A Novel Fundus Image Reading Tool for Efficient Generation of a Multi-dimensional Categorical Image Database for Machine Learning Algorithm Training

• Published in: Journal of Korean medical science Vol. 33; no. 43; pp. e239 - 12
• Main Authors: Park, Sang Jun, Shin, Joo Young, Kim, Sangkeun, Son, Jaemin, Jung, Kyu-Hwan, Park, Kyu Hyung
• Format: Journal Article
• Language: English
• Published: Korea (South) The Korean Academy of Medical Sciences 22.10.2018; 대한의학회
• Subjects: Databases, Factual; Fundus Oculi; Humans; Machine Learning; Original; Republic of Korea; Retina - diagnostic imaging; 의학일반; Republic of Korea; Retina Fundus Image; Reading Tool; Machine Learning; Deep Learning; Grader
• ISSN: 1011-8934; 1598-6357; 1598-6357
• DOI: 10.3346/jkms.2018.33.e239

We described a novel multi-step retinal fundus image reading system for providing high-quality large data for machine learning algorithms, and assessed the grader variability in the large-scale dataset generated with this system. A 5-step retinal fundus image reading tool was developed that rates image quality, presence of abnormality, findings with location information, diagnoses, and clinical significance. Each image was evaluated by 3 different graders. Agreements among graders for each decision were evaluated. The 234,242 readings of 79,458 images were collected from 55 licensed ophthalmologists during 6 months. The 34,364 images were graded as abnormal by at-least one rater. Of these, all three raters agreed in 46.6% in abnormality, while 69.9% of the images were rated as abnormal by two or more raters. Agreement rate of at-least two raters on a certain finding was 26.7%-65.2%, and complete agreement rate of all-three raters was 5.7%-43.3%. As for diagnoses, agreement of at-least two raters was 35.6%-65.6%, and complete agreement rate was 11.0%-40.0%. Agreement of findings and diagnoses were higher when restricted to images with prior complete agreement on abnormality. Retinal/glaucoma specialists showed higher agreements on findings and diagnoses of their corresponding subspecialties. This novel reading tool for retinal fundus images generated a large-scale dataset with high level of information, which can be utilized in future development of machine learning-based algorithms for automated identification of abnormal conditions and clinical decision supporting system. These results emphasize the importance of addressing grader variability in algorithm developments.