Development and multicenter validation of an AI driven model for quantitative meibomian gland evaluation

Li Li; Kunhong Xiao; Taichen Lai; Kunfeng Lai; Jiawen Lin; Zongyuan Ge; Lingyi Liang; Hao Huang; Xiaoshan Zhang; Li Liu; Yujie Wang; Xianwen Shang; Mingguang He; Ying Xue; Zhuoting Zhu

doi:10.1038/s41746-025-01753-5

Development and multicenter validation of an AI driven model for quantitative meibomian gland evaluation

NPJ Digit Med. 2025 Jul 4;8(1):403. doi: 10.1038/s41746-025-01753-5.

Authors

Li Li^{1

2

3}, Kunhong Xiao⁴, Taichen Lai³, Kunfeng Lai⁵, Jiawen Lin^{5

6}, Zongyuan Ge^{7

8}, Lingyi Liang⁹, Hao Huang¹⁰, Xiaoshan Zhang¹¹, Li Liu¹², Yujie Wang^{1

2}, Xianwen Shang^{13

14}, Mingguang He^{15

16}, Ying Xue¹⁷, Zhuoting Zhu^{18

19

20}

Affiliations

¹ Centre for Eye Research Australia, Royal Victorian Eye and Ear Hospital, East Melbourne, VIC, Australia.
² Department of Surgery (Ophthalmology), The University of Melbourne, Melbourne, VIC, Australia.
³ Department of Ophthalmology, Shengli Clinical Medical College of Fujian Medical University, Fujian Provincial Hospital, Fuzhou University Affiliated Provincial Hospital, Fuzhou, China.
⁴ Department of Ophthalmology and Optometry, Fujian Medical University, Fuzhou, China.
⁵ College of Computer and Data Science, Fuzhou University, Fuzhou, China.
⁶ Engineering Research Center of big data intelligence, Ministry of Education, Fuzhou University, Fuzhou, China.
⁷ Monash e-Research Centre, Monash University, Melbourne, VIC, Australia.
⁸ Monash Medical AI Group, Monash University, Melbourne, VIC, Australia.
⁹ State Key Laboratory of Ophthalmology, Zhongshan Ophthalmic Center, Sun Yat-sen University, Guangdong Provincial Key Laboratory of Ophthalmology and Visual Science, Guangzhou, China.
¹⁰ Department of Ophthalmology, Zhuzhou Hospital Affiliated to Xiangya School of Medicine, Central South University, Zhuzhou, China.
¹¹ Dongguan Huaxia Eye Hospital, Dongguan, China.
¹² Huaxia Eye Hospital Group Putian Eye Hospital, Putian, China.
¹³ School of Optometry, The Hong Kong Polytechnic University, Hong Kong, China.
¹⁴ Research Centre for SHARP Vision, The Hong Kong Polytechnic University, Hong Kong, China.
¹⁵ School of Optometry, The Hong Kong Polytechnic University, Hong Kong, China. mingguang.he@unimelb.edu.au.
¹⁶ Research Centre for SHARP Vision, The Hong Kong Polytechnic University, Hong Kong, China. mingguang.he@unimelb.edu.au.
¹⁷ Department of Ophthalmology, Shengli Clinical Medical College of Fujian Medical University, Fujian Provincial Hospital, Fuzhou University Affiliated Provincial Hospital, Fuzhou, China. 502873740@qq.com.
¹⁸ Centre for Eye Research Australia, Royal Victorian Eye and Ear Hospital, East Melbourne, VIC, Australia. lisa.zhu@unimelb.edu.au.
¹⁹ Department of Surgery (Ophthalmology), The University of Melbourne, Melbourne, VIC, Australia. lisa.zhu@unimelb.edu.au.
²⁰ Department of Ophthalmology and Optometry, Fujian Medical University, Fuzhou, China. lisa.zhu@unimelb.edu.au.

Abstract

This multicenter retrospective study developed and validated an AI driven model for automated segmentation and quantitative evaluation of meibomian glands using infrared meibography images acquired by the Keratograph 5M device. A total of 1350 infrared meibography images were collected and annotated for model training and validation. The model demonstrated high segmentation performance, with an Intersection over Union of 81.67% (95% Confidence Interval [CI]: 81.03-82.31) and accuracy of 97.49% (95% CI: 97.38-97.62), outperforming conventional algorithms. The agreement was observed between AI-based and manual gland grading (Kappa = 0.93) and gland counting (Spearman r = 0.9334). Repeatability analysis confirmed the model's stability, and external validation across four independent centers yielded consistent results with AUCs exceeding 0.99. This AI tool offers a standardized, efficient, and objective method for meibography image analysis, which may improve diagnostic precision and assist in the clinical management of meibomian gland dysfunction across diverse populations.

Abstract

Grants and funding