Convolutional Neural Network for the Detection of Cervical Myelomalacia

Convolutional Neural Network for the Detection of Cervical Myelomalacia on Magnetic Resonance Imaging

Deep learning technology has been used increasingly in spine surgery as well as in many medical fields. However, it is noticed that most of the studies about this subject in the literature have been conducted except of the cervical spine. In this study, we aimed to demonstrate the effectiveness of the deep learning algorithm in the diagnosis of cervical myelomalacia compared to conventional diagnostic methods.

Artificial neural networks, a machine learning technique, have been used in several industrial and research fields increasingly. The development of computational units and the increasing amount of data led to the development of new methods on artificial neural networks

Study Overview

• Status: Completed
• Conditions: Cervical Myelopathy
• Intervention / Treatment: Diagnostic test: Convolutional Neural Network

Detailed Description

Cervical myelopathy (CM) is a frequent degenerative disease of the cervical spine that occurs as a result of compression of the spinal cord. In evaluating of this disease and determining treatment options, the patient's clinic and radiological modalities should be evaluated together.

The current imaging procedures for CM are plain roentgenograms, computed tomography and magnetic resonance imaging (MRI). However, MRI in CM is more valuable in evaluating of the disc, spinal cord and other soft tissues compared to other imaging methods. Artificial intelligence technologies also used in many health applications such as medical image analysis, biological signal analysis, etc. In this study, we aimed to demonstrate the effectiveness of the deep learning algorithm in the diagnosis of cervical myelomalacia compared to conventional diagnostic methods.

• Study Type: Observational
• Enrollment (Actual): 125

Contacts and Locations

Study Locations

• Turkey
  • Fatih
    • Istanbul, Fatih, Turkey, 34093
      • Istanbul University

Participation Criteria

Eligibility Criteria

• Ages Eligible for Study: 32 years to 77 years (ADULT, OLDER_ADULT)
• Accepts Healthy Volunteers: N/A
• Genders Eligible for Study: All

• Sampling Method: Probability Sample

Study Population

The participants are aged 30-80 years, who have cervical myelomalacia that proved in MRI.

Description

Inclusion Criteria:

• the patients with classical cervical myelomalacia sypmtoms such as neck pain and stiffness, weakness and clumsiness at the upper extremities or gait difficulties and radiological findings of spinal compression
• 30-80 years age.

Exclusion Criteria:

• Patients with a previous history of cervical spinal surgery and has a systematic disease (rheumatologic or neural disease) .

Study Plan

How is the study designed?

Number of groups / cohorts

2

Cohorts and Interventions

Group / Cohort	Intervention / Treatment
cervical myelopathy; MR images of patients with cervical myelopathy	Diagnostic test: Convolutional Neural Network; Convolutional neural networks, a machine learning technique, have been used in several industrial and research fields increasingly. The development of computational units and the increasing amount of data led to the development of new methods on artificial neural networks. Deep learning (DL) is a multi-layered neural network in which feature extraction is done automatically. It extends traditional neural networks by adding more hidden layers to the network architecture between the input and output layers to model more complex and nonlinear relationships.
normal; normal section of the MRI of patients with cervical myelopathy	Diagnostic test: Convolutional Neural Network; Convolutional neural networks, a machine learning technique, have been used in several industrial and research fields increasingly. The development of computational units and the increasing amount of data led to the development of new methods on artificial neural networks. Deep learning (DL) is a multi-layered neural network in which feature extraction is done automatically. It extends traditional neural networks by adding more hidden layers to the network architecture between the input and output layers to model more complex and nonlinear relationships.

What is the study measuring?

Primary Outcome Measures

Outcome Measure	Measure Description	Time Frame
The value of confusion matrix accuracy for sagittal views	It is a specific table layout that allows visualization of the performance of an algorithm.	1 day
The value of confusion matrix accuracy for axial views	It is a specific table layout that allows visualization of the performance of an algorithm.	1 day

Collaborators and Investigators

• Sponsor: Istanbul University
• Investigators: Principal Investigator: Hakan Yilmaz, Karabuk University, Faculty of Engineering; Principal Investigator: Murat Korkmaz, Istanbul University, Faculty of Medicine

Study record dates

Study Major Dates

• Study Start (ACTUAL): April 15, 2021
• Primary Completion (ACTUAL): April 22, 2021
• Study Completion (ACTUAL): April 22, 2021

Study Registration Dates

• First Submitted: March 11, 2021
• First Submitted That Met QC Criteria: March 11, 2021
• First Posted (ACTUAL): March 15, 2021

Study Record Updates

• Last Update Posted (ACTUAL): June 1, 2021
• Last Update Submitted That Met QC Criteria: May 27, 2021
• Last Verified: May 1, 2021

More Information

Terms related to this study

• Keywords: deep learning; cervical myelopathy; neural network
• Additional Relevant MeSH Terms: Central Nervous System Diseases; Nervous System Diseases; Hematologic Diseases; Spinal Cord Diseases; Bone Marrow Diseases
• Other Study ID Numbers: KAEK/2020.07.129

Plan for Individual participant data (IPD)

• Plan to Share Individual Participant Data (IPD)?: YES
• IPD Plan Description: It can be shared after publication
• IPD Sharing Time Frame: after publication
• IPD Sharing Supporting Information Type: STUDY_PROTOCOL; SAP; ICF

Drug and device information, study documents

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