AI-Assisted 3D Ultrasound for Rapid Diagnosis of Ocular Trauma Injuries
Abstract
Eye injuries in Service Members are on the rise, mainly due to the increasing rate of utilizing improvised explosive devices (IEDs). Explosions create fast-moving shrapnel that may penetrate the eye, injuring very delicate structures such as the retina, a very important part of the eye in charge of creating visual messages sent to the brain. These injuries need to be immediately diagnosed to prevent long-term damages or permanent vision loss. The CT scans currently used as standard tools in military hospitals may not be able to show some small injuries such as separation of retina or may fail to detect small foreign bodies, such as pieces of wood or bone, that have penetrated the eye. In both situations, immediate surgery is needed to avoid long-term consequences. Fortunately, ultrasonography, a technology based on high-pitch sounds and known for its high level of safety, can detect the necessary details that need to be seen in these eye injuries and has been widely used in civilian applications. However, conventional ultrasound systems only provide two-dimensional (2D) views of the eye only from one cross-section. In many cases, to be able to detect injuries, ultrasonography needs to be performed in many directions. This technique, which increases the manipulation on the injured eyes, is discouraged to avoid potential further damages. Additionally, 2D ultrasonography requires experience and can only be operated by a highly skilled user. We propose a new technology to obtain 3D ultrasonography images of the entire eye using one fast scan, hence enabling this highly favored imaging method for a very critical application in the military. We use artificial intelligence to make highly complex processing possible on a small portable ultrasound suitable for battlefield applications. In Aim 1 of this project we will develop a 3D ultrasound system by carefully modifying an advanced ultra-high frequency but general-purpose ultrasound scanner. In Aim 2, we will develop artificial intelligence to assist with interpretation of the images for rapid diagnosis of critical cases that need immediate evacuations, e.g., retinal detachment. In Aim 3, we will test and validate our methods by extensive experiments. We will use animal eyes to be able to replicate injuries that might be seen during military traumas. These studies will include eyes that are removed from dead pigs for research purposes. We will also create realistic models of human military eye injuries in 15 rabbits under anesthesia. In all these studies we will apply our new 3D ultrasonography as well as other methods, including clinical CT scans, to assess the improvements achieved by this project. The short-term goal of this project is to develop new technologies that will fill a gap in the management of eye injuries in military personnel due to battlefield traumas and validate them on non-human examples. Our long-term goal is to make these technologies ready for human use. We envision a compact 3D ultrasonography system that after clinical validation can be used in military hospitals on injured Service Members to avoid misdiagnosis of conditions that may lead to permanent vision loss.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Dec 05, 2021
- Source ID
- W81XWH2110659
Entities
People
- Mahdi Bayat
Organizations
- Case Western Reserve University
- United States Army