A recent large-scale study published on July 12 found that artificial intelligence (AI) shows promise in reducing the workload of radiologists in outpatient settings, specifically in the analysis of chest X-rays. The study, conducted by scientists from Siemens Healthineers, involved testing an algorithm on radiographs from over 14, 000 patients at Zwanger-Pesiri Radiology LLP, an outpatient imaging center in the northeast U. S. The algorithm, called "AI NAD Analyzer, " was able to correctly classify cases with no actionable findings with a sensitivity of 29% and a yield of nearly 21%. Its specificity, or ability to distinguish images with disease from those without, was approximately 99%, with a miss rate of only 0. 3% cases. The AI system did not overlook any critical findings and missed significant findings in just 0. 06% of cases.

The authors of the study noted that in an outpatient population, AI could identify 20% of chest radiographs as containing no actionable disease, which could potentially allow for a streamlined reading protocol, improving efficiency and reducing daily workload for radiologists. The researchers also emphasized the importance of workflow integration and caution against potential biases caused by incorrect AI results. They suggested that future iterations of the system could combine the AI NAD Analyzer with other AI systems to create a more comprehensive decision support system. To read the full study, click on the link provided.