《产后生长和早产儿视网膜病变(G-ROP)筛查标准》(简化版)在严重早产儿视网膜病变筛查中的应用

Application of the simplified postnatal growth and retinopathy of prematurity screening criteria in detecting severe retinopathy of prematurity

目的评估简化版产后生长和早产儿视网膜病变(G-ROP)筛查标准检测严重早产儿视网膜病变(ROP)的准确性,并与现行的筛查标准及本研究推荐的筛查标准进行比较。方法回顾性选取2015年1月至2022年12月在四川省人民医院进行的ROP筛查的早产儿1286例为研究对象。所有受检者均在散瞳后使用双目间接眼底镜或广角数码儿童视网膜成像系统(Retcam)按常规操作进行眼底检查。详细收集受检者出生体重、出生孕周、性别以及眼底检查结果等临床资料。根据《中国早产儿视网膜病变筛查指南(2014年)》对眼底病变进行诊断及治疗(筛查标准为BW<2000 g,或GA<32周)。简化版G-ROP筛查标准定义为:出生体重<1051 g,或出生孕周<28周;推荐的筛查标准定义为:出生体重1500 g,或出生孕周≤30周。应用受试者工作特征(ROC)曲线对筛查标准检测严重ROP的性能进行分析。结果共302例(23.5%,302/1286)受检者符合简化版G-ROP筛查标准,其中159例(52.6%,159/302)为严重ROP。共134例(45.7%,134/293)严重ROP被漏诊。ROC分析结果显示,简化版G-ROP筛查标准检测严重ROP患儿的敏感性与特异性分别为37.54%与88.52%,AUC为0.630;《中国早产儿视网膜病变筛查指南(2014年)》筛查严重ROP患儿的敏感性与特异性分别为98.90%与18.40%,AUC为0.587;推荐筛查标准检测严重ROP患儿的敏感性与特异性分别为91.80%与47.00%,AUC为0.694。推荐筛查标准较《中国早产儿视网膜病变筛查指南(2014年)》减少380例(29.5%,380/1286)受检者,仅7例(2.4%,7/293)严重ROP患儿漏诊。结论简化版G-ROP筛查标准对严重ROP检测的准确性欠佳。将筛查标准调整为BW≤1500 g,或GA≤30周更适合我国ROP的筛查策略。

Objective To evaluate the accuracy of simplified postnatal growth and retinopathy of prematurity(G-ROP)screening criteria in detecting severe retinopathy of prematurity(ROP)and compare it with the current screening criteria and the recommended screening criteria in this study.Methods A total of 1286 premature infants screened for ROP at Sichuan Provincial People’s Hospital from January 2015 to December 2022 were retrospectively selected.All patients underwent routine fundus examination using the binocular indirect ophthalmoscope or wide-angle digital pediatric retinal imaging system(Retcam)after pupil dilation.Detailed clinical data,such as birth weight(BW),gestational age(GA),gender,and fundus examination results,were collected.Fundus lesions were diagnosed and treated according to the Chinese Guidelines for Screening Retinopathy of Prematurity(2014)(the screening criteria were BW<2000 g or GA<32 weeks).The simplified G-ROP screening criteria were defined as BW<1051 g or GA<28 weeks.The recommended screening criteria were defined as BW≤1500 g or GA≤30 weeks.The performance of screening criteria for detecting severe ROP was analyzed using the receiver operating characteristic(ROC)curve.Results A total of 302 infants(23.5%,302/1286)met the simplified G-ROP screening criteria,of which 159(52.6%,159/302)had severe ROP.A total of 134 infants(45.7%,134/293)with severe ROP were misdiagnosed.The ROC analysis results showed that the sensitivity and specificity of the simplified G-ROP screening criteria for detecting severe ROP were 37.54%and 88.52%,respectively,with an area under the curve(AUC)of 0.630.The sensitivity and specificity for detecting severe ROP according to the Chinese Guidelines for Screening Retinopathy of Prematurity(2014)were 98.9%and 18.4%,respectively,with an AUC of 0.587.The sensitivity and specificity for detecting severe ROP according to the recommended screening criteria were 91.8%and 47.0%,respectively,with an AUC of 0.694.The recommended screening criteria had reduced the number of screened infants by 380(29.5%,380/1286)compared to the Chinese Guidelines for Screening Retinopathy of Prematurity(2014),and only 7(2.4%,7/293)with severe ROP were misdiagnosed.Conclusion The simplified G-ROP screening criteria have poor accuracy in detecting severe ROP.Adjusting the criteria to BW≤1500 g or GA≤30 weeks is more suitable for the screening strategy of ROP in China.