ACR's Connect and AI-LAB technical framework

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Health Informatics

Background:

• Artificial intelligence (AI) has significantly advanced medical imaging analysis but remains underutilized in clinical settings.
• Successful AI development requires collaboration between clinicians, data scientists, and access to quality datasets.
• Current AI adoption is hindered by a lack of accessible platforms for development and validation.

Purpose of the Study:

• To develop a free, vendor-neutral software suite, ACR Connect, to democratize artificial intelligence (AI) for all individuals and institutions.
• To create a collaborative AI ecosystem across clinical and technical domains.
• To enable non-technical users to participate in AI education and model development.

Main Methods:

• ACR Connect offers educational resources, tools for dataset annotation, model building, and evaluation.
• It provides an interface for collaboration and federated learning without data transfer.
• The AI-LAB application within ACR Connect allows local data investigation while maintaining security.

Main Results:

• ACR Connect enables users to investigate AI models using local data securely via the AI-LAB application.
• The platform facilitates non-technical users' participation in AI model evaluation and training within a collaborative network.
• It supports external validation and collaborative, distributed training of AI models.

Conclusions:

• ACR Connect provides a foundational platform for democratizing AI in healthcare.
• The software empowers non-technical users to engage in AI development and education.
• It fosters a collaborative AI ecosystem essential for advancing clinical applications.