Using Big Data and Machine Learning Approaches to Discover Prognostic Biomarkers and Drugs for Neuropathic Pain in Chronic SCI
Abstract
Chronic neuropathic pain ranks among the top three secondary complications that significantly impact the lives of individuals following spinal cord injuries (SCIs). It is estimated that 50 percent to 80 percent of SCI patients will experience neuropathic pain within six months of their injury, and as of now, there exists no effective treatment. Research has consistently shown that early and preemptive interventions are the most successful means of alleviating pain symptoms. Consequently, the development of predictive models capable of forecasting neuropathic pain months in advance could provide clinicians with a valuable tool to address this critical medical complication. In pursuit of this objective, our study aims to construct a predictive model using three distinct types of acute care data. By utilizing data from the TRACK-SCI database, we have identified all SCI patients for whom we possess chronic neuropathic pain status information. We have been collecting and analyzing the following data points: 1) acute care clinical variables, 2) gene expression levels from white blood cells obtained within 24 hours of the injury, and 3) routinely collected blood analytes for the entire duration of their hospital stay. We are currently employing machine learning algorithms to utilize the aforementioned acute care data in order to develop mathematical models that can estimate the probability of an SCI patient developing neuropathic pain between six and twelve months after their injury.
Document Details
- Document Type
- Technical Report
- Publication Date
- Aug 01, 2023
- Accession Number
- AD1219313
Entities
People
- Nikolaos Kyritsis
Organizations
- University of California, San Francisco