Application of WINROP to predict severe retinopathy of prematurity in a multi-ethnic cohort in Singapore

Aim: To investigate the validity of WINROP use in multi-ethnic population in a tertiary centre in Singapore. Methods: Birth weight, gestational age, and weekly weight measurements of four hundred two preterm infants (<32 weeks gestation) born between year 2011 and 2019 were entered into WINROP algorithm. Based on their weekly weight gain, WINROP algorithm would signal an alarm if the infant is at risk for type 1 ROP requiring treatment. The WINROP result is then compared with the ophthalmological findings. All the infants were screened based on the hospital ROP screening protocol. The negative predictive value, positive predictive value, sensitivity and specificity were calculated. Results: Among all the infants enrolled, 31 infants developed type 1 ROP. WINROP successfully signalled 23 out of 31 correctly. The calculated sensitivity was 74.2 % and specificity was 48.0 %. The positive predictive value was 10.6% and negative predictive value was 95.7%. Conclusions: Our study showed that when WINROP is applied to our multi-ethnic population, it has a moderate sensitivity of 74.2% and a high predictive negative value of 95.7%. We would not recommend it to be used a solitary screening method but it can be used to better risk stratify preterm infants at risk of type 1 ROP, particularly in resource limited settings.