Retrospective validation of the postnatal growth and retinopathy of prematurity (G-ROP) and Colorado retinopathy of prematurity (CO-ROP) models in a Turkish cohort

Background. The aim of this study was to investigate the effectiveness of the Postnatal Growth and Retinopathy of Prematurity (G-ROP) and Colorado Retinopathy of Prematurity (CO-ROP) models in predicting the risk of Retinopathy of Prematurity (ROP) in preterm infants at a tertiary ROP diagnostic and treatment center. Methods. The G-ROP and CO-ROP models were applied to the study group using the data obtained. The sensitivity and specificity of both models were then calculated. Results. One hundred and twenty-six infants were included in the study. When the G-ROP model was applied to the study group, the model's sensitivity at detecting any stage ROP was 88.7%, while it was 93.3% for the treated group. The specificity of the model was 10.9% for any stage ROP, and 11.7% for the treated group. For the CO-ROP model in the same study group, the sensitivity at detecting any stage ROP was 87.3%, while it was 100% for the treated group. The CO-ROP model's specificity was 40% for any stage ROP, and 27.9% for the treated group. When cardiac pathology criteria were introduced to both models, the sensitivity of the G-ROP and CO-ROP models increased to 94.4% and 97.2%, respectively. Conclusions. It was found that the G-ROP and CO-ROP models are simple and effective models for predicting any degree of ROP development, but that they are unable to be 100% accurate. When the models were modified by introducing cardiac pathology criteria, it was observed that they began to produce more accurate results. Studies with larger groups are needed in order to assess the applicability of the modified criteria.