Machine Learning Analysis of Ultrasound Images for the Investigation of Thoracolumbar Myofascial Pain and Therapeutic Efficacy of Hydrodissection
Abstract
TECHNICAL ABSTRACT LAY ABSTRACT Low back pain (LBP) is a condition with high prevalence among both Service Members and Veterans and is one of the leading causes of disability worldwide. In the two-year period from 2016 to 2018, 34.7% of all active-duty Soldiers received care for low back pain. In this project, we will recruit individuals with or without LBP, collect information on their demographics and pain status and perform ultrasound imaging of their thoracolumbar fascia (TLF), a soft tissue complex implicated in LBP. As many individuals with LBP have restrictions (tightness) of the TLF and/or alterations in tissue structure or density, medical imaging is an appealing approach to investigate TLF changes. Ultrasound in particular is convenient, does not use ionizing radiation, is relatively cost effective, and is effective in detecting soft tissue alterations as a point-of-care approach. We will quantify lower back flexibility in order to examine potential relationships between pain, function, and imaging characteristics. We will perform detailed computer-based image analyses to quantify metrics describing image texture. We seek to identify imaging biomarkers (signatures) critical to the identification, diagnosis, treatment, and monitoring of low back pain. We will further analyze the collected images using artificial intelligence (machine learning) methods in order to develop a computer code which can objectively and rapidly identify image features that can accurately diagnose whether a particular image is that of an individual with LBP (or not). Lastly, we will investigate the clinical efficacy of a technique known as hydrodissection, which involves the application of fluid to the TLF to separate its tissue layers. We will track patients over time including measurements of pain, function, and ultrasound imaging. The primary objectives of the proposed project are to (1) develop reliable, quantitative image analysis approaches to objectively classify images from subjects with active or latent TLF pain from those without pain and (2) adapt these quantitative analysis tools to investigate and predict the efficacy of hydrodissection as a novel therapeutic treatment for chronic low back pain. Our research will provide new insights and tools for monitoring LBP in civilian and military populations. The methodologies developed are expected to (a) improve the accuracy of monitoring LBP and (b) demonstrate the therapeutic efficacy of a new approach to rapidly improve (lessen) LBP which in turn will substantially improve the health and well-being of Service Members.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Jan 04, 2024
- Source ID
- HT94252311076
Entities
People
- Vincent Wang
Organizations
- United States Army
- Virginia Tech