Imaging Data Saves Lives: AI-Enabled Worklist Prioritization as a Triage Tool and Learning Engine in Large-Scale Combat Operations

To the Editor: Remondelli et al. make a compelling case that the U.S. military trauma system’s dependence on delayed, human-powered documentation is incompatible with the operational tempo of large-scale combat operations (LSCO).1 Their proposed ecosystem, centered on passive data capture, scalable edge computing, and AI-enabled triage and decision support, is the correct architecture for future conflict. One high-density, time-sensitive data stream receives no dedicated attention within that framework: the radiology imaging queue. This omission matters not only for acute casualty triage but for a second function the authors identify as foundational, namely the ability of the trauma system to learn rapidly from injury patterns and course-correct clinical practice before the next casualty arrives. At the point of acute care, the imaging queue is a triage decision node. Remondelli et al. note that when providers face task saturation, documentation and secondary processes are abandoned in favor of direct patient care.1 The same constraint governs radiology throughput: a single military radiologist managing an unranked worklist during a mass casualty event will sequence studies in arrival order, not acuity order, absent automated support. The critical hemothorax and the minor soft tissue injury compete for the same interpretive slot. AI-driven worklist prioritization eliminates this competition by re-sequencing the queue based on detected injury severity without requiring any provider action. This is the class of passive, provider-independent decision support Remondelli et al. advocate, applied to imaging data.