The Quantitative Study of the Habenula Based on Multi-channel Cascaded Neural Network and the Establishment of the Prediction Model of the Curative Effect in Patients With Depression

NCT05872607 | Unknown

• 1 other identifier: zlei19810830
• Study Type: observational
• Target: 300
• Locations: 1 country
• Sites: 1

Brief Summary

Depression is the second leading cause of disease burden in our country. It has serious effects on the physical and mental health of human beings, and about 30% of patients with depression are unresponsive or respond poorly to antidepressant treatment. Clinical practice is in a tough position of wanting objective measures of assessing depression. The applicant and her team have devoted many years to the basic and clinical research on habenular nucleus (Hb) accumulating a significant amount of experience from animal experiments and patients' magnetic resonance (MR) studies. These studies have demonstrated that the habenular nucleus is the key target area that is responsible for the pathophysiological changes in depression as well as its treatment. Small volumes and unsatisfactory contrast have been knotty problems in the MR imaging of Hb. In addition, time-consuming manual segmentation and lack of quantitative standards in conventional studies has impeded the advancement of Hb research. Fortunately, the development of high-resolution multi-parametric quantitative MR imaging and the extensive use of artificial intelligence (AI) technology in medical imaging can just provide powerful support for the imaging, segmentation and quantification of Hb. This project proposes to use high resolution MR anatomy of Hb combined with multimodal fusion to 1) construct a model for automatic 3D segmentation of Hb MR images based on the densely connected multichannel dilated convolutional neural networks; 2) sift out the quantitative imaging signatures related to the antidepressants' efficacy using the radiomics methodology, and in combination with clinical information, construct an individualized prediction model for treatment efficacy. Overall, this study focuses on the translation of basic research to clinical application in the hope of providing quantifiable objective imaging markers in clinical practice, facilitating clinical decision-making and bringing about individualized precise diagnosis and treatment.

Conditions

Depression

Keywords

Depression; Habenula; Convolution neural network; Segmentation; Curative effect prediction

Outcome Measures

Primary Outcomes (1)

• Multimodal MR data of habenula

  To extract high-dimensional features of multimodal image data including MP2RAGE, T1MAP, T2 SPACE, DIR, QSM, T2 \*

  Baseline (Before medication)

Eligibility Criteria

• Sex: all
• Healthy Volunteers: Yes
• Age Groups: Child (0-17), Adult (18-64), Older Adult (65+)
• Sampling Method: Probability Sample

Study Population

Premiere depressed patients

You may qualify if:

• Premiere depressed patients

You may not qualify if:

• Drug therapy modality change、Incomplete MR image data

Sponsors & Collaborators

• The First Hospital of Jilin University: lead

Study Sites (1)

The First Hospital of Jilin University

Changchun, Jilin, 130000, China

RECRUITING

MeSH Terms

Conditions

Depression

Condition Hierarchy (Ancestors)

Behavioral Symptoms; Behavior

Central Study Contacts

Lei Zhang, doctor

CONTACT

15804300150; zlei99@jlu.edu.cn

Study Design

• Study Type: observational
• Observational Model: OTHER
• Time Perspective: PROSPECTIVE
• Sponsor Type: OTHER
• Responsible Party: SPONSOR

Study Record Dates

• First Submitted: May 15, 2023
• First Posted: May 24, 2023
• Study Start: January 1, 2022
• Primary Completion: December 31, 2024
• Study Completion: December 31, 2024
• Last Updated: May 24, 2023

Record last verified: 2023-02

Locations

CN (1)