AI-Assisted 3D Ultrasound for Rapid Diagnosis of Ocular Trauma Injuries
Abstract
Traumatic injuries to the eye, such as those in the military, require immediate diagnosis to prevent vision loss. These impacts can alter anatomical integrity of the eye causing, for example, retinal detachment. Removal of penetrating intraocular foreign bodies (IOFBs) requires clear visualization for surgical removal. Traumatic injuries coincide with edema and bleeding; thus, limiting the utility of optic based methods such as direct ophthalmoscopy or OCT. Small foreign bodies, may not be well visible in CT imaging, even at highest resolutions. Ultrasound has been used for diagnosis and treatment of ocular injuries. Conventional 2D ultrasound requires skilled users and are subject to large variability. We propose to develop a novel 3D ultrasound system to greatly facilitate diagnosis and treatment of ocular injuries. Our gentle-touch approach via single sweep scanning minimizes the risk of additional injuries. We have created a prototype of the system resulting in anatomically correct volumes of ex vivo eyes. Initial deep learning results show promise in automated analysis of 3D data.
Document Details
- Document Type
- Technical Report
- Publication Date
- Aug 01, 2023
- Accession Number
- AD1216834
Entities
People
- Mahdi Bayat
Organizations
- Case Western Reserve University