Artificial Intelligence-assisted Screening of Malignant Pigmented Tumors on the Ocular Surface

Completed

• Conditions: Eye Neoplasms; Orbital Neoplasms; Conjunctival Neoplasms
• Interventions: Diagnostic Test: screening system for ocular surface malignant tumors

• Registration Number: NCT05645341
• Lead Sponsor: Sun Yat-sen University

Brief Summary

Rare diseases generally refer to diseases whose prevalence rate is lower than 1 / 10 000 and the number of patients is less than 140000. Rare diseases are generally faced with the dilemma of a lack of qualified doctors, difficulty in large-scale screening, and a lack of rapid and effective channels for medical treatment. Studies have shown that 42% of patients say they have been misdiagnosed, and each patient with a rare disease needs to go through an average of eight doctors in seven years to see a corresponding rare disease specialist. More importantly, most rare diseases seriously affect the health and quality of life of patients. The ocular surface malignant tumor is a typical rare disease, and its incidence is less than 1 / 100000. The ocular surface not only affects the patient's appearance, but also damages the visual function, and the malignant tumor may even affect life. These uncommon malignant tumors are often hidden in the common black nevus on the eye surface, which is easy to be ignored and has great potential risks. With the improvement of people's living standards, people start to pay attention to rare diseases.

In recent years, the rapid development of digital technology has also provided new opportunities for the prevention and treatment of rare diseases. Our team established the database of rare ophthalmopathy in China in the early stage, which provided a solid foundation for the digitization of precious clinical data. This study intends to develop an intelligent screening system for ocular surface malignant tumors, using the mobile phone for real-world verification and scale screening, and explore it to improve the ability of doctors to diagnose and treat rare diseases. This study is expected to improve the ability to screen malignant tumors on the ocular surface and provide a novel model for the universal screening of rare diseases.

Detailed Description

Not available

Study Timeline

• Study First Submitted: 2022-12-02
• Study First Submitted QC: 2022-12-02
• Study Start: 2022-12-05
• Study First Posted: 2022-12-09
• Primary Completion: 2023-06-05
• Study Completion: 2023-06-05
• Status Verified: 2024-08
• Last Update Submitted: 2024-08-01
• Last Update Posted: 2024-08-05

Recruitment & Eligibility

• Status: COMPLETED
• Sex: All
• Target Recruitment: 535

Inclusion Criteria

• Dark-brown lesions on the ocular surface are found: i.e. ocular surface malignant melanoma, ocular basal cell carcinoma, conjunctival nevus, eyelid nevus, sclera pigmentation, benign eyelid keratosis

Exclusion Criteria

• Non-pigmented ocular surface tumors: pterygium, corneal dermoid tumor, meibomian gland cyst, cataract, blepharitis, etc.
• The image quality does not meet the clinical requirements.

Study & Design

• Study Type: OBSERVATIONAL
• Study Design: Not specified

Arm && Interventions

Group	Intervention	Description
Eligible participants for smartphone-based ocular surface tumors diagnosis	screening system for ocular surface malignant tumors	-

Primary Outcome Measures

Name	Time	Method
Area under the curve (AUC)	2024.1	Measure of the ability of a binary classifier to distinguish between malignent and benign.

Secondary Outcome Measures

Name	Time	Method
Screening coverage	2024.1	Count the number of people who have successfully received and read knowledge about ocular surface pigmented tumors on each offline and online platform.
Sensitivity, specificity and accuracy	2024.1	The study will assess the sensitivity and specificity of the CaptureTumor (CaT) system under various conditions.
Referral efficiency	2024.1	For cases where the system judges that it is necessary to go to the hospital for further diagnosis and treatment, two or more researchers will conduct a diagnostic review first. If further diagnosis and treatment is really needed, the subject will be contacted and told to go to the hospital for treatment by phone, text message, etc., and continue to follow up. Finally, the duration of diagnosis (screening time to pathological diagnosis time), visit distance, number of visits before diagnosis, and the proportion of referred patients in all subjects were counted.
Human-machine collaboration performance	2024.1	Doctors with different seniority were asked to diagnose the test set with and without assistance from the intelligent screening system, and the accuracy in the two cases were calculated and compared.

Trial Locations

Locations (1)

Zhognshan Ophthalmic Center, Sun Yat-sen University; 🇨🇳 Guangzhou, Guangdong, China