Automated detection of type 1 ROP, type 2 ROP and A-ROP based on deep learning.

PURPOSE: To provide automatic detection of Type 1 retinopathy of prematurity (ROP), Type 2 ROP, and A-ROP by deep learning-based analysis of fundus images obtained by clinical examination using convolutional neural networks. MATERIAL AND METHODS: A total of 634 fundus images of 317 premature infants born at 23-34 weeks of gestation were evaluated. After image pre-processing, we obtained a rectangular region (ROI). RegNetY002 was used for algorithm training, and stratified 10-fold cross-validation was applied during training to evaluate and standardize our model. The model's performance was reported as accuracy and specificity and described by the receiver operating characteristic (ROC) curve and area under the curve (AUC). RESULTS: The model achieved 0.98 accuracy and 0.98 specificity in detecting Type 2 ROP versus Type 1 ROP and A-ROP. On the other hand, as a result of the analysis of ROI regions, the model achieved 0.90 accuracy and 0.95 specificity in detecting Stage 2 ROP versus Stage 3 ROP and 0.91 accuracy and 0.92 specificity in detecting A-ROP versus Type 1 ROP. The AUC scores were 0.98 for Type 2 ROP versus Type 1 ROP and A-ROP, 0.85 for Stage 2 ROP versus Stage 3 ROP, and 0.91 for A-ROP versus Type 1 ROP. CONCLUSION: Our study demonstrated that ROP classification by DL-based analysis of fundus images can be distinguished with high accuracy and specificity. Integrating DL-based artificial intelligence algorithms into clinical practice may reduce the workload of ophthalmologists in the future and provide support in decision-making in the management of ROP.

One-year refractive outcomes after intravitreal bevacizumab versus laser photocoagulation for retinopathy of prematurity.

PURPOSE: To investigate the effects of laser photocoagulation (LPC) and intravitreal bevacizumab (IVB) therapy used in the treatment of retinopathy of prematurity (ROP) on the first age refraction values in our center. METHODS: The preterm infants who received LPC (Group I) and IVB therapy (Group II) for ROP were evaluated, and the refraction results were retrospectively compared. RESULTS: The study included 86 eyes of 45 infants with a mean birth week of 26.5 ± 2.1 weeks and a mean birth weight of 904 ± 223 g. Treatments were administered up to a mean PMA of 36.0 ± 2.4 and 35.3 ± 2.6 weeks in Group I and Group II, respectively. In the follow-up examinations, 1-year spherical, cylindrical, and spherical equivalent (SE) values were 0.1 ± 2.2 D, - 1.2 ± 0.9 D, and - 0.5 ± 2.0 D in Group I and 1.3 ± 1.7 D, - 1.1 ± 0.8 D, and 0.8 ± 1.7 D in Group II, respectively (P = 0.018 for spherical; P = 0.772 for cylindrical, and P = 0.009 for SE). The mean spherical power and SE were significantly higher in Group II for zone II disease (p = 0.005 and p = 0.002). In addition, according to the ROP stage, infants with Stage 3 ROP were found to be significantly more myopic than infants in Stage 2 ROP in Group I (p = 0.03). CONCLUSION: In conclusion, this study supports that even 0.625 mg IVB for ROP causes less myopia compared to LPC. Consistent with the literature, it was observed that the stage and zone of ROP had a significant effect on the development of myopia.

Development of myopia in laser-treated ROP infants: prematurity or laser photocoagulation?

PURPOSE: To investigate the relationship between the development of corrected 1-year-old refraction values and risk factors in preterm infants who underwent laser photocoagulation (LPC) therapy for retinopathy of prematurity (ROP). METHODS: The ophthalmic examination findings of preterm neonates who had developed Type I ROP and received LPC therapy were evaluated retrospectively. The association between spherical equivalent (SE) values and clinical findings were analyzed by using multivariable linear regression analysis. RESULTS: The study included 157 eyes of 80 neonates with a mean birth week of 27.1 ± 2.2 weeks (23 to 32 weeks) and a mean birth weight of 995 ± 273 g (565 to 1760 g). The treatments were administered on an average of 36.8 ± 2.7 (32 to 45 weeks) postmenstrual age. LPC treatment was applied bilaterally to 77 of the 80 neonates included in the study, and unilaterally to 3 of them. The mean ± standard deviation of the SE value was 0.31 ± 1.89 diopters (D) (- 8.00 to 4.63 D) according to the results of the 1-year corrected age refraction examination. In univariate analysis, no significant association between GA, BW, and ROP zone and SE value, while the number of laser spots (ß = - 0.27 ± 0.00 D, p = 0.00) and stage 3 ROP (ß = - 0.29 ± 0.37 D, p = 0.00) were significantly associated with the SE value. In multivariable linear regression analysis, a significant association between number of laser spot, stage 3 ROP and SE value (ß = - 0.25 ± 0.00 D, p = 0.01 for number of laser spot, ß = - 0.28 ± 0.36 D, p = 0.00 for stage 3 ROP). CONCLUSION: In conclusion, this study supports that stage of ROP and the number of laser spots count applied in photocoagulation treatment for ROP is significantly correlated with degree of myopia (p < 0.05).