Title Page:

Full Length Title:

Fundus autofluorescencein rubella retinopathy: correlation with photoreceptor structure and function

Running Title:

Autofluorescence in rubella retinopathy

Authors:

*Danuta M Bukowska, PhD1

*SueLingWan, MBBS (Hons), FRANZCO1,2

Avenell Chew, MBBS1

Enid Chelva, BSc (Hons)3

Ivy Tang, BSc (Hons) MOrth1

David A Mackey, MD, FRANZCO1

Fred K Chen, PhD, MBBS (Hons), FRANZCO1,2

* Equal first authors

Affiliations:

1. Centre for Ophthalmology and Visual Science, Lions Eye Institute, The University of Western Australia, Western Australia

2. Department of Ophthalmology, Royal Perth Hospital, Perth, Western Australia

3. Department of Medical Technology and Physics, Sir Charles Gairdner Hospital, Perth, Western Australia

Corresponding author:Fred K Chen, Centre for Ophthalmology and Visual Science

Lions Eye Institute, 2 Verdun Street, Nedlands, WA 6009

Email:

Funding source:NH&MRC Early Career Fellowship (APP1054712 , (FK Chen)

Ophthalmic Research Institute of Australia (DM Bukowska)

The authors report no conflicts of interest

Keywords:

adaptive optics, autofluorescence, congenital rubella syndrome, microperimetry, scanning laser ophthalmoscopy,spectral domain optical coherence tomography

Summary Statement:

We describe fundus autofluorescence features of 4 patients with rubella retinopathy. Hypoautofluorescent lesions on blue and near-infrared excitation did not correlate with attenuation of ellipsoid or interdigitation zone. There was no significant reduction in retinal sensitivity in regions with hypoautofluorescence. We discuss the mechanisms for autofluorescence features in rubella retinopathy.

ABSTRACT:

Purpose

To illustrate hypoautofluorescent and hyperautofluorescentaltered fundus autofluorescencelesions in rubella retinopathy and to investigate their relationships with photoreceptor structure and function using multimodal imaging.

Methods

We report 4 cases ofrubella retinopathy aged 8, 33, 42 and 50 years.All patients had dilated clinical fundus examination; wide-field color photography; blue, green and near-infrared autofluorescence imaging and spectral-domain optical coherence tomography (OCT). Twopatients also underwentmicroperimetry and adaptive optics imaging.En face OCT, cone mosaic and microperimetry were co-registered with autofluorescence images. We explored the structure-function correlation.

Results

All 4 patients had a “salt-and-pepper” appearance on dilated fundus examination and wide-field color photography. There werevariable-sized patches of hypoautofluorescence on both blue and near-infrared excitation in all 4 patients. Wave-guiding cones were visible and retinal sensitivity was intact over these regions.There was no correlation between hypoautofluorescence and regions of attenuated ellipsoid and interdigitation zones. Hyperautofluorescent lesionswere also noted and some of theseweredue to acquired pseudo-vitelliform lesions.

Conclusions:

Patchy hypoautofluorescenceon near-infrared excitation is a characteristic can be a unique feature of rubella retinopathy. This may be due to abnormal melanin production or loss of melanin within retinal pigment epithelium (RPE) cells harboring persistent rubella virus infection. Preservation of the ellipsoid zone, wave-guiding cones and retinal sensitivity within hypoautofluorescent lesions suggest that these RPE changes have only mild impact on photoreceptor cell function.

INTRODUCTION

Congenital rubella syndrome (CRS) imposes a heavy burden on both the affected individual and their community. CRS affects almost 100,000 children every year worldwide despite successful vaccination programs.1Moreover, vaccination programs do not exist in all countries, leaving many fetuses exposed to the risk of developing CRS.

A "salt-and-pepper" fundus appearance is commonly associated with CRS due to disruption of the normal embryogenesis of the retinal pigment epithelium (RPE). Histologically, these lesions correspond to RPE clumping and migration.2,3These RPE abnormalitiesy are particularly well-visualized on short wavelength (blue light) fundus autofluorescence (AF) imaging as stippled signal within the macula.4

Loss of AF or hypoautofluorescence (hypo-AF) is often associated with photoreceptor damage and retinal dysfunction.5 However, tThe "salt-and-pepper" retinopathy in CRS has been described as a non-progressive pigmentary retinopathy with that does not impair normal vision and or retinal electrophysiology response.

However this is not consistent, as some reports describe correlation between of hypoautofluorescence (hypo-AF) is with RPE and photoreceptor damage.5 Herein, using multimodal retinal imaging we examinede the relationship of between blue, green and near-infrared hypo-fundus AF,with and photoreceptor structure and function in4 cases of rubella retinopathy. We found disturbance of the interdigitation zone, formation of pseudo-vitelliform lesions and wide spread fundus AF abnormalities in congenital rubella. These features suggest abnormal interaction between photoreceptors and the RPE and melanin metabolism.

METHODS

Four patients aged (8, 33, 42 and50 years (Cases 1, 2, 3 and 4 respectively) old) with rubella retinopathy underwent Early Treatment Diabetic Retinopathy Study (ETDRS) visual acuity testing,complete ophthalmic examination by a retinal specialist (FKC) and multimodal imaging. The 50 year old patient was male and the other 3 were female. The diagnosis of rubella retinopathy was based on positive clinical signs together with a history of maternal infection during the first trimester. One patientThe 50yo male did not have this history available but had congenital sensorineural deafness and clinical features consistent with rubella retinopathy. No patient in this series had a history of using any medication known to cause retinal toxicity, such asincluding hydroxychloroquine, deferoxamine and chlorpromazine., etc

Wide-field color and medium wavelength AF imaging (excitation: 532 nm; green light, detector: 570-780 nm) were performed using a scanning laser ophthalmoscope (OptosP200MATx, Optos plc, Dunfermline, UK).Spectral domain optical coherence tomography (SD-OCT) and short wavelength (excitation: 486 nm; blue light, detector: 500-750 nm) and long wavelength near-infrared (excitation: 786 nm, near-infrared light, detector: 810-900 nm)AF imaging were performed using the Spectralis HRA+OCT (Heidelberg Engineering, Heidelberg, Germany). One patient also had retinal electrophysiology (incorporating the International Society for Clinical Electrophysiology of Vision standards) and two patients hadfundus-controlled microperimetry and adaptive optics (AO) retinal imaging(see below).

Spectralis images were captured in high-resolution mode (1,536 x 1,536 pixels). SD-OCT images were acquired concurrently with near-infrared reflectance imaging by using the following protocols: (1) a dense raster scan of the foveal region (20° x 10°) consisting of 97 horizontal lines (line separation of 30µm) and (2) macular volume scan of 30 x 25° consisting of 61 horizontal lines (lines separation of 61 µm).Each B-scan linewas obtained by averaging 9 single frames. The en-face minimum projection intensity reflectivity maps of the ellipsoid and interdigitation zones within the foveal region were created using the HEYEX 3D viewing module in Spectralis (version 6.0.9.0). We usedthe lexicon proposed by the International Nomenclature for Optical Coherence Tomography Consensus to describe structural changes in the OCT scan.6AF images were acquired with a 30°x 30° field of viewcentered on the fovea. One hundredAF frames were averaged in automated real time (ART) mode to acquire the final high definition AF images. If near-infrared AF could not be acquired, combined angiography mode (the simultaneous fluorescein and indocyanine green angiography mode) was used.7

Microperimetry was performed using the fundus-controlled microperimeter MAIA (Centervue, Padova, Italy). Pupils were dilated with tropicamide 1% and phenylephrine 2.5%. The sessions were conducted in a darkened room prior to retinal imaging. Each eye was tested with a 10-2 (68 loci) stimulus grid using a 4-2 staircase strategy to determine retinal sensitivity thresholds to the nearest 1dB.

Adaptive optics (AO) retinal imaging was performed through dilated pupils using a commercially available instrument, the rtx1camera (Imagine Eyes, Orsay, France), based on a flood-illumination ophthalmoscopy system. Each AO image frame is4° x 4° and consecutive images were acquired with 50% overlap in area of adjacent frames. MosaicJ plugin (ImageJ,Image Processing and Analysis in Java, NIH, Bethesda, USA) was used to manually create a wide-field montage of the AO images. Cone density maps were generated using AOdetect software, provided by Imagine Eyes. A wide-field cone density map was created using coordinates describing the relative positions of AO image frames that were montaged usingthe MosaicJ software.

RESULTS

Brief Case Summary

CASE 1: An 8-year-old female had pigmentary changes in both fundi. Her visual acuity was 20/32 in each eye. Anterior segment examination revealed no lens opacities and intraocular pressures (IOPs) were 16 mmHg in both eyes. Examination findings remained stable over a 10-month period. Axial lengths were 21.34 and 21.32 mm in the right and left eye, respectively. Her mother gave a history of vaccination and had active rubella infection at 10 weeks’ gestation in South Africa and the patient has not been tested for hearing impairment but has had no developmental delay in language or speech.

CASE 2: A 33-year-old female was referred for assessment of ‘retinitis pigmentosa’ despite a lack of history of nyctalopia. She had a "salt-and-pepper" appearance of both fundi with perifoveal pseudo-vitelliform lesions in the left eye. Her visual acuity was 20/20 in each eye and IOPs were 19 mmHg in both eyes. She had no cataracts. Axial lengths were 23.73 and 23.38 mm in the right and left eye, respectively. Full-field and pattern electroretinography were within normal limits. Her clinical status remained stable over 12 months. Her Australian-born mother had been diagnosed with active rubella infection at 6 weeks’ gestation in Australia and doe not recall prior vaccination The patient was diagnosed with mild sensorineural hearing loss at the age of 6 years old.

CASE 3: A 42-year-old female was found to have pigmentary changes in her retina during a ophthalmic examination as part of a cutaneous melanoma clinical trial in which she received dabrafenib and trametinib. She was unable to tolerate these and withdrew from the study within a few months of enrolment. She had no history of nyctalopia. Visual acuities were 20/18 and 20/35 in the right and left eye, respectively. IOPs were 18 mmHg in both eyes and she had a bilateral “salt-and-pepper” retinal appearance. She had no lenticular changes. Axial lengths were 22.79 and 22.61 mm in the right and left eye, respectively. There was no change in her clinical status over a 24-month period. Her Australian-born mother developed rubella infection at 3 months’ gestation in Australia and the patient has a history of sensorineural hearing impairment. The vaccination status of the patient’s mother is unknown.

CASE 4: A 50-year-old male was referred for diabetic retinopathy screening. He had congenital sensorineural deafness. Visual acuities were 20/25 and IOPs were 16 mmHg in both eyes. In addition to a bilateral “salt-and-pepper” appearance of the retina, he also had temporal perifoveal lipid exudates and occasional microaneurysms from mild non-proliferative diabetic retinopathy. There was no macular edema and the lens was clear in each eye. The retinopathy did not change over a 6-month period. It was note possible to ascertain or verify a history of maternal rubella infection during gestation because his parents were deceased. He was born at a time before rubella vaccination was available to his mother.

Multimodal Imaging

Wide-field pseudo-color SLO imaging(Optos) demonstratedmottled pigmentary changes extending from the macula to the temporal equator in all 4 patients.Green light induced AF imaging (Optos) showed speckled hypo-AF in the macular region extending to the temporal equator to variable degrees (Figure 1). Stippling of AF signal wasmore prominent on near-infraredAF imaging (Spectralis) than on blue -wave-AF imaging in 3 of the 4 patients(Figure 2) or green AF imaging (Figure 3). There were several hyperautofluorescent (hyper-AF) lesions scattered throughout the posterior pole on blue light AF imaging (Spectralis) and some of these correlated with hyper-reflective sub-retinal pseudo-vitelliform lesions or drusen-like lesions on SD-OCT (Figure 2). Case 4 also had hyper-reflective lesions within the outer retina representing hard exudate from resolving diabetic macular edema. The hypo-AF lesions seen with near-infrared excitation were hyper-reflective on near- infrared reflectance (NIR) imaging (Figure 3). . It was not possible to capture near-infrared AF images in one patient (case 3)case 3 despite using the combined dual wavelength angiography mode provided by the Spectralis device. There were several hyperautofluorescent (hyper-AF) lesions scattered throughout the posterior pole on blue light AF imaging (Spectralis) and some of these correlated with irregularities of the interdigitation zone or hyper-reflective sub-retinal pseudo-vitelliform lesion on SD-OCT (Figure 2). Case 4 also had hyper-reflective lesions within the outer retina representing hard exudate from resolving diabetic macular edema (Figure 2).

In general, the ellipsoid zone was intact except in the region of pseudo-vitelliform lesions,but there were patches of the retina affected by attenuation of the interdigitation zone as seen in the dense raster SD-OCT slices.Tofurther investigate the relationship ofthe ellipsoid and interdigitation zone attenuation reflectivity profiles with hypo-AF patches seen on near-infrared AF, en face intensity maps of the ellipsoid and interdigitation zones were created (Figure 43). In general, the ellipsoid zone was intact except in the regions of pseudo-vitelliform lesions, but there were separate patches of the retina affected by attenuation of the interdigitation zone as seen in the dense raster SD-OCT slices. Co-registration between theseen face intensity mapsand near-infrared AF imagesshowed no obvious correlation between regions of ellipsoid or interdigitation zone attenuation and loss of near-infrared AF signal.

Cases 2and 3 also had microperimetry and AO imagingin the foveal region (Figure 54). Co-registration between the cone mosaic montages with near-infrared AF images showed that the wave-guiding cone outer tips werestill visiblein regions of hypo-AF. Retinal sensitivity was preserved over regions of hypo-AF (Figure 54). Case 2 also had normal light rise in electro-oculography and amplitudes and latencies on full-field electroretinography (Figure 65). Cases 2 and 3 had follow-up imaging performed over the course of 12 and 24 months,respectively, and neither patient there was nodemonstrated any deterioration in visual acuity, retinal sensitivity on microperimetry or progression of the retinal lesions on examination and AF imaging (Figure 7).

DISCUSSION

Fundus AF is a useful diagnostic tool in the differential diagnosis of various types of pigmentary retinopathies such asinherited retinal dystrophy, immune or infectious chorioretintis and drug toxicities.5Signal generated from the retina in this imaging modality is based on the autofluorescentproperties of molecules such as fluorophores within degraded photoreceptor outer segments in the subretinal space or lipofuscin and melaninwithin the RPE and melanin within melanocytes of the choroidchoroidal melanocytes.8,9 Although fundus The pattern of fluorescenceAF in congenital rubella on fundus autofluorescence has previously been characterizedbeen described, this was limited to the use of blue light excitation in the macular region and there was no correlation with photoreceptor cell function and structure, but this was using only blue light excitation.Using multimodal imaging, we demonstrated speckledhypo-AFwith blue, green and near-infrared lightexcitation in patients with rubella retinopathy. On wide-field imaging, we documented extension of These these AF abnormalities extended to the equator temporally. There was a significant difference in the pattern of AF , they are not congruent between excitation wavelengths and they the lesions on AF do did not correlate with ellipsoid or interdigitation zone attenuation on SD-OCT. Focal and punctate hyper-AF correlated with pseudo-vitelliform lesions and hyper-reflective depositson over the RPE.Microperimetry demonstrated normal retinal sensitivity and AO imaging showed wave-guiding coneswere detected over regions areas of hypo-AF.

The 4 cases were referred with varied diagnoses ranging fromidiopathic retinal pigmentation(case 1), presumed retinitis pigmentosa (case 2), cutaneous melanoma chemotherapy toxicity screening (case 3) and diabetic retinopathy screening (case 4). The first 3 cases had confirmed history of maternal gestational rubella consistent with vertical transmission but none of them had cataract or a history of cardiac disease and only twohad sensorineural hearing impairment. Retinal toxicity from inhibitors of the serine/threonine-protein kinase B-RAF (dabrafenib) and mitogen-activated protein kinase (trametinib) have been described10,11. While the former is known to cause uveitis,the latter has been associated with central serous retinopathy, retinal vein occlusion and pigment epithelial detachment; none of these features were present in Case 3. Case 4 has retinal features of a “salt-and-pepper” retinopathy and congenital hearing impairment without confirmed history of maternal gestational rubella infection. Although congenital syphilis is a possible differential diagnosis, he had no other features of the Hutchinson triad such as interstitial keratitis, saddle nose, mulberry molars and peg-shaped incisors. We did not have the opportunity to exclude this diagnosis with serology. None of these patients had nyctalopia or loss of the photoreceptor layer on SD-OCT to suggest that they may have a rod cone dystrophy associated with Usher syndrome.The individual referred with presumed retinitis pigmentosa also had normal full-field electrophysiology, making the diagnosis of a rod cone dystrophy very unlikely.

Goldberg et al. illustrated unique short wavelength (blue light)AF features in 4 cases of presumed rubella retinopathyaged 14 to 40 years old4. Perifoveal stippledAF was noted and one case had subfoveal choroidal neovascularisation. We broadened the description of AF imaging features in rubella retinopathy by illustrating extension of speckled AF signal loss to the temporal equatorial region of the retinaon green light excitation using the Optos wide field camera. Furthermore, near-infrared AF (possible successfully acquirerd in 3 of the 4 cases) showed more numerous and largerhypo-AF lesions in the macular region compared to bluelight AF. Discordance in the distribution of hypo-AFlesions between blue and near-infrared light excitationhas been reported in age-related macular degeneration7,12, inherited retinal dystrophy13–16, central serous retinopathy17,18 and inflammatory retinopathy19.Most of these studies also demonstrated a relationship between hypo-AF and the loss of ellipsoid zone integrity and reduced function on microperimetry or visual field examinations12–14,16,20. Duncker et al. postulated that preservation of blue light autofluorescence in regions of reduced near-infrared AF signal in Stargardt macular dystrophy is due to photoreceptor debris over a region wheredead RPE has been lost.14.In retinitis pigmentosa, the increased foveal near-infrared AF is thought to be due to increased melanolysosomeproduction by RPE located underneathcones with abnormal but preserved cone outer segments and intact synaptic pedicles9,21–23. Conversely, the reduced near-infrared AF outside the foveal island of vision may be due toloss of melanin and melanolysosomes in RPE adjacent to residual cone cell bodies without outer segments or synaptic pedicles15,23,24. If this hypothesis also holds true for rubella retinopathy, we would expect to see reduced retinal sensitivity on microperimetry, lack of wave-guiding coneson AO imaging and loss of the ellipsoid and interdigitation zoneson SD-OCT in the regions of hypo-AF on near-infrared excitation.