RESUMEN
PURPOSE: To estimate the effect of atropine eyedrops at different concentrations for myopia control in children. METHODS: We conducted a Bayesian random-effects network meta-analysis based on randomized controlled trials (RCT). Primary outcomes include changes in spherical equivalent error (SER) and changes in axial length (AL), mean difference (MD) together with 95% credible interval (CrI) were used to evaluate the efficacy. RESULTS: 28 RCTs (6608 children) were included in this review. Comparing ten atropine eyedrops (0.0025%, 0.005%, 0.01%, 0.02%, 0.025%, 0.05%, 0.1%, 0.25%, 0.5% and 1% concentrations) with the placebo, the MDs and 95%CrIs of changes in SER are -0.006 (-0.269, 0.256) D, 0.216 (-0.078, 0.508) D, 0.146 (0.094, 0.199) D, 0.167 (0.039, 0.297) D, 0.201 (0.064, 0.341) D, 0.344 (0.251, 0.440) D, 0.255 (0.114, 0.396) D, 0.296 (0.140, 0.452) D, 0.331 (0.215, 0.447) D, and 0.286 (0.195, 0.337) D, respectively. The MDs and 95%CrIs of changes in AL are -0.048 (-0.182, 0.085) mm, -0.078 (-0.222, 0.066) mm, -0.095 (-0.130, -0.060) mm, -0.096 (-0.183, -0.009) mm, -0.083 (-0.164, -0.004) mm, -0.114 (-0.176, -0.056) mm, -0.134 (-0.198, -0.032) mm, -0.174 (-0.315, -0.061) mm, -0.184 (-0.291, -0.073) mm, and -0.171 (-0.203, -0.097) mm, respectively.Whether evaluated by SER or AL, 1% concentration ranks first in efficacy, but the risk of photophobia is 17 times higher than 0.01% concentration. CONCLUSIONS: 0.01% or higher concentration atropine eyedrops are effective for myopia control, while 0.0025% and 0.005% concentrations may not. As the concentration increases, the effect tends to increase, 1% concentration may have the strongest effect.
RESUMEN
OBJECTIVE: To evaluate the performance of two lightweight neural network models in the diagnosis of common fundus diseases and make comparison to another two classical models. METHODS: A total of 16,000 color fundus photography were collected, including 2000 each of glaucoma, diabetic retinopathy (DR), high myopia, central retinal vein occlusion (CRVO), age-related macular degeneration (AMD), optic neuropathy, and central serous chorioretinopathy (CSC), in addition to 2000 normal fundus. Fundus photography was obtained from patients or physical examiners who visited the Ophthalmology Department of Beijing Tongren Hospital, Capital Medical University. Each fundus photography has been diagnosed and labeled by two professional ophthalmologists. Two classical classification models (ResNet152 and DenseNet121), and two lightweight classification models (MobileNetV3 and ShufflenetV2), were trained. Area under the curve (AUC), sensitivity, specificity, accuracy, positive predictive value, and negative predictive value were used to evaluate the performance of the four models. RESULTS: Compared with the classical classification model, the total size and number of parameters of the two lightweight classification models were significantly reduced, and the classification speed was sharply improved. Compared with the DenseNet121 model, the ShufflenetV2 model took 50.7% less time to make a diagnosis on a fundus photography. The classical models performed better than lightweight classification models, and Densenet121 showed highest AUC in five out of the seven common fundus diseases. However, the performance of lightweight classification models is satisfying. The AUCs using MobileNetV3 model to diagnose AMD, diabetic retinopathy, glaucoma, CRVO, high myopia, optic atrophy, and CSC were 0.805, 0.892, 0.866, 0.812, 0.887, 0.868, and 0.803, respectively. For ShufflenetV2model, the AUCs for the above seven diseases were 0.856, 0.893, 0.855, 0.884, 0.891, 0.867, and 0.844, respectively. CONCLUSION: The training of light-weight neural network models based on color fundus photography for the diagnosis of common fundus diseases is not only fast but also has a significant reduction in storage size and parameter number compared with the classical classification model, and can achieve satisfactory accuracy.
