| Topic | Key findings | Source (first author, journal) | Publication date | Identifier (DOI or NCT) | URL |
|---|---|---|---|---|---|
| Classification | ICROP was initially adopted in 1984, expanded in 1987, amended in 2005, and revised in 2021; ICROP3 retains zone, stage, and extent, and adds posterior zone II plus recognition of a continuous spectrum from normal to pre-plus and plus disease (pqac-00000013) | Saidasheva, *Russian Pediatric Ophthalmology* | 2022-05-03 | DOI: 10.17816/rpoj100683 | https://doi.org/10.17816/rpoj100683 |
| Epidemiology | In the US, ROP incidence rose from 4.4% in 2003 to 8.1% in 2019; reported prevalence: Poland 15.1% (2016–2019) and 15.6% (2012–2021), Italy 38% (2017–2020), Netherlands 28.3% (2017), Portugal 23.8% (2012–2020) (pqac-00000010) | Modrzejewska, *Journal of Clinical Medicine* | 2024-07-10 | DOI: 10.3390/jcm13144034 | https://doi.org/10.3390/jcm13144034 |
| Epidemiology | Over 2010–2022, 141,550 infants received ROP screening in Germany; mean annual incidence was 3.5% in premature infants and 19.6% in screened infants; 2.0% of screened infants received treatment; infants with birth weight ≥1500 g were 35.2% of the screening population but only 0.4% of ROP stage 3–5 cases (pqac-00000016) | Krohne, *Archives of Disease in Childhood* | 2025-05-30 | DOI: 10.1136/archdischild-2024-327133 | https://doi.org/10.1136/archdischild-2024-327133 |
| Risk factors | Highest-risk infants were those born at or before 30 weeks gestation or with birth weight ≤1,500 g; incidence and severity were inversely related to gestational age and birth weight (pqac-00000011) | Dablouk, *Cureus* | 2024-11-08 | DOI: 10.7759/cureus.73286 | https://doi.org/10.7759/cureus.73286 |
| Risk factors | Most significant factors highlighted were preterm birth, low birth weight, and exposure to variable levels of exogenous oxygen; additional factors noted included infections, neonatal sepsis, hyperglycemia, insulin resistance, and low IGF-1 (pqac-00000010) | Modrzejewska, *Journal of Clinical Medicine* | 2024-07-10 | DOI: 10.3390/jcm13144034 | https://doi.org/10.3390/jcm13144034 |
| Oxygen targets | Clinical trial data cited as suggesting saturation levels of 90–95% are safer than 85–89% in preterm infants (pqac-00000010) | Modrzejewska, *Journal of Clinical Medicine* | 2024-07-10 | DOI: 10.3390/jcm13144034 | https://doi.org/10.3390/jcm13144034 |
| Oxygen targets | NeOProM recommended 91–95%; in a South Africa survey, 70.2% of infants were receiving oxygen, 81.2% of those received blended oxygen, 61.2% had saturations >95%, and only 27.1% were within the recommended 91–95% range (pqac-00000015) | Thomas, *South African Journal of Child Health* | 2024-10 | DOI: 10.7196/SAJCH.2024.v18i3.1994 | https://doi.org/10.7196/SAJCH.2024.v18i3.1994 |
| Oxygen delivery/monitoring | In Mexico, in 2011 only 38% of NICUs had adequate staffing; 25.5% of cots had blenders and 80.1% had saturation monitors; among 153 observed infants, 53% had SpO2 ≥96% and 8% ≤89%; in 2023, monitored infants were 75% vs 79% in 2011, and hyperoxia rates fell slightly from 54% to 49% (pqac-00000008) | Zepeda Romero, *BMC Nursing* | 2024-08 | DOI: 10.1186/s12912-024-02227-x | https://doi.org/10.1186/s12912-024-02227-x |
| Screening-AI performance | External validation used 2,530 examinations from 843 infants for training/calibration and tested on 6,245 examinations from 1,545 infants (SUNDROP) and 5,635 examinations from 2,699 infants (AECS); mtmROP/type 1 ROP prevalence was 5.9%/1.2% in SUNDROP and 6.2%/2.5% in AECS (pqac-00000012) | Coyner, *JAMA Ophthalmology* | 2024-03-07 | DOI: 10.1001/jamaophthalmol.2024.0045 | https://doi.org/10.1001/jamaophthalmol.2024.0045 |
| Screening-AI performance | Examination-level AUROCs were 0.896 and 0.985 in SUNDROP and 0.920 and 0.982 in AECS for mtmROP and type 1 ROP, respectively; patient-level sensitivity for future type 1 ROP was 100% in both datasets (pqac-00000012) | Coyner, *JAMA Ophthalmology* | 2024-03-07 | DOI: 10.1001/jamaophthalmol.2024.0045 | https://doi.org/10.1001/jamaophthalmol.2024.0045 |
| Treatment comparisons | Meta-analysis of 21 studies (6 RCTs, 15 non-randomized studies) including 6,152 eyes found anti-VEGF had higher recurrence than laser photocoagulation (RR 2.14, 95% CI 1.06–4.33); bevacizumab had longer retreatment interval than laser (SMD 0.89, 95% CI 0.61–1.17); no significant difference in retreatment success between bevacizumab and laser (pqac-00000011) | Dablouk, *Cureus* | 2024-11-08 | DOI: 10.7759/cureus.73286 | https://doi.org/10.7759/cureus.73286 |
| Treatment comparisons | Conbercept showed lower recurrence risk than laser photocoagulation and ranibizumab (RR 0.47, 95% CI 0.39–0.58); aflibercept showed higher recurrence risk than bevacizumab (RR 12.61, 95% CI 6.43–24.73); recurrences occurred significantly later after anti-VEGF than after laser (pqac-00000011) | Dablouk, *Cureus* | 2024-11-08 | DOI: 10.7759/cureus.73286 | https://doi.org/10.7759/cureus.73286 |
| Treatment trends | In Germany, treatment preference shifted from laser coagulation accounting for 46.2% in 2015 to anti-VEGF therapy accounting for 83.7% in 2022; a 2020 revision reducing gestational-age screening cutoff from <32 to <31 weeks decreased annual screening numbers by 25.8% (pqac-00000016) | Krohne, *Archives of Disease in Childhood* | 2025-05-30 | DOI: 10.1136/archdischild-2024-327133 | https://doi.org/10.1136/archdischild-2024-327133 |
| Clinical trial | BUTTERFLEYE was a randomized, controlled, multicenter phase 3 trial of intravitreal aflibercept vs laser photocoagulation; enrollment was 127; started 2019-10-30; primary completion 2022-08-18; primary endpoint was percentage of participants with absence of active ROP and unfavorable structural outcomes through week 52 of chronological age (pqac-00000017, pqac-00000018) | Regeneron, *ClinicalTrials.gov* | Results posted 2023-07-20 | NCT04101721 | https://clinicaltrials.gov/study/NCT04101721 |
| Clinical trial | Key eligibility included gestational age at birth ≤32 weeks or birth weight ≤1500 g; treatment-naive ROP including Zone I stage 1 plus/2 plus/3 non-plus/3 plus, Zone II stage 2 plus/3 plus, or AP-ROP; secondary objectives included need for a second treatment modality, recurrence, safety, and tolerability (pqac-00000017) | Regeneron, *ClinicalTrials.gov* | Results posted 2023-07-20 | NCT04101721 | https://clinicaltrials.gov/study/NCT04101721 |
| Model organism / mechanism support | OIR models were developed in different species; the oxygen-induced retinopathy mouse model is the most widely used and has been critical for understanding pathological retinal angiogenesis and identifying anti-angiogenic therapies; molecular regulators highlighted include VEGF, lipids, and amino acids (pqac-00000014) | Maurya, *Biomedicines* | 2024-08-23 | DOI: 10.3390/biomedicines12091937 | https://doi.org/10.3390/biomedicines12091937 |


*Table: This table compiles the main quantitative findings, identifiers, and study links for retinopathy of prematurity from the gathered evidence. It is useful as a quick-reference artifact for epidemiology, risk, oxygen management, AI screening, treatment comparison, and clinical trial documentation.*