Perifoveal interdigitation zone loss in hydroxychloroquine toxicity leads to subclinical bull's eye lesion appearance on near-infrared reflectance imaging

Abstract

Purpose: To characterize the ultrastructural and functional correlates of hydroxychloroquine (HCQ)-induced subclinical bull's eye lesion seen on near-infrared reflectance (NIR) imaging.

Methods: An asymptomatic 54-year-old male taking HCQ presented with paracentral ring-like scotoma, abnormal multifocal electroretinography (mfERG) and preserved ellipsoid zone on optical coherence tomography (OCT). Dense raster OCT was performed to create en face reflectivity maps of the interdigitation zone. Macular Integrity Assessment (MAIA) microperimetry and mfERG findings were compared with NIR imaging, en face OCT, retinal thickness profiles and wave-guiding cone density maps derived from flood-illumination adaptive optics (AO) retinal photography.

Results: The bull's eye lesion is an oval annular zone of increased reflectivity on NIR with an outer diameter of 1450 µm. This region corresponds exactly to an area of preserved interdigitation zone reflectivity in en face OCT images and of normal cone density on AO imaging. Immediately surrounding the bull's eye lesion is an annular zone (3°-12° eccentricity) of depressed retinal sensitivity on MAIA and reduced amplitude density on mfERG. Wave-guiding cone density at 2° temporal was 25,400 per mm². This declined rapidly to 12,900 and 1200 per mm² at 3° and 4°.

Conclusion: Multimodal imaging illustrated pathology in the area surrounding the NIR bull's eye, characterized by reduced reflectance, wave-guiding cone density and retinal function. Further studies are required to investigate whether the bull's eye on NIR imaging and en face OCT is prominent or consistent enough for diagnostic use.

Keywords: Adaptive optics; Bull’s eye maculopathy; En face optical coherence tomography; Fundus autofluorescence; Microperimetry; Multifocal electroretinography.

Fig. 1

10–2 Macular Integrity Assessment (MAIA) microperimetry showing measured retinal sensitivity values in the right (a) and left (b) maculae in 2015. The location of fixation is shown by the dots scattered at the foveal center. Color scales for the symbols are shown below the image. Normal sensitivity is between 25 and 35 deciBel (dB)

Fig. 2

Multifocal electroretinography trace arrays of the right (a) and left (b) eye displayed in a retinal view perspective (S = superior, T = temporal, N = nasal and I = inferior) showing reduced amplitude densities in rings 3 and 4 in both eyes. Scalar plots for amplitude densities in right (c) and left (d) eyes and implicit times in right (e) and left (f) eyes are shown in nanoVolt (nV)/degree² and millisecond (ms) scales (color bars)

Fig. 3

A 30° × 30° short-wavelength fundus autofluorescence (SWAF) imaging of the right eye in hydroxychloroquine (HCQ) toxicity case (a) and a control healthy subject (b) showing no obvious differences in autofluorescence signal from the foveal and perifoveal region. Near-infrared reflectance (NIR) imaging from the patient (c) shows an annular oval region of increased reflectance with an inner diameter of 410 µm and an outer diameter of 1450 µm horizontally and 430 and 1410 µm vertically. This bull’s eye ring lesion is absent in the healthy control (d). En face visualization of the interdigitation zone (IDZ) on optical coherence tomography (OCT), as shown in the insert (red lines demarcate the IDZ located between ellipsoid and retinal pigment epithelium), shows a region of preserved IDZ reflectivity coinciding with the NIR bull’s eye lesion in the case of HCQ toxicity (e), but there is no such ring in controls (f)

Fig. 4

Overlay of the montage of adaptive optics (AO) cone images over near-infrared reflectance in the patient with toxicity (a) and a healthy subject (b) with a horizontal scale (yellow) marking the spacing between one degree from the foveal center. Insert shows zoomed-in images of cone reflexes at 2°, 3° and 4° eccentricities demonstrating the difference in cone packing at 3° and 4°, but not at 2°. Bright round signals represent wave-guiding cone outer segments. Overlay of cone density color map on near-infrared reflectance in the patient with toxicity (c) and a healthy subject (d) showing a central zone of reduced density due to inability of the rtx1 AO camera to resolve cones at the foveal center. Peak cone density is normally found at 2° eccentricity, and this is also seen in the eye with the bull’s eye lesion. However, the density of wave-guiding cones drops off rapidly from 3° eccentricity, outside the region of the bull’s eye lesion as seen on near-infrared reflectance. Color scale of cone density is shown below

Fig. 5

Graphical display of retinal sensitivity in deciBel (a, b), amplitude density of multifocal electroretinography (c, d), retinal thickness on optical coherence tomography (e, f) and cone density on adaptive optics retinal imaging (g, h) in the horizontal (left side) and vertical (right side) meridians. X-axis shows eccentricity in degrees from foveal center. Corresponding near-infrared reflectance images (i, j) showing the extent of the bull’s eye lesion. Bars represent −2 to +2 standard deviation of the normative data. Red circles = right eye data. Blue circles = left eye data. X-axis location of the measured value corresponds to exact location or the center of the zone of measurement. Retinal thickness is taken from central subfield, inner ring and outer ring of the Early Treatment for Diabetic Retinopathy Study grid