Article Impact Level: HIGH Data Quality: STRONG Summary of Journal of Imaging Informatics in Medicine. https://doi.org/10.1007/s10278-024-01200-z Dr. Neil Chatterjee et al.
Points
- The study developed an AI orchestrator to integrate AI algorithms into clinical radiology workflows across a multi-site university healthcare system.
- The orchestrator was tested by processing 991 abdominal CT scans over 60 days for hepatic steatosis screening, with an average turnaround time of 2.8 minutes.
- Standardized data formats and quality control measures ensured seamless integration of AI results into existing clinical workflows without disrupting radiologists’ routines.
- The orchestrator demonstrated scalability, allowing AI deployment across multiple locations with minimal impact on existing processes.
- The framework offers a scalable model for integrating AI into healthcare, potentially reducing barriers to AI implementation and enhancing diagnostic efficiency in clinical settings.
Summary
The study aimed to address the challenge of integrating artificial intelligence (AI) into clinical radiology workflows by developing an AI orchestrator. This orchestrator was designed to streamline the deployment and usage of AI algorithms across an extensive, multi-site university healthcare system. The authors conducted an opportunistic screening for hepatic steatosis using abdominal CT scans to demonstrate the effectiveness of their orchestrator. Over 60 days, 991 abdominal CT scans were processed at multiple locations, with an average turnaround time of 2.8 minutes. Integrating AI results and quality control images into the clinical workflow was achieved seamlessly, utilizing standardized data formats for all inputs and outputs.
The methodology employed in this study emphasized the use of standardized data formats, allowing the AI orchestrator to process images and results efficiently while fully integrating with existing clinical infrastructure. Quality control measures ensured the AI results were accurate and suitable for clinical use, making them readily accessible to radiologists without disrupting their routine. The study demonstrated that AI deployment could be effectively scaled across multiple locations with minimal impact on existing workflows.
The framework developed by the authors provides a scalable model for incorporating AI into clinical settings. By adapting the orchestrator, other healthcare systems could integrate various AI algorithms for different medical purposes, enhancing the efficiency of diagnostic procedures and clinical decision-making. The study highlights that such an approach can significantly reduce the barriers to implementing AI in clinical environments, potentially broadening the use of AI across healthcare enterprises.
Link to the article: https://link.springer.com/article/10.1007/s10278-024-01200-z
References Chatterjee, N., Duda, J., Gee, J., Elahi, A., Martin, K., Doan, V., Liu, H., Maclean, M., Rader, D., Borthakur, A., Kahn, C., Sagreiya, H., & Witschey, W. (2024). A cloud-based system for automated ai image analysis and reporting. Journal of Imaging Informatics in Medicine. https://doi.org/10.1007/s10278-024-01200-z
