Application of artificial intelligence reading label system in diabetic retinopathy grading training of junior ophthalmologists and medical students / 中国医师杂志

ABSTRACT

Objective:To evaluate the efficiency of using artificial intelligence reading label system in diabetic retinopathy (DR) grading training among junior ophthalmologists and medical students.Methods:520 diabetic fundus images were randomly divided into 8 groups with 65 images in each group. 13 junior ophthalmologists and medical students were selected as the research objects. Each of them read 8 groups of pictures and evaluated the DR grading of each fundus image. The sensitivity, specificity and diagnostic test consistency (Q-kappa value) of grading results were analyzed with the DR grading given by 3 senior ophthalmologists as the gold standard. The average Q-kappa values of 13 subjects were compared between the first four times and the last four times.Results:Through 8 round reading, the average Q-kappa was elevated from 0.67 to 0.81. Average Q-kappa of round 1 to 4 was 0.77, and average Q-kappa of round 5 to 8 was 0.81. The participants were divided into two groups. Participants in group 1 were junior ophthalmologists and participants in group 2 were medical students. Average Q-kappa of group 1 was elevated from 0.71 to 0.76. Average Q-kappa of group 2 was elevated from 0.63 to 0.84.Conclusions:The artificial intelligence reading label system was a useful tool in training junior ophthalmologists and medical students in doing diabetic retinopathy grading.