AI-Assisted Collaborative Reporting In Radiology: A Novel Framework For DICOM Structured Reporting Integration

Abstract

The integration of artificial intelligence into radiology practice has been significantly hindered by interoperability challenges between AI systems and existing clinical workflows. This article presents a novel framework for AI-assisted collaborative reporting that enables direct generation of DICOM Structured Reports from deep learning models, eliminating traditional format conversion barriers. The system architecture integrates a DICOM Gateway, AI Inference Engine, Structured Reporting Module, and Collaborative Review Interface to transform medical image analysis results into standardized clinical documentation. Multi-institutional evaluation across academic medical centers and community hospitals demonstrated substantial workflow efficiency improvements and enhanced documentation quality. The framework achieved complete DICOM compliance and universal PACS compatibility while preserving radiologist autonomy through collaborative review interfaces. Technical performance validation confirmed clinically acceptable accuracy levels for anatomical segmentation and pathology detection. Clinical workflow assessment revealed significant time reduction in report preparation with improved finding detection sensitivity and measurement consistency. User acceptance evaluation indicated strong satisfaction with system usability and clinical utility. Quality improvements included enhanced template compliance, standardized terminology usage, and comprehensive documentation practices. The implementation addresses critical barriers to AI adoption in healthcare by providing seamless integration with existing infrastructure without requiring workflow disruption or extensive training. The framework establishes a foundation for next-generation AI-assisted healthcare technologies that prioritize interoperability and clinical workflow compatibility.