Early prediction of severe retinopathy of prematurity in infants born at < 32 weeks’ gestation: a multicenter prediction model study

Severe retinopathy of prematurity (ROP) is a vision-threatening condition among infants born at < 32 weeks’ gestation. We aimed to develop and internally validate an early prediction model for severe ROP using routinely available clinical variables. This prospective multicenter cohort study included infants born at < 32 weeks’ gestation admitted to 28 hospitals in Shenzhen, China, between January 2022 and December 2023. The prediction target was the conditional risk of severe ROP among infants surviving to ROP assessment. Severe ROP was defined as stage 3 or higher according to the International Classification of Retinopathy of Prematurity, Third Edition. Candidate predictors were maternal, perinatal, and early neonatal variables available at birth or within the first 24 h. A ridge logistic regression model was developed. Performance was assessed using discrimination, calibration, Brier score, decision curve analysis, bootstrap optimism correction, and leave-one-center-out internal-external validation. Among 1,453 initially identified infants, 112 died before ROP assessment and 5 duplicate records were excluded, leaving 1,336 infants for prediction modeling. Severe ROP occurred in 179 infants (13.4%). The apparent area under the receiver operating characteristic curve (AUC) was 0.844 (95% confidence interval [CI], 0.812–0.876). After 500 bootstrap resamples, the optimism-corrected AUC was 0.831, with an optimism-corrected Brier score of 0.089 and calibration slope of 1.192. In leave-one-center-out internal-external validation, the pooled AUC was 0.789 (95% CI, 0.749–0.827), with a Brier score of 0.094 and calibration slope of 1.031. The model provides investigational early risk stratification among infants surviving to ROP assessment. It must complement, not replace, guideline-recommended ROP screening. Independent external validation and prospective clinical-impact evaluation are required before clinical implementation. Trial registration: Chinese Clinical Trial Registry, ChiCTR2400090262; registered on September 26, 2024, after study commencement (retrospectively registered).

Authors

Institutions

Publication Details

Journal
BMC Pediatrics
Published
2026-11-20
DOI
https://doi.org/10.1186/s12887-026-08061-1
Primary Topic
Retinopathy of Prematurity Studies
Type
article
Field-Weighted Citation Impact
0.00

Funders

Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Early prediction of severe retinopathy of prematurity in infants born at < 32 weeks’ gestation: a multicenter prediction model study

Zhangbin Yu, Ying Liu, Gen Lu, YanPing Guo et al.
2 links
BMC Pediatrics
Retinopathy of Prematurity Studies
article

Early prediction of severe retinopathy of prematurity in infants born at < 32 weeks’ gestation: a multicenter prediction model study

Zhangbin Yu, Ying Liu, Gen Lu, YanPing Guo, Guobing Chen, Rui Su, Huaqing Zhang, Jingjing Chen, Xudong Yan
article en

Abstract

Severe retinopathy of prematurity (ROP) is a vision-threatening condition among infants born at < 32 weeks’ gestation. We aimed to develop and internally validate an early prediction model for severe ROP using routinely available clinical variables. This prospective multicenter cohort study included infants born at < 32 weeks’ gestation admitted to 28 hospitals in Shenzhen, China, between January 2022 and December 2023. The prediction target was the conditional risk of severe ROP among infants surviving to ROP assessment. Severe ROP was defined as stage 3 or higher according to the International Classification of Retinopathy of Prematurity, Third Edition. Candidate predictors were maternal, perinatal, and early neonatal variables available at birth or within the first 24 h. A ridge logistic regression model was developed. Performance was assessed using discrimination, calibration, Brier score, decision curve analysis, bootstrap optimism correction, and leave-one-center-out internal-external validation. Among 1,453 initially identified infants, 112 died before ROP assessment and 5 duplicate records were excluded, leaving 1,336 infants for prediction modeling. Severe ROP occurred in 179 infants (13.4%). The apparent area under the receiver operating characteristic curve (AUC) was 0.844 (95% confidence interval [CI], 0.812–0.876). After 500 bootstrap resamples, the optimism-corrected AUC was 0.831, with an optimism-corrected Brier score of 0.089 and calibration slope of 1.192. In leave-one-center-out internal-external validation, the pooled AUC was 0.789 (95% CI, 0.749–0.827), with a Brier score of 0.094 and calibration slope of 1.031. The model provides investigational early risk stratification among infants surviving to ROP assessment. It must complement, not replace, guideline-recommended ROP screening. Independent external validation and prospective clinical-impact evaluation are required before clinical implementation. Trial registration: Chinese Clinical Trial Registry, ChiCTR2400090262; registered on September 26, 2024, after study commencement (retrospectively registered).

BMC Pediatrics
Jinan University (CN), Southern University of Science and Technology (CN), ShenZhen People’s Hospital (CN), Guangzhou Women and Children Medical Center (CN), Peking University Shenzhen Hospital (CN), Guangzhou Medical University (CN)
Shenzhen Science and Technology Innovation Program
Good health and well-being
Retinopathy of Prematurity Studies
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon