Automated Integration of AI Results into Radiology Reports Using Common Data Elements

Area of Science:

• Medical Imaging
• Artificial Intelligence in Medicine
• Radiology Informatics

Background:

• Artificial intelligence (AI) offers potential to enhance diagnostic accuracy and workflow efficiency in radiology.
• Integrating AI-derived measurements into radiology reports requires standardized data formats for seamless incorporation.
• Common Data Elements (CDEs) provide a standardized framework for interoperable health information.

Purpose of the Study:

• To describe the application of CDEs for embedding AI-derived measurements into radiology reports.
• To define a set of CDEs for liver and spleen volume and attenuation measurements.
• To demonstrate the use of CDEs in clinical practice for AI data transfer.

Main Methods:

• An AI system segmented liver and spleen on non-contrast CT images.
• AI-generated measurements were recorded as CDEs using the DICOM-SR framework.
• Automated systems extracted CDEs, incorporated them into reports, and sent images to PACS.

Main Results:

• The AI system successfully segmented abdominal organs and generated volume and attenuation measurements.
• CDE labels and values were successfully extracted from AI data and integrated into radiology reports.
• The process enabled transmission of AI-generated image series to PACS.

Conclusions:

• Radiology CDEs provide a practical framework for recording and transferring AI-generated data.
• This approach enhances communication, facilitates research, and improves decision support systems.
• CDEs ensure consistency, interoperability, and clarity in reporting AI findings across healthcare systems.