Pediatric fundus exam can be limited due to lack of cooperation. Widefield digital imaging is capable of capturing images in less than 1 minute using a noncontact system.1 Previous studies have demonstrated comparable ability of widefield imaging to detect pathology compared with indirect ophthalmoscopy in adults.2 Here, we aimed to determine if high quality ultra-widefield (UWF) images can be reliably obtained from dilated pediatric patients, and compared them to indirect ophthalmoscopy.
Patients and Methods
This study was approved by the Institutional Review Board of the University of California San Diego and followed the Declaration of Helsinki. In this study, 173 patients between the ages of 3 years and 17 years were enrolled for retinal evaluation and UWF retina photography at the University of California, San Diego from January to December 2016.
Each physician in the study completed survey questions before photography, after the dilated examination, and after viewing the UWF photo. Prior to photography, the patient's ethnicity, pre-exiting retinal diagnosis, and cooperativity (low, medium, or high) were measured. The patient received dilation with a combination drop comprised of cyclopentolate 2%, tropicamide 1%, and phenylephrine 10% in a 6:3:2 ratio.
After retinal photos were taken, the physician measured pupil size using the standardized ruler on a pocket reading card. Media clarity was measured as poor, moderately hazy, or clear. After completing the dilated examination, the area of the retina evaluated was drawn on a schematic fundus drawing including retinal landmarks (the macula, vortex veins, and anterior retina). After reviewing the UWF image, the physician answered questions aimed at evaluating the helpfulness of the image for clinical decision decision-making.
The physicians who performed the fundus examination and answered the survey were both pediatric ophthalmologists and retina specialists. Many more patients were enrolled from the pediatric ophthalmology clinic than the retina clinic (Table 1).
Using the California fluorescein angiography/indocyanine green angiography machine (Optos, Dunfermline, UK), color composite red and green laser photographs were obtained by one of four licensed and trained ophthalmic photographers with at least 5 years of experience. The Optos Retinal Imaging System was used on a table with adjustable height, and the child's head was guided into the headrest while the photographer took the photograph. Although the patient was encouraged to open his or her eyes, no lid-opening devices were used. If the upper lid obscured the view and, if tolerated by the child, the lid was gently elevated. Images were taken 5 to 10 minutes after the patient received one dilating drop in each eye.
Image Area Analysis
The Optos UWF imaging system, a scanning ophthalmoscope technology, utilizes an ellipsoid mirror to capture a retinal image spanning a maximum of 200 internal degrees.4 Green and red lasers produce pseudocolor, high-quality retinal images. The Optomap software produces a stereographic projection of the fundus by formatting the image to maximally fit a 24-mm sphere.4 The pixels of this conformal image of the three-dimensional fundus are projected upon a two-dimensional plane located at the eye's equator to produce an image viewable on a flat surface, such as a computer screen. The resulting image is bijective, allowing for the accurate measurement of retinal areas through this program by accounting for axial length.5 The area of the retina evaluated on the UWF image was calculated using the Optomap software freehand tool by a physician who was not involved in the clinical care of the patient.
Image Quality Analysis
Retinal image quality was then measured by a physician masked to the patient's age. Image quality was graded on a quantitative scale of 0 to 4. A quality score of 0 corresponded to “No image obtained.” A score of 1 corresponded to “less than two quadrants of the retina visualized or retinal details visible in 0% to 25% of the photo area,” a score of 2 corresponded to “two to four quadrants of the retina visualized or retinal details visible in 25% to 50% of the photo area,” a score of 3 corresponded to “all four quadrants visualized and retinal details visible in 50% to 75% of the photo area, and a score of 4 corresponded to “all four quadrants visualized and retinal details visible in more than 75% of the photo area.” Each eye was given a separate score.
Clinical Area Measurement
An image grader took the hard copy of the clinical drawing with the retinal area evaluated on indirect ophthalmoscopy and transposed this onto the patient's digital retinal image using the drawing tool in the Optomap software to determine the retinal area evaluated during ophthalmoscopy (Figure 2B). Area measurements were performed by two separate image graders. Disagreements were adjudicated by a third image grader.
Methodology for retinal area measurements. (A) Ultra-widefield photograph (UWF) in a 4-year-old girl with left optic nerve hypoplasia. (B) Clinical drawing with area drawn on examination form by clinician. (C) Transposition of the clinical drawing freestyle to the UWF image. The area of retina evaluated with clinical exam was generated by the Optomap software. (D) UWF photograph with evaluated retina circled with the freestyle drawing tool. The retinal area measurement was generated by the Optomap software.
Measurements were compared between groups using linear mixed models with subject as a random effect to account for the fact that subjects had more than one eye included in the study. When analyzing the effect of retinal diagnosis on additional clinical information, a chi-square test was utilized with one data point per subject. The relationship between observed patient cooperativity and average image quality was assessed using the chi-square test, logistic regression, and Spearman's rank correlation coefficient. All statistical analyses were performed with R statistical software.6
A total of 99 male and 74 female patients aged between 3 and 17 years were enrolled. The 6- to 8-year-old age group accounted for 30% of total enrolled patients. Fifty-three patients (31%) had an existing retinal diagnosis, and the majority (146 of 173) patients were enrolled from the pediatric ophthalmology clinic (Table 1).
UWF images of at least one eye were obtained in 91% of enrolled patients in the 3- through 5-year-old age group and 98% of patients in the 6- through 8-year-old age group. Images of both eyes were obtained in 80% of patients in the 3- through 5-year-old age group, 94% of patients in the 6- through 8-year-old group, and 100% of patients in the 9- to 11-, 12- to 14-, and 15- to 17-year-old age groups. The primary reason cited for a patient's failure to complete the image capture of one or both eyes was lack of patient cooperation. Patient cooperativity scores increased with age, and image quality scores were positively associated with age (P < .001; Rho = 0.375) (Table 2). Logistic regression confirmed that high cooperativity leads to highest image quality (P < .001). Images obtained from patients ages 3 through 5 years were of the lowest quality (2.64 ± 1.43), whereas the highest quality images were obtained from patients in the 15-to-17-year-old age group (3.84 ± 0.59). Representative photographs of varying image quality are shown (Figure 1).
Patient Cooperativity and Image Quality Separated by Patient Age Group
Ultra-widefield photographs of varying quality. (A) Quality score 2, given that only three quadrants were visualized. (B) Quality score 3, as all four quadrants were visualized but only 50% to 75% of the retinal area had retinal details visible. (C) Quality score 4, with all four quadrants visualized and more than 75% of the quadrant has retinal details visible.
The area of the retina evaluated was measured from the dilated examination after being transposed with the Optomap software (Figure 2). When separated by age, the area of retina evaluated on UWF photograph was significantly larger than with the dilated examination in the 3-to-5, 6-to-8, 9-to-11, and 12-to-14-year-old age groups (P < .001 for all groups) (Figure 3).
Retina area measurements on dilated exam versus ultra-widefield (UWF) photograph by age. Mean retinal areas on the dilated clinical exam (Clin ex) are compared to the UWF photograph (Optos) with standard deviations presented with error bars.
The UWF photograph did not evaluate more of the retina when compared with the dilated exam in patients with media opacities. There was no difference in the retinal area on UWF images with and without a retinal diagnosis. However, using a linear mixed model, the retinal area evaluated on clinical examination was more in patients with a retinal diagnosis (P < .001.)
The majority (60%) of physician responses in this study indicated that the UWF retinal photograph did not provide additional examination information when compared with the clinical exam (Supplemental Table 3). Although reviewing the UWF image after indirect ophthalmoscopy did not alter the referring physician's clinical diagnosis 98% of the time, it increased physician confidence in the peripheral retina exams in 44% of patients and helped avoid the need for three examinations under anesthesia in this population. If a patient had a pre-existing retinal diagnosis, the photograph offered additional clinical information 53% of the time, compared to 62.5% of the time if the patient did not have a retinal diagnosis. In 22% (38 of 173) of patients, the UWF image provided data that was difficult for the ordering physician to interpret. The UWF image was helpful with family counseling in 17% of cases. There was no significant difference in area of the retina evaluated on UWF photography when patients of varying pupil size and ethnicity were compared.
Ultra-Widefield Photograph Clinical Utility Evaluation by Physicians
This prospective study found that significantly more retinal area can be evaluated with UWF fundus photography than with the in-office dilated examination in all age groups. Although only 69% of patients in the 3- to 5-year-old age group were considered to be either moderately or highly cooperative, images were obtained in 91% of enrolled patients in this age range. High quality images can be obtained in children in the 3- to 5-year-old age group, and, as expected, image quality scores, age, and cooperativity scores are all significantly associated. The strong correlation between the physician's assessment of patient cooperativity and image quality indicates that the dataset is in this population is reliable.
Given the large (22%) percentage of patients in which the retinal photograph revealed an artifact or a finding that was difficult to interpret for the ordering physician, the authors do not believe that the UWF photograph is sufficient to replace indirect ophthalmoscopy for retinal evaluation. The photograph led to several false positive findings that, after examination were proven to not be present. This highlights the value of the photograph as an adjunctive tool, but not as a replacement for physical exam.
There are several limitations of this study. First, although the physician scoring the retinal image quality and area was masked to the clinical examination, the physician transposing the clinical examination sheet to the Optomap software image was not masked to the patient's UWF fundus photograph. It was necessary to use the patient's own fundus photograph to transpose the drawing to get accurate an area calculation because the software takes into account the size of the patient's eye. To reduce this bias, area measurements were performed by two separate image graders. Second, this study was not designed to determine if the UWF fundus photograph is as accurate as the clinical examination in diagnosis. Although the area of the retina evaluated may be greater, the authors did not do a masked evaluation of the retinal diagnosis, so no conclusions about the use of the UWF photograph alone as a diagnostic tool can be made.
Most ophthalmologists participating in this study indicated that having access to these images did not significantly change their patients' diagnosis or treatment plan. Images can be readily obtained in the vast majority of patients, and the photograph increased physician confidence in the peripheral retinal exam for pediatric ophthalmologists.
In summary, our results indicate that high-quality UWF fundus photography is possible even in very young children. This tool allows for evaluation of more peripheral retina than an in-office dilated fundoscopic examination in patients ages 3 years to 12 years presenting to pediatric ophthalmology clinics. Because it often produced data that were difficult to interpret, it is a useful screening and adjunctive clinical tool but cannot replace an examination by indirect ophthalmoscopy. In the retina clinics, the greatest benefits are in documentation of the retinal exam, family counseling, and monitoring disease activity. It also prevented examinations under anesthesia in several patients. UWF imaging in the pediatric population is a feasible and effective adjunctive tool in the pediatric population. It may allow for more substantial evaluation of patients. may lead to the earlier diagnosis of treatable retinal disease.
- Patel CK, Fung THM, Muqit MMK, et al. Non-contact ultra-widefield retinal imaging and fundus fluorescein angiography of an infant with incontinentia pigmenti without sedation in an ophthalmic office setting. J AAPOS. 2013;17(3):309–311. doi:10.1016/j.jaapos.2012.12.152 [CrossRef]23607978
- Pandya A, Friberg T, Eller A. Optos non-mydriatic widefield imaging vs. clinical dilated fundus exam for retinal diagnosis and management. Invest Ophthalmol Vis Sci. 2002;43:2860.
- Shoughy SS, Arevalo JF, Kozak I. Update on wide- and ultra-widefield retinal imaging. Indian J Ophthalmol. 2015;63(7):575–581. doi:10.4103/0301-4738.167122 [CrossRef]26458474
- Tan CS, Chew MC, Hemert J Van, Singer MA, Bell D, Sadda SR. Measuring the precise area of peripheral retinal non-perfusion using ultra-widefield imaging and its correlation with the ischaemic index. Br J Ophthalmol. 2016;100(2):235–239. doi:10.1136/bjophthalmol-2015-306652 [CrossRef]
- Sagong M, van Hemert J, Olmos de Koo LC, Barnett C, Sadda SR. Assessment of accuracy and precision of quantification of ultra-widefield images. Ophthalmology. 2015;122(4):864–866. doi:10.1016/j.ophtha.2014.11.016 [CrossRef]25576995
- R Development Core Team. R: A Language and Environment for Statistical Computing. Available at https://www.gbif.org/tool/81287/r-a-language-and-environment-for-statistical-computing. Updated February 10, 2015.
- Gupta V, Al-Dhibi HA, Arevalo JF. Retinal imaging in uveitis. Saudi J Ophthalmol. 2014;28(2):95–103. doi:10.1016/j.sjopt.2014.02.008 [CrossRef]24843301
- Madhusudhan S, Beare N. Wide-field fluorescein angiography in wet age-related macular degeneration. Scientific World Journal. 2014;2014:536161. doi:10.1155/2014/536161 [CrossRef]25379537
- Muqit MMK, Marcellino GR, Henson DB, et al. Optos-guided pattern scan laser (Pascal)-targeted retinal photocoagulation in proliferative diabetic retinopathy. Acta Ophthalmol. 2013;91(3):251–258. doi:10.1111/j.1755-3768.2011.02307.x [CrossRef]
|Age Groups (Years of Age)|
| South Asian||10|
| East Asian||11|
| African American||6|
| Declined to state||4|
|Pre-Existing Retinal Diagnosis|
|Clinic Patient Recruited From|
| Pediatric ophthalmology||146|
Patient Cooperativity and Image Quality Separated by Patient Age Group
|Age Group (Years of Age)||3–5 (n = 46)||6–8 (n = 52)||9–11 (n=35)||12–14 (n = 20)||15–17 (n = 20)|
|Age Group (Years of Age)||3–5 (n = 74)||6–8 (n = 98)||9–11 (n = 68)||12–14 (n = 38)||15–17 (n = 38)|
Ultra-Widefield Photograph Clinical Utility Evaluation by Physicians
|Age Group (Years of Age)|