Skip to content

A Multicenter Study on Meibomian Gland Imaging Radiomics Subtyping and Intervention Strategies Based on Agentic Artificial Intelligence

A Multicenter Study on Meibomian Gland Imaging Radiomics Subtyping and Intervention Strategies Based on Agentic Artificial Intelligence

Status
Active, not recruiting
Phases
Unknown
Study type
Observational
Source
ChiCTR
Registry ID
ChiCTR2600120107
Enrollment
Unknown
Registered
2026-03-09
Start date
2026-03-09
Completion date
Unknown
Last updated
2026-03-16

For informational purposes only — not medical advice. Sourced from public registries and may not reflect the latest updates. Terms

Conditions

Meibomian Gland Dysfunction (MGD) and associated Dry Eye Disease.

Interventions

Dry Eye Syndrome Observation Group:None

Sponsors

Fuzhou University Affiliated Provincial Hospital
Lead Sponsor

Eligibility

Sex/Gender
All
Age
4 Years to 80 Years

Inclusion criteria

Inclusion criteria: 1. Subjects aged between 4 and 80 years old. 2. Meets the clinical diagnostic criteria for meibomian gland dysfunction (MGD). 3. The subjects (or their legal guardians) are able to cooperate in completing the ophthalmic image collection and subsequent follow-up. 4. Voluntarily participate in this study and sign the informed consent form.

Exclusion criteria

Exclusion criteria: 1. Those who have previously received targeted treatment for dry eye or meibomian gland-related conditions. 2. Those with other severe ocular surface or eye diseases, such as blepharitis, entropion, glaucoma, etc. 3. Those with severe systemic diseases, such as severe liver or kidney dysfunction or malignant tumors. 4. Those with a history of using drugs that may affect the stability of the tear film (such as isotretinoin, etc.)

Design outcomes

Primary

MeasureTime frame
Accuracy of AI classification model;Sensitivity of AI classification model;Cohen’s Kappa between AI and expert classification;Area Under the Curve (AUC) of AI classification model;Specificity of AI classification model;

Secondary

MeasureTime frame
Concordance of AI recommendations with clinicians;Cross-center robustness;F1-score of AI model;SHAP feature contribution;Brier score;Intraclass Correlation Coefficient (ICC) of radiomic features;Accuracy of Grad-CAM lesion localization;

Countries

China

Contacts

Public ContactLi Li

Fuzhou University Affiliated Provincial Hospital

lili_js@fjsl.com.cn+86 591 88618523

Outcome results

None listed

Source: ChiCTR (via WHO ICTRP) · Data processed: Mar 20, 2026