A new study published in the Journal of American Medical Association unveiled the artificial intelligence (AI) model which used only photos from smartphones, showed good performance in correctly diagnosing ptosis, strabismus, and myopia.
Early detection of pediatric eye disorders is a global concern where conventional screening methods are costly and time-consuming, requiring hospitals and ophthalmologists. Artificial intelligence might make it easier to diagnose eye diseases in a home environment by evaluating children's eye health using smartphone images. Qin Shu and colleagues carried out this research to create an AI model that can recognize ptosis, myopia, and strabismus from cellphone photos.
Children having a diagnosis of myopia, strabismus, or ptosis were included in this cross-sectional study, which took place at the Department of Ophthalmology at Shanghai Ninth People's Hospital between October 1, 2022 and September 30, 2023. A deep learning-based model for the diagnosis of ptosis, strabismus, and myopia was created. Sensitivity, specificity, accuracy, negative predictive values (NPV), positive predictive values (PPV), negative likelihood ratios (N-LR), positive likelihood ratios (P-LR),the area under the curve (AUC), and the F1-score were used to evaluate the model's performance. The effect of each region on the model was visually and analytically evaluated using GradCAM++. To confirm the model's generalizability, subgroup analyses by age and sex were carried out.
The model was constructed using 1419 pictures altogether, taken from 476 patients (225 female [47.27%] and 299 [62.82%] aged between 6 and 12 years). Of these, 473 binocular photos were utilized to diagnose strabismus, while 946 monocular images were used to diagnose myopia and ptosis. In terms of myopia, strabismus, and ptosis, the model showed high sensitivity. During sex subgroup analysis, the model performed as well in diagnosing eye abnormalities in children that were male and female. The ability to recognize eye diseases varied depending on the age group.
This cross-sectional study discovered that an AI-based detection algorithm performed well in properly recognizing myopia, strabismus, and ptosis using just smartphone photos. These findings imply that it can help families test their children for myopia, strabismus, and ptosis, allowing for early detection and lowering the risk of visual impairment and severe difficulties caused by delayed screening.
Reference:
Shu, Q., Pang, J., Liu, Z., Liang, X., Chen, M., Tao, Z., Liu, Q., Guo, Y., Yang, X., Ding, J., Chen, R., Wang, S., Li, W., Zhai, G., Xu, J., & Li, L. (2024). Artificial Intelligence for Early Detection of Pediatric Eye Diseases Using Mobile Photos. In JAMA Network Open (Vol. 7, Issue 8, p. e2425124). American Medical Association (AMA). https://doi.org/10.1001/jamanetworkopen.2024.25124