Yen Kimberly G, Hess Ditte, Burke Barbara, Johnson Rose Anne, Feuer William J, Flynn John T
Department of Ophthalmology, Moran Eye Center, University of Utah, Salt, Lake City, Utah, USA.
J AAPOS. 2002 Apr;6(2):64-70.
Labor-intensive screening of infants in the neonatal intensive care units is the only way presently to detect retinopathy of prematurity (ROP). Our purpose is to determine if RetCam 120 images (Massie Research Laboratories, Inc, Dublin, Calif), acquired by a neonatal nurse, can be used to screen for ROP by performing 2 screening sessions, at 32 to 34 weeks' (examination 1) and 38 to 40 weeks' (examination 2) postconceptional age.
RetCam examinations were performed by a nurse on infants at examination 1 and examination 2 intervals. At the same time, an examination was performed by an experienced ophthalmologist. Masked readers evaluated the digital images for the presence of ROP and, if ROP was present, estimated the risk of that eye progressing to prethreshold or threshold disease. The data were compared to the eye's clinical course.
A total of 46 eyes were assessed at examination 1 and 50 eyes at examination 2 from July 1, 1999, to December 15, 1999. For detecting ROP, the sensitivity and specificity were 46% and 100% for examination 1 and 76% and 100% for examination 2. Sensitivity and specificity of predicting prethreshold was 64% and 97%, respectively, for examination 1 and 2. Sensitivity for predicting ROP threshold at examination 1 was 0% (only 1 photo was available for grading of sensitivity) and specificity for predicting ROP threshold at examination 1 was 95%. At examination 2, sensitivity and specificity were 100%.
The RetCam examination had insufficient sensitivity to be recommended as a substitute for indirect ophthalmoscopy in screening for ROP. Reasons for low sensitivity are the technical limitation of the camera design itself, which creates difficulty in photographing the peripheral retina in small eyes, and the need for a lid speculum better adapted to the contact camera optical system design. Both of these issues are being addressed as part of an ongoing project to study the feasibility of employing telemetry of digital fundus images from remote, underserved neonatal intensive care units to ophthalmologists capable of diagnosing ROP.
在新生儿重症监护病房对婴儿进行劳动密集型筛查是目前检测早产儿视网膜病变（ROP）的唯一方法。我们的目的是确定由新生儿护士获取的RetCam 120图像（马西研究实验室，加利福尼亚州都柏林）是否可用于通过在孕龄32至34周（检查1）和38至40周（检查2）进行两次筛查来筛查ROP。
护士在检查1和检查2的间隔对婴儿进行RetCam检查。同时，由经验丰富的眼科医生进行检查。蒙面阅片者评估数字图像中是否存在ROP，如果存在ROP，则估计该眼进展为阈值前或阈值疾病的风险。将数据与眼睛的临床病程进行比较。
从1999年7月1日至1999年12月15日，检查1共评估了46只眼，检查2共评估了50只眼。对于检测ROP，检查1的敏感性和特异性分别为46%和100%，检查2的敏感性和特异性分别为76%和100%。预测阈值前的敏感性和特异性在检查1和检查2时分别为64%和97%。检查1时预测ROP阈值的敏感性为0%（仅1张照片可用于敏感性分级），检查1时预测ROP阈值的特异性为95%。检查2时，敏感性和特异性均为100%。
RetCam检查的敏感性不足以推荐作为筛查ROP时间接检眼镜的替代方法。敏感性低的原因是相机设计本身的技术限制，这使得在小眼睛中拍摄周边视网膜困难，以及需要一个更适合接触式相机光学系统设计的眼睑撑开器。作为一个正在进行的项目的一部分，这两个问题都正在得到解决，该项目旨在研究将来自偏远、服务不足的新生儿重症监护病房的数字眼底图像遥测给能够诊断ROP的眼科医生的可行性。
