PERFORMANCE OF A NOVEL MULTIMODAL LARGE LANGUAGE MODEL IN INTERPRETING MEIBOMIAN GLANDS QUANTITATIVELY AND QUALITATIVELY

To evaluate the performance of a novel multimodal large language model in interpreting quantitative and qualitative alterations in Meibomian glands through meibography images.

Ege University Department of Ophthalmology Cornea and Ocular Surface subspecialty

A total of 228 meibography images were analyzed manually by the same researcher and an assessment was performed in terms of gland drop out ratio and morphological abnormalities. Meibomian gland loss was graded from 0 (no loss) to 3 (>2/3 loss of total gland area). One-hundred and sixty images, comprising 40 images per grade, were included in the study. Claude 3.5 Sonnet, a multimodel large language model, developed by Anthropic (California, United States) was utilized to investigate its performance in evaluating meibography images across different grades.

Claude 3.5 Sonnet showed high performance in grading Meibomian gland dropout, correctly scoring 97.5%, 92.5%, 95%, and 85% of images in Grades 0, 1, 2, and 3, respectively. The model showed a tendency to classify Grade 3 images as Grade 2 in 15% of cases. In addition, Claude 3.5 Sonnet showed remarkable performance in detecting morphological abnormalities, including heterogeneous lumen diameters, lumen tortuosity, shortened lumen length, and hyperreflective gland residues. The model detected all of the 48 manually identified morphological abnormalities accurately. In 12 images, initially classified as morphologically normal by the manual assessment, the model reported additional subtle abnormalities such as mild tortuosity, inconsistent spacing, and minor gland fusion.

Claude 3.5 Sonnet showed promising results in interpreting meibography images, detecting morphological abnormalities and discriminating normal Meibomian glands from abnormal. This study shows that multimodal large language models such as Claude 3.5 Sonnet have a potential in visual based assessments and might serve as a complementary educational tool in ophthalmology clinics. The model's ability to perform detailed morphological evaluations and respond to further questions provides a tailored learning experience for young ophthalmologists.