1485. Mobile Bedside Ultrasound (mBSUS) and Use of an Artificial Intelligence Algorithm for Diagnosis of Pediatric Pneumonia in Limited Resource Settings

BACKGROUND: In low and middle-income countries (LMICs) pneumonia is by far the leading cause of death among children < 5 years of age. A key factor is the challenge of pneumonia diagnosis. Chest X-Ray is the gold standard for pneumonia diagnoses but exposes children to ionizing radiation and is mainly restricted to hospital settings. advances in artificial intelligence (AI) render possible the automated interpretation of mobile bedside US (mBSUS) images on a smartphone, obviating the need for a radiologist. Ultraspund findings in pneumonia [Image: see text] Artificial intelligence feature recognition [Image: see text] METHODS: We measured the accuracy of mBSUS for the diagnosis of pneumonia using chest X-Ray as the gold standard. Children 1-59 mo presenting at the University Teaching Hospital in Lusaka, Zambia with ages ranging from aged 1-59 months and meeting WHO criteria for severe/very severe pneumonia were enrolled. Clinical data is collected in RedCap. Digital X-Rays were done at the University Teaching Hospital and saved as JPEG images. Pulmonary mBSUS images are taken using a butterfly, a mobile device system, and stored in the butterfly iCloud of the Butterfly app and transmitted to an iOS phone or tablet. Images are stored locally and saved to a secured/encrypted cloud platform for remote viewing with a HIPAA (Health Insurance Portability and Accountability Act) compliant secure cloud. Images are currently extracted from the clips stored in the butterfly icloud, radiologists annotate the images that have abnormal findings and they are then sent to the AI lab where they are analyzed and organized to build a platform of similar images that could be recognized by the machine learning system. Imaging correlation CXR Vs mobile bedside ultrasound mBSUS [Image: see text] Butterfly ultrasound system [Image: see text] RESULTS: Of the 11 patients enrolled so far, ll have been having ultrasound images that correlated with chest x-ray findings. In three of those patients, the ultrasound has shown pulmonary findings not recognized or hardly seen on chest x-ray. The artificial intelligence lab is developing a pull of images that will be used to recognize patterns of consolidation from mBSUS images. Protocol fro obtaining images [Image: see text] CONCLUSION: Mobile pulmonary ultrasound mBSUS is a feasible, non radiation technique that could be used in limited-resource settings to diagnose pneumonia in children. Images obtained from mBSUS can be used to build a pattern of recognition based on consolidation findings. DISCLOSURES: All Authors: No reported disclosures