FLORetina 2025: Weight and PMA play central roles in preterm foveal development

Melissa Yuan, MD, talked with Ophthalmology Times Europe about her presentation, "Prospective Foveal Light Reflex Development in Prematurity: Roles of Weight and Post Menstrual Age." She presented this talk during the 2025 FLORetina meeting and International Congress on OCT Angiography, En Face OCT, and Advances in OCT (ICOOR), held from 4 to 7 December in Florence, Italy.

Modern Retina: How do weight and postmenstrual age (PMA) interact to influence foveal maturation in preterm infants?

Melissa Yuan, MD: PMA was the strongest predictor, while weight showed a relationship with foveal and an interaction with PMA as well. Each additional week of PMA increased odds of foveal light reflex (FLR) presence by 44.6%, while each 200g increase in weight independently increased odds by 23.1%.

This suggests that foveal maturation isn't solely determined by chronological age but that nutritional status and overall growth also contribute.

MR: What imaging techniques were most effective in capturing these developmental changes? Were there techniques you thought might be effective that weren't?

Yuan: We used wide-field digital retinal imaging with the 120° FOV 3nethra NEO camera. This was very effective for longitudinal tracking. We found this technique a great fit for the more resource-limited setting we were using it in, given that it is non-invasive, portable, and works well with telemedicine protocols. OCT theoretically offers superior microstructural detail but has limitations in premature infants. Additionally, OCT use is limited by cost and availability, particularly in low- and middle-income countries where ROP rates are increasing.

We found that the FLR combined with pigmentary changes was quite reliable for foveal identification across serial imaging. The progression from no macular definition to pigmentary changes to partial and then complete FLR was consistently identifiable on fundus photography.

MR: Did your team identify thresholds or critical windows for optimal visual outcomes?

Yuan: We did not look at visual outcomes in this study. We did, however, find some thresholds for anatomic development. In eyes screened before 40 weeks PMA, the 2100g threshold was associated with significantly higher odds of FLR development. This weight threshold appears to represent a critical mass for foveal maturation.

Regarding PMA, we found that 75% of eyes showed FLR by 36 weeks PMA and 50% by 34 weeks. The mean PMA at first FLR was 35 0/7 weeks for low-risk ROP and 35 2/7 weeks for high-risk ROP, suggesting this 34-36 week window represents a critical developmental period.

MR: How might these findings inform neonatal care or nutritional strategies?

Yuan: The finding that absolute weight independently predicts FLR development beyond PMA represents a potentially modifiable variable for FLR and potentially functional outcomes. This suggests that improvements in nutrition could help identify FLR and subsequently more accurately identify Zone I.

Currently, many NICUs focus primarily on preventing growth failure but may not consider the ophthalmologic implications of suboptimal weight gain. Achieving the 2100g threshold by 36-37 weeks PMA could represent a nutritional target that serves dual purposes, aiding in general health and retinal maturation.

MR: What are the global challenges to applying findings from this research?

Yuan: The challenges are substantial, particularly in low- and middle-income countries where ROP rates are increasing as neonatal intensive care units develop capacity to save younger and smaller infants. Resource allocation is a major hurdle. Also, our patient cohort came from West Africa, so generalizability to other racial and ethnic groups and different neonatal care practices needs validation.

Imaging access is another barrier. While we demonstrated that widefield digital retinal imaging-based telemedicine is feasible and cost-effective, some regions may lack this technology, and OCT remains inaccessible in many parts of the world.

MR: Could these parameters serve as predictive markers for long-term visual function?

Yuan: This is an important future research direction. Our model achieved 84.8% predictive accuracy for FLR presence using only PMA and weight, which demonstrates robust predictive capability for this early developmental milestone.

The logical next step is determining whether the timing of FLR development correlates with visual acuity. Previous longitudinal studies following preterm infants into adolescence and adulthood have associated low birthweight, low gestational age, and postnatal ROP with disruptions in macular development and functional consequences. We're particularly interested in whether infants who achieve FLR earlier or at higher weights have better long-term outcomes.

MR: Are there any innovations in ROP management you'd like to see discussed more broadly among retina specialists?

Yuan: Our team believes that Zone I identification would benefit from standardization. Current International Classification of Retinopathy of Prematurity (ICROP) definitions rely on identifying the macular center, which is problematic when FLR is absent early in screening.

Our next study directly addresses this by having trained graders estimate FLR position and Zone I on images before and after FLR development to quantify how much grader variability is attributable to FLR visibility.

Artificial intelligence applications represent another innovation. While significant work has focused on improving grading of plus disease and stage through deep learning, zone assessment has been neglected. AI models could potentially predict foveal location before FLR development, reducing zone misclassification.

We also would benefit as a field from better integration of telemedicine into standard ROP care. Our GHANAROP program demonstrates feasibility by screening 30+ neonatal units across Ghana's Ashanti region through telemedicine partnership.

The nutritional intersection with ophthalmology is also understudied. Prospective interventional trials evaluating whether optimized nutrition protocols reduce ROP severity or improve visual outcomes would help us better understand a potentially modifiable risk factor.

MR: What insights do you hope to gain from the intersection of adult and pediatric retinal research at this event?

Yuan: I am excited about learning how artificial intelligence and imaging innovations developed for adult retinal diseases can be adapted for ROP. The algorithms for vascular analysis, ultra-widefield imaging protocols, and automated grading systems being used for diabetic retinopathy and AMD may be translated to neonatal screening.

I'm particularly interested in emerging imaging biomarkers. Retinal thickness parameters or vascular density metrics from OCT-A that predict functional outcomes in adults might be translatable to the pediatric population. The pharmacologic management discussions will be valuable since the same anti-VEGF agents used in adult disease are now being applied to ROP, but we're still learning about optimal dosing, timing, and long-term effects in developing eyes.

I also want to learn from telemedicine implementations in adult retina practices. Many have developed sophisticated remote screening programs for diabetic retinopathy with lessons about workflow optimization, quality control, and grader training that could strengthen our global ROP efforts. Understanding the full continuum of retinal vascular development and disease from neonatal development through adulthood is a particular interest of mine.