A Scar Recognition Software for Chronic Spinal Cord Injury (SCI)

In Vivo Optimization and Clinical Application of a Scar Recognition Software for Chronic Spinal Cord Injury (SCI)

To construct and validate a software to recognize scar for patients with chronic SCI based on multimodal MRI.

Study Overview

• Status: Not yet recruiting
• Conditions: Spinal Cord Injury
• Intervention / Treatment: Other: MRI

Detailed Description

Spinal cord injury (SCI) is a kind of serious neurologic damage caused by violence to the spinal cord, resulting in various functions of the body below the injury level, including motor, sensory, sphincter, and reflex dysfunction in varying degrees, usually resulting in permanent and irreversible functional loss or paralysis of patients. The treatment of SCI is an essential problem in the world. In the past decades, experimental research on SCI involves genes, proteins, cells, and tissues, and has made great progress. However, these studies mainly focus on the SCI at the early stage, rather than the later stage. The reason is that in the later stage, scar formed by glial cells and fibroblasts in the injured area is a physical and chemical barrier, which inhibits the regeneration and myelination of nerve axons and results in inhibiting spinal cord repairment. Therefore, before the treatment of chronic SCI, the key problem is to distinguish glial scar tissue from normal tissue and eliminate its influence.

As glial scar inhibits axon regeneration, eliminating glial scar is necessary for the repair of the injured spinal cord. In recent years, a large number of experimental studies have been carried out to destroy the process of glial scar formation after SCI by enzyme digestion and antibody. Though these methods reduced glial scar, residual glial scars were reported in animal experiments. Compared to biochemical methods, surgical resection of glial scar tissue is a relatively simple and effective method to eliminate glial scars. Due to the limited regeneration ability of nerves after SCI, it is important to identify scar tissue accurately before operations to avoid surgical injury to normal tissue, which is also the premise of further research and clinical application of various interventional treatment methods.

Magnetic resonance imaging (MRI) is one of the most commonly used non-invasive imaging techniques to evaluate the degree of injury and therapeutic effect of SCI. Nemours MRI studies on SCI show the impact of SCI on the central nervous system from the structural and functional level and prove the potential application value of MRI in assisting doctors in the diagnosis of SCI. A small number of previous studies have used magnetization transfer imaging, and diffusion tensor imaging to detect glial scar tissue, showing the potential application value of these images in differentiation between glial scar and surrounding normal spinal cord. However, because glial cells, which constitute glial scar, are also important components of normal spinal cord tissue, previous studies only identified glial scar from a single aspect, such as tissue type, macromolecular component, or water molecular diffusion strength. Therefore, their specificities were unsatisfactory. Relative methods were unable to identify glial scar accurately and finally resulted in difficulty in treatment arrangement and evaluation of prognosis, which hinders the development of SCI treatment research.

Combing multimodal MRI, including conventional MRI and diffusion MRI, with supervised machine learning makes accurate glial identification in chronic SCI possible. multimodal MRI can depict the differences between scar tissue and non-scar tissue from the aspects of cell composition, water molecular dispersion, structural complexity, etc. Comparing to MRI with a single model, multimodal MRI provides more specific features. Machine learning, a way to construct robust and accurate models, can mine the quantitative relationship between imaging features and clinical diagnosis results, reveal MRI feature markers of the glial scar, to improve the accuracy of identification. The research work, combined with medicine, imaging, and artificial intelligence technology, is expected to solve the problem of accurate and non-invasive identification of glial scar in chronic SCI, which has potential application value for laboratory research and clinical treatment of chronic SCI.

• Study Type: Observational
• Enrollment (Anticipated): 25

Contacts and Locations

• Study Contact: Name: Mengze Zhang; Phone Number: 18600393607; Email: zmzforever@pku.edu.cn
• Study Contact Backup: Name: Ouyang Hanqiang; Email: ouyanghanqiang@bjmu.edu.cn

Participation Criteria

Eligibility Criteria

• Ages Eligible for Study: Child; Adult; Older Adult
• Accepts Healthy Volunteers: N/A
• Genders Eligible for Study: All

• Sampling Method: Non-Probability Sample

Study Population

The patients came from the Peking University Third Hospital.

Description

Inclusion Criteria:

• (Prospective part) compliance to MRI scan
• (Prospective part) no MRI contraindication
• (Retrospective part) available conventional MRI data
• clinical diagnosis of SCI (the course of disease≥1 year)

Exclusion Criteria:

• prior head or neck surgery or accompanying diseases with neurologic deficits and/or symptoms including multiple sclerosis, motor neuron disease, or spinal cord tumor
• images with motion artifact

Study Plan

How is the study designed?

Design Details

• Observational Models: Cohort
• Time Perspectives: Other

Number of groups / cohorts

2

Cohorts and Interventions

Group / Cohort	Intervention / Treatment
Training; random splitting based on random sequences generated by engineers to train and optimize a machine learning model	Other: MRI; conventional MRI and diffusion MRI
Testing; random splitting based on random sequences generated by engineers to evaluate the performance of the model	Other: MRI; conventional MRI and diffusion MRI

What is the study measuring?

Primary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Performance of the fitted model	positive predictive value (PPV)	through study completion, an average of 2 year
Performance of the fitted model	sensitivity (SE)	through study completion, an average of 2 year
Performance of the fitted model	Dice coefficient score (DSC)	through study completion, an average of 2 year

Collaborators and Investigators

• Sponsor: Peking University Third Hospital
• Investigators: Principal Investigator: Huishu Yuan, Peking University Third Hospital

Study record dates

Study Major Dates

• Study Start (Anticipated): September 1, 2021
• Primary Completion (Anticipated): December 1, 2022
• Study Completion (Anticipated): June 1, 2023

Study Registration Dates

• First Submitted: June 29, 2021
• First Submitted That Met QC Criteria: July 1, 2021
• First Posted (Actual): July 8, 2021

Study Record Updates

• Last Update Posted (Actual): July 8, 2021
• Last Update Submitted That Met QC Criteria: July 1, 2021
• Last Verified: September 1, 2020

More Information

Terms related to this study

• Additional Relevant MeSH Terms: Central Nervous System Diseases; Nervous System Diseases; Trauma, Nervous System; Spinal Cord Diseases; Wounds and Injuries; Spinal Cord Injuries
• Other Study ID Numbers: M2020400，M2020356

Plan for Individual participant data (IPD)

• Plan to Share Individual Participant Data (IPD)?: No

Drug and device information, study documents

• Studies a U.S. FDA-regulated drug product: No
• Studies a U.S. FDA-regulated device product: No