Blue-light reflectance imaging in macular telangiectasia type 2 (MacTel) using two different SLO systems.

Abstract

Purpose : MacTel is a disease of the central retina, leading over time to photoreceptor degeneration and potentially to bilateral loss of central vision. One of the earliest signs of MacTel is a pattern of abnormal parafoveal reflectance, thought to originate in the inner retina due to increased scattering caused by structural reorganisation in these layers. Reflective scanning laser ophthalmoscopic (SLO) imaging using short wavelength light (blue light reflectance, BLR imaging) provides a superior tool for recording this phenomenon. Our aim was to investigate how different SLO systems compare in BLR imaging in MacTel.

Methods : Eyes of participants with a diagnosis of MacTel were selected from the cohort of the MacTel Study on the basis of the availability of confocal blue-light reflectance (BLR) images collected using the two SLO systems of interest. Images were acquired using a Heidelberg Retina Angiograph II and a Heidelberg Spectralis SLO system (Heidelberg Engineering GmbH, Heidelberg, Germany) using the 'red free' setting (utilizing a 488nm blue laser for illumination of the retina). Images were aligned using I2K Retina (v1.3.8, DualAlign™ LLC) and analyzed in Adobe Photoshop v CS6. The lateral extent of hyper-reflectivity was delineated and measured (expressed in pixels) and their relative areas and locations were compared. Agreement between measurements was assessed by calculation of Spearman's coefficient of rank correlation using MedCalc (v12.7.0.0, MedCalc Software BVBA, Ostend, Belgium).

Results : Full image sets were available for 21 eyes of 11 patients (5 females, 6 males), ranging in age between 46-71 years, (mean=58.2 years, SD=7.3 years). MacTel disease severity ranged from stage 1-5 on the Gass & Blodi scale. Agreement of area measurements of BLR hyper-reflectivity was good, Spearman's coefficient of rank correlation (ρ) = 0.868 (95%CI 0.664 - 0.951, p=<0.0001, n=17) In two eyes no hyper-reflective area was apparent and in another two eyes the hyper-reflective area could not be delineated accurately. A specific pattern was discernible: Spectralis images often contained central round artifacts whereas HRA2 images demonstrated an hourglass-shaped reduction in luminance.

Conclusions : While the correlation of BLR lesion extent is good, studies aiming for a quantitative analysis of BLR intensity must take into account the technical differences between HRA2 and Spectralis SLO devices.

This abstract was presented at the 2019 ARVO Annual Meeting, held in Vancouver, Canada, April 28 - May 2, 2019.