Purchase this article with an account.
Tian Xia, Ben Szirth, Albert S Khouri; Anterior segment smartphone imaging quality in telemedicine. Invest. Ophthalmol. Vis. Sci. 2016;57(12):5930.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Smartphone imaging in ophthalmic telemedicine has gained popularity. Limited studies have assessed image quality acquired through smartphones. We performed an evaluation of image quality of anterior segment images acquired with smartphones.
Smartphone anterior segment images acquired without adapter assistance using various model smartphones during telemedicine were collected. Image quality was assessed on a 1-5 scale by three masked clinicians of varying training levels (evaluator 1 ophthalmologist, 2 resident, 3 medical student). Quality grading scale was as follows: grade 1 (of no clinical value, barely visible), 2 (only a portion visible, not able to evaluate findings clearly), 3 (majority visible, able to evaluate abnormalities), 4 (near perfect image, able to evaluate most subtle findings), and 5 (perfect image, able to evaluate all subtle findings). Data on smartphone model, acquisition technique, image transmission method were obtained. Kappa coefficient was calculated to determine inter-observer image quality score agreement. Mean, standard deviation (SD), and student t- test were performed.
53 images were included. Smartphone models used for acquisition were: 7.5% HTC, 7.5% iPhone 4s, 15% iPhone 5, 2% iPhone 5c, 55% iPhone 6, and 13% iPhone 6s. Imaging techniques included: 92% through slit lamp biomicroscope, 4% through surgical microscope and 4% non-magnified. 80% of images were transferred through text message and 20% through email. All images were graded on iPhone screen with 1136 x 640 pixels standard resolution. Mean±SD of quality score were 4.05±0.79, 3.98±0.88, and 3.59±0.92, for evaluators 1, 2, and 3, respectively. The inter-observer image quality score: kappa coefficient of evaluator 1 and 2 was 0.22±0.097, showing fair agreement. There was a difference in mean image quality between evaluator 1 vs 3 and evaluator 2 vs. 3 (p<0.05). There was a difference in the image quality among different types of acquiring phones (p<0.05). There was no difference in imaging quality of emailed vs. texted images.
Successful and good quality imaging may be acquired with smartphones for use in telemedicine. Variable image quality by grader and smartphone highlights the need for standards in telemedicine and raises clinician awareness of pitfalls in relying on non-validated emerging trends in smartphone imaging.
This is an abstract that was submitted for the 2016 ARVO Annual Meeting, held in Seattle, Wash., May 1-5, 2016.
This PDF is available to Subscribers Only