Abstract | BACKGROUND: DESIGN: Retrospective study in a quaternary public neonatal intensive care unit. PARTICIPANTS: METHODS: Retrospective chart and photo review were performed on participants screened by both serial wide-field digital retinal imaging and concurrent binocular indirect ophthalmoscopy. Review of captured digital photos was performed independently by a masked reader. Using the binocular indirect ophthalmoscopy findings as the gold standard, the efficacy of wide-field digital retinal imaging in detecting treatment-requiring retinopathy of prematurity, defined as type 1 prethreshold disease, was determined. MAIN OUTCOME MEASURES: RESULTS: Treatment-requiring retinopathy of prematurity was detected in 11 infants by both binocular indirect ophthalmoscopy examination and telemedicine images taken at the same visit. Wide-field digital retinal imaging has a sensitivity of 100% (95% CI: 76.2-100%) and a specificity of 97.9% (95% CI: 93.4-99.7%) in detecting infants with treatment-requiring retinopathy of prematurity. Positive and negative predictive values of wide-field digital retinal imaging were 84.6% (95% CI: 57.8-97.3%) and 100% (95% CI: 96.9-100%), respectively. CONCLUSIONS:
|
Authors | Shuan Dai, Kent Chow, Andrea Vincent |
Journal | Clinical & experimental ophthalmology
(Clin Exp Ophthalmol)
Vol. 39
Issue 1
Pg. 23-9
(Jan 2011)
ISSN: 1442-9071 [Electronic] Australia |
PMID | 20796264
(Publication Type: Journal Article)
|
Copyright | © 2010 The Authors. Clinical and Experimental Ophthalmology © 2010 Royal Australian and New Zealand College of Ophthalmologists. |
Topics |
- Birth Weight
- False Positive Reactions
- Female
- Gestational Age
- Humans
- Image Processing, Computer-Assisted
- Infant, Newborn
- Infant, Premature
- Intensive Care Units, Neonatal
- Male
- Ophthalmoscopy
- Photography
- Predictive Value of Tests
- Reproducibility of Results
- Retina
(pathology)
- Retinopathy of Prematurity
(classification, diagnosis, therapy)
- Retrospective Studies
- Sensitivity and Specificity
- Telemedicine
|