Prognosis for a quicker diagnosis

By Cheryl Guttman Krader

Reviewed by Paolo S. Silva, MD

Boston-Advances in imaging technology means patients may benefit from faster image acquisition and disease evaluation.

Nonmydriatic ultra-wide-field imaging (Optos P200MA, Optos) compares favorably with dilated ETDRS protocol fundus photography and dilated fundus examination in determining the severity of both diabetic retinopathy (DR) and diabetic macular edema (DME), according to the results of a clinical validation study performed by researchers at the Beetham Eye Institute, Joslin Diabetes Center, Harvard Medical School, Boston.

The ultra-wide-field retinal imaging system combines scanning laser (green 532-nm and red 633-nm) ophthalmoscope technology with a large ellipsoidal mirror to image up to 200° (82%) of the retina in a single, 0.25-second scan. A high-resolution mode (ResMax, Optos) scans a 100° area of the retina (25%), but offers increased image resolution (11 µm).

“Compared with ETDRS 30° color fundus photography, the new imaging technology has several advantages as it avoids the need for pupil dilation, offers a wider imaging field, has faster acquisition time, and may penetrate better through media opacity and smaller pupils,” said Paolo S. Silva, MD, lead author of a published paper on the study [Am J Ophthalmol. 2012;154:549-559] and assistant chief, Center for Ocular Telehealth, Beetham Eye Institute.

“If our findings on its use for evaluating DR are confirmed in larger studies in broader diabetic populations, nonmydriatic ultra-wide-field imaging may prove to be beneficial in research and clinical settings,” Dr. Silva said. “To determine whether additional prognostic information can be obtained from peripheral retinal lesions, ultra-wide-field images of diabetic and non posterior pole will need to be evaluated carefully over time in a defined population. These peripheral findings may or may not provide additional prognostic information for DR progression.”

Validation study

The validation study enrolled 103 patients with type 1 or 2 diabetes who were attending a regularly scheduled clinic appointment. At a single visit, each patient underwent imaging of both eyes using the nonmydriatic ultra-wide-field technology (100° and 200° fields) as well as dilated ETDRS photography and dilated fundus examination. The nonmydriatic imaging was performed first by a certified imager, the ETDRS photographs were acquired by an ETDRS certified photographer, and the dilated fundus examination was performed by a retina specialist.

All of the images acquired were graded for DR and DME severity by two independent masked readers; any discrepancies between the two readers were adjudicated by a third masked retina specialist.

ETDRS photography was completed for 200 eyes and the readings showed the group represented the full spectrum of DR from none (12.5%) to proliferative (2.5%); the majority of eyes were graded as having mild or moderate nonproliferative DR (54%). The majority of eyes had no DME (57%) and nearly one-fourth were determined to have clinically significant DME.

Results from grading of the ultra-wide-field 100° images showed exact agreement with the ETDRS photography-based interpretations in 84% of eyes and there was a difference of just 1 severity level in another 7%. When compared with clinical examination, the readings from the nonmydriatic ultra-wide-field images were an exact match in 70% of eyes and were within 1 level in 93%.

The ultra-wide-field images had 99% sensitivity for detecting and identifying DR on ETDRS photographs and 100% specificity. Sensitivity and specificity for detecting and identifying PDR on ETDRS photographs using the ultra-wide-field imaging was 100% and 92%, respectively.

The researchers also noted that acquisition of the nonmydriatic ultra-wide-field images took less than half the time needed for ETDRS photography, even excluding the time for pupil dilation.

Although the ultra-wide-field imaging captures peripheral lesions, their implications on DR severity are unknown at this time.

“Current knowledge of DR progression rates along with recommendations for DR management is based on lesions located within the 7 standard 30° fields following the ETDRS imaging protocol,” he said. “It is uncertain how peripheral lesions would impact these recommendations.”

Challenges to overcome

Dr. Silva acknowledged that the imaging technology has some limitations. The cost is substantially higher compared with traditional methods of retinal imaging, and there is the potential for imaging artifacts not seen in traditional fundus photography due to the mirror, lashes, motion, and reflections.

In addition, the composite color image is based on red and green wavelengths only. Due to the curvature of the globe and the mirror, in 200° images there is some loss of resolution beyond the superior and inferior vascular arcades. Due to the large contiguous area of retina that is acquired in a single image, it is necessary to zoom in to sufficient magnification in order to obtain an adequate view necessary to appreciate subtle retinal lesions.

“With the 200° image there is good resolution along a central horizontal band that encompasses the disc and macula and extends to the temporal and nasal periphery,” Dr. Silva concluded.OT

FYI

Paolo S. Silva, MD

E-mail: paoloantonio.silva@joslin.harvard.edu

Dr. Silva discloses no relevant financial interests. Grant funding was provided to the Joslin Diabetes Center for performance of the study.