| Enhancing Efficiency in Low-Risk Chest X-ray Reporting: A Comparative Study of Manual, Template-Based, and AI-Generated Methods
Marek Řehoř, Šimon Kličník, Jakub Dandár, Daniel Kvak |
||
| ABSTRACT: Efficient and accurate chest X-ray (CXR) reporting is essential in radiology, especially for quickly identifying low-risk cases to prioritize more complex ones. This study investigates the time efficiency of three CXR reporting methods: manual, template-based, and AI-generated, focusing specifically on low-risk CXR evaluations in a radiology department. Results show that manual reporting, which requires free-text documentation, takes significantly longer than other methods, with average mean times per study of 96.4 seconds (RAD1), 91 seconds (RAD2), and 70.8 seconds (RAD3). In contrast, the structured, template-based approach reduced these times to 32.9 seconds (RAD1), 32 seconds (RAD2), and 48.8 seconds (RAD3), representing an average efficiency improvement of 53.93% compared to manual reporting. More The AI-generated method yielded the shortest mean times per study at 27.7 seconds (RAD1), 31.9 seconds (RAD2), and 33.8 seconds (RAD3), with an average reduction of 62.82% compared to manual reporting. In conclusion, AI-generated reporting offers substantial time savings and maintains high accuracy, indicating strong potential to enhance radiology workflow efficiency. This study supports the integration of AI into routine CXR reporting, enabling radiologists to focus more on complex cases. Future research should explore the long-term impacts and further improvement of AI algorithms to optimize radiology practices. Keywords: Artificial Intelligence, Chest X-ray Reporting, Diagnostic Reporting Methods, Radiology, Time Efficiency, Workflow EfficiencyDOI: 10.35191/medsoft_2024_1_36_kvak |
















