Smart Chest X-ray Worklist Prioritization using Artificial Intelligence:\n A Clinical Workflow Simulation

Abstract

The aim is to evaluate whether smart worklist prioritization by artificial\nintelligence (AI) can optimize the radiology workflow and reduce report\nturnaround times (RTAT) for critical findings in chest radiographs (CXRs).\nFurthermore, we investigate a method to counteract the effect of false negative\npredictions by AI -- resulting in an extremely and dangerously long RTAT, as\nCXRs are sorted to the end of the worklist.\n We developed a simulation framework that models the current workflow at a\nuniversity hospital by incorporating hospital specific CXR generation rates,\nreporting rates and pathology distribution. Using this, we simulated the\nstandard worklist processing "first-in, first-out" (FIFO) and compared it with\na worklist prioritization based on urgency. Examination prioritization was\nperformed by the AI, classifying eight different pathological findings ranked\nin descending order of urgency: pneumothorax, pleural effusion, infiltrate,\ncongestion, atelectasis, cardiomegaly, mass and foreign object. Furthermore, we\nintroduced an upper limit for the maximum waiting time, after which the highest\nurgency is assigned to the examination.\n The average RTAT for all critical findings was significantly reduced in all\nPrioritization-simulations compared to the FIFO-simulation (e.g. pneumothorax:\n35.6 min vs. 80.1 min; p $<0.0001$), while the maximum RTAT for most findings\nincreased at the same time (e.g. pneumothorax: 1293 min vs 890 min; p\n$<0.0001$). Our "upper limit" substantially reduced the maximum RTAT all\nclasses (e.g. pneumothorax: 979 min vs. 1293 min / 1178 min; p $<0.0001$).\n Our simulations demonstrate that smart worklist prioritization by AI can\nreduce the average RTAT for critical findings in CXRs while maintaining a small\nmaximum RTAT as FIFO.\n

Ähnliche Arbeiten