This prospective longitudinal cohort study (BabySTEPS, n=72 very preterm infants, mean GA 27.6 weeks) evaluated whether bedside OCT measures of retinal microanatomy at 36 weeks' postmenstrual age (PMA) — including RNFL thickness, choroidal thickness, inner retinal thickness, and macular edema — predicted neurodevelopmental outcomes (BSID-III, CBCL, M-CHAT-R) at 2 years' corrected age.
Greater RNFL thickness was independently associated with higher BSID-III Motor scores (+7.50 points per 10-µm increase; P<.001) and Cognitive scores (+3.71 points; P=.02), lower autism risk (M-CHAT-R: −0.64 per 10 µm; P=.03), and fewer internalizing behavior problems (CBCL: −2.25 per 10 µm; P=.04). Adding RNFL thickness to standard infant factors boosted Motor score prediction from R²=0.36 to 0.53 (gain +0.17). Greater choroidal thickness also predicted better Motor scores (+4.84 points per 100 µm; P=.03) but added no incremental value beyond RNFL alone.
- Single-center study using a **research-only OCT device** not yet commercially available, limiting immediate clinical translation. - 26 of 98 surviving infants (26.5%) were lost to follow-up; those lost were slightly less preterm and heavier, potentially underestimating impairment burden. - Multiple comparisons were not adjusted, raising the risk of false-positive associations for secondary behavioral outcomes.
Bedside OCT RNFL thickness measured during routine ROP examinations at ~36 weeks PMA may serve as an early, low-stress biomarker to flag very preterm infants most at risk for motor and cognitive delays — enabling earlier enrollment in developmental therapies. Pending independent validation and a commercially available device, clinicians should consider RNFL thickness data as a supplement to, not replacement for, standard neurodevelopmental risk assessment.
Explore related topics
