VCOM Research Day Program Book 2023

Medical Student Research Biomedical

01 Image Segmentation: Identifying and Quantifying Peripheral Nerve Fascicles for Clinical Applications in Nerve Injury Repair

Malek Zahed; Kelly Roballo Corresponding author: mzahed@vcom.edu

Edward Via College Of Osteopathic Medicine-Virginia Campus

Peripheral nerve injuries are debilitating injuries that can leave patients immobile and at risk for comorbid conditions. Peripheral nerve repair and all its subtypes give patients the chance to regain function. Yet, those that regain full function are very few and it is not uncommon to remain severely debilitated, even after attempted repair. The prognostic indicators for recovery are multifaceted and complex and one of the factors includes the ability to assimilate a nerve graft into a transected in-vivo nerve. This allows axons within a fascicle to create new connections. A major component of this process is to select a graft that most resembles the internal organization of fascicles within a nerve. This has proven inefficient

and impractical thus far. Image segmentation is a form of image processing that isolates aspects of an image to identify certain objects and to further analyze those objects. This form of processing is used across many fields from security and military weapons to studying microscopic structures of materials and microbiologic organisms. Yet, this technology also can further advance clinical medicine. This research delves into the use of image segmentation to isolate and identify fascicles and their organizational layout within a nerve to improve nerve grafting with respect to repairing peripheral nerve transections and injuries. This research uses the Anaconda-Python platform to build and write a program that has the ability to

segment and identify fascicles within a cross-section of a nerve image or ultrasound. Our hypothesis is that image segmentation can be efficiently and practically used to accurately identify and quantify fascicles within peripheral nerves in a clinical setting. The preliminary results show that our program segments fascicles from a microscopic image of a nerve cross section acquired from a cadaver with 87% accuracy.

56