Abstract
Purpose: :
To determine whether review of dynamic flickering images, generated by digital registration of images from different time points, can improve accuracy and speed of detecting plus disease progression in retinopathy of prematurity (ROP) compared to the standard method of reviewing static side-by-side images.
Methods: :
A set of 15 de-identified wide-angle retinal image pairs from infants who developed plus disease were uploaded to a secure study website. Image pairs representing plus disease progression were taken at least 1 week apart, and image pairs representing controls were taken on the same day. Dynamic flickering image pairs were obtained using digital registration software (EyeIC, Narberth, PA). Ten ROP experts independently reviewed each image pair on the study website, initially displayed as static side-by-side images and subsequently displayed as dynamic flickering pairs. Experts were asked to identify which image in each pair exhibited increased severity, or to state the images were identical. Speed of diagnosis was measured using a computer timestamp, and accuracy of detecting change in ROP severity was measured using examination date and ophthalmoscopic findings as a reference standard.
Results: :
Diagnostic speed was significantly faster by by review of dynamic flickering image pairs (24.7±8.3 seconds) than by static side-by-side pairs (40.3±18.3 seconds) (p=0.002). For static side-by-side image pairs, experts made an accurate response in a mean (SD, %) of 11.4 (1.65, 76%) out of 15 image pairs. For dynamic flickering image pairs, experts made an accurate response in a mean (SD, %) of 11.3 (1.70, 75%) of 15 image pairs. There was no statistically significant difference in the accuracy of identifying plus disease progression between the two methods.
Conclusions: :
Identification of plus disease progression was significantly faster by review of dynamic flickering image pairs than by review of standard static images, although there was no difference in accuracy. Dynamic flickering with digital registration may warrant further investigation as a diagnostic method for ROP progression.
Keywords: retinopathy of prematurity • imaging/image analysis: clinical