integration.

Main Results:

• The new CIRCUS platform was successfully installed on Linux servers at two clinical facilities.
• CIRCUS DB has registered a total of 1954 cases.
• CIRCUS CS is actively being used with four Docker-based CAD plug-ins.

Conclusions:

• A new, enhanced version of the CIRCUS system has been successfully developed and implemented.
• The platform's successful deployment in two clinical settings demonstrates its viability.
• The project is planned for release as an open-source software initiative to promote wider adoption and collaboration.