Artificial Intelligence in the Identification of Lung Contusions Through Chest Radiological Examination in Blunt Thoracic Trauma

What this Trial Is About

This research aims to evaluate how well a computer-based system called LUNIT can interpret chest X-rays to identify lung contusions in patients with blunt chest injuries. The study compares LUNIT's readings to those of radiologists and uses chest high-resolution CT scans as the gold standard. Lung injuries from chest trauma can lead to serious complications and require careful monitoring to prevent respiratory failure. The study is observational and retrospective, involving patients who visited the Emergency Department of IRCCS AOU of Bologna between June 2014 and June 2024. Included patients had blunt thoracic trauma and underwent both a standard chest X-ray and a chest HRCT within 48 hours. Radiologists will analyze the chest X-rays with and without the support of the LUNIT software, allowing comparison of the system's ability to detect lung contusions against standard interpretations. Participants' existing medical images and reports will be reviewed without altering their care. The study measures how accurately LUNIT detects lung contusions compared to radiologist readings, using chest HRCT as the reference. This evaluation will help determine if LUNIT can reduce the need for CT scans, lowering radiation exposure and potentially speeding up emergency care management. The study is expected to last about one year until mid-2025.

AI in the Identification of Lung Contusions Through Chest Radiological Examination in Blunt Thoracic Trauma