A white woman in her 60s presented with a juxtafoveal choroidal nevus (arrowheads in part C). The adaptive optics image of the boundary of the lesion depicts 250 × 320-µm samples analysis for the en face image (parts E1 and E5), cell density (parts E2 and E6), intercellular spacing (parts E3 and E7), and Voronoi polygon (parts E4 and E8).
A white woman in her 50s presented with an extrafoveal choroidal nevus (arrowheads in part C). The adaptive optics image of the boundary of the lesion depicts 250 × 320-µm samples analysis for the Voronoi polygon (parts D1 and D6), intercellular spacing (parts D2 and D5), cell density (parts D3 and D8), and en face image (parts D4 and D7).
An African American man in his 50s presented with a peripapillary choroidal nevus (arrowheads in part B). En face adaptive optics imaging highlights (yellow- and orange-dashed square) the 250 × 320-µm samples analysis in parts B1 (top left, en face image; top right, cell density; bottom left, intercellular spacing; and bottom right, Voronoi polygon) and B2 (top right, en face image; top left, cell density; bottom right, intercellular spacing; and bottom left, Voronoi polygon).
eTable. Available Clinical Data and Adaptive Optics Analysis on Patients
Customize your JAMA Network experience by selecting one or more topics from the list below.
Rodrigues MW, Correa ZM, Say EA, et al. Photoreceptor Arrangement Changes Secondary to Choroidal Nevus. JAMA Ophthalmol. 2016;134(11):1315–1319. doi:10.1001/jamaophthalmol.2016.3633
Although mostly asymptomatic, patients with choroidal nevi carry a moderate risk for malignant transformation and visual loss. A novel noninvasive imaging assessment could change the current clinical evaluation of choroidal nevi.
Three patients with a recent diagnosis of choroidal nevi underwent a novel adaptive optical assessment that detected potential photoreceptor abnormalities in the retina overlying the choroidal nevi.
Conclusions and Relevance
Adaptive optics imaging may provide high-resolution en face images of retinal structural changes in the photoreceptor mosaic overlying the choroidal nevi. Cone attenuation may be an important component of structural damage in choroidal nevi and may correlate and possibly predict functional visual loss.
Choroidal nevi are relatively common among white individuals.1 Malignant transformation occurs in only 1 of 8845 individuals annually.2 Current knowledge suggests these benign intraocular tumors threaten vision only if located under the fovea.3
Clinical evaluation of choroidal nevi has included baseline ophthalmic examination, ultrasonography, fundus photography, occasional optical coherence tomography (OCT), and follow-up every 6 to 12 months.3 The following 2 noninvasive tools have been incorporated for such evaluation: enhanced-depth spectral domain OCT (EDI SD-OCT) and fundus autofluorescence.4 Current fundus autofluorescence techniques provide limited information on choroidal tumors owing to faint or barely detectable autofluorescence.5 Conversely, a known factor associated with activity in melanocytic choroidal tumors, subretinal fluid,3 is detected by EDI SD-OCT, although it is occasionally overlooked clinically and ultrasonographically.6
This study takes an alternative approach with noninvasive en face imaging acquired by an adaptive optics (AO) system to analyze the retina overlying the choroidal nevi. Thus, we report novel findings on photoreceptor density and arrangement overlying choroidal nevi and consider their effect on retinal function.
This study adheres to the Declaration of Helsinki7 and was approved by the ethics and research committee of the University of São Paulo, Ribeirão Preto School of Medicine, Ribeirão Preto, Brazil. Informed written consent was obtained from all patients before testing.
All patients underwent a comprehensive ophthalmic examination consisting of best-corrected visual acuity, dilated ophthalmoscopic examination, fundus photography (CR-1 Mark II; Canon), near-infrared reflectance scanning laser ophthalmoscopy (Spectralis; Heidelberg Engineering), and EDI SD-OCT (Spectralis; Heidelberg Engineering). Two patients also underwent microperimetry using the expert examination strategy from MAIA (Macular Integrity Assessment; CenterVue SpA). We used a predefined grid of 37 points in 10° of macular coverage (stimulus intensity range, 0-36 dB). Threshold sensitivities at each predefined point were calculated using a staircase 4-2 strategy.
We used an AO device (RTX-1; Imagine Eyes) to evaluate all eyes. This commercial prototype obtains high-resolution imaging of cones using a noncoherent flood-illuminated design with an 850-nm central illumination wavelength and a 4° × 4° imaging field of view (1.2 × 1.2 mm of the retina) with a 600-mm focusing range. Images were processed using manufacturer-provided software (AO Detect, version 0.1; Imagine Eyes). We analyzed the spatial distribution of the point coordinates in terms of intercellular spacing, local cell density, and the number of nearest-neighbor cells.
All patients had 0.00 logMAR visual acuity (Snellen, 20/20) and normal anterior segments in both eyes. Results of the ophthalmoscopic examination of the fellow eyes were unremarkable. The involved eye revealed a melanocytic choroidal nevus in all patients (Figures 1A, 2A, and 3A). Patients 1 and 3 presented with macular lesions measuring 4.0 mm in horizontal basal diameter and 4.0 and 2.5 mm, respectively, in vertical basal diameter. Patient 2 had extramacular lesions measuring 2.0 × 2.0 mm in vertical and horizontal basal diameter. The posterior nevus edge in patient 3 touched the disc, which is considered a risk factor for growth.8
All nevi were hyperreflective using near-infrared reflectance imaging (Figures 1B, 2B, and 3B). The nevus appeared as a highly reflective band with shadowing of deeper structures by EDI SD-OCT (black-dashed squares in Figures 1C, 2C, and 3B) in each patient. The AO imaging (orange-dashed squares in Figures 1E, 2D, and 3B [insert 2]) demonstrated an irregular cone mosaic that became discontinuous with enlarged hyperreflective and hyporeflective polygonal shapes overlying the nevus. The unaffected retina revealed cone photoreceptors as continuous bright hyperreflective dots (yellow-dashed squares in Figures 1E, 2D, and 3B [insert 1]).
Representative 250 × 320-µm sampling windows (yellow- and orange-dashed squares) from uninvolved and nevus areas were matched to spatial distribution and analyzed for intercellular spacing range value of 15 to 20 µm, local cell density (range, 20 000-30 000 cones/mm2), and number of neighboring cells (approximately 6) using the AO software images to create color scale graphic representation. Measurements obtained in patients 1, 2, and 3 revealed mean cone densities of 9055, 3967, and 11 597 cones/mm2, respectively, in uninvolved retina. Over the nevus area, mean cone densities were 457, 650, and 786 cones/mm2, respectively. Proportions of 6-sided Voronoi polygons9 were 35.7%, 26%, and 42.8% in the uninvolved retina, respectively. Nevus area proportions were 22.7%, 25.5%, and 25.4%, respectively. Intercellular spacing was 11.25, 16.58, and 10.03 µm, respectively, in uninvolved retina and 32.11, 39.01, and 32.28 µm overlying the nevi, respectively. Clinical data and AO analysis for all 3 patients (with conversion from micrometer to arcminute spacing) are available in the eTable in the Supplement.
Microperimetry in patients 1 and 3 showed an abnormal macular integrity index of 98.9 and 99.9, respectively (0-40 indicates normal; 41-60, suspicious; and 61-100, abnormal).10 The mean threshold was 25.0 and 21.0 dB, respectively (>25 to 36 dB indicates normal; 24-25 dB, suspicious; and <24 dB, abnormal).10 A sensitivity map of patients 1 (Figure 1D) and 3 (Figure 3C) suggested reduced retinal function near the choroidal nevus.
This study shows disturbance of the cone mosaic pattern over the choroidal nevus. However, recognition of the limitations of AO imaging and its correlation with photoreceptor integrity is important. First, we cannot assume that cone mosaic disturbance occurs solely in overlying nevi. For example, white patches were verified on AO images from patients 1 and 2 and may represent increased retinal pigment epithelial scatter seen in drusen.11 Second, the disturbed cone mosaic on AO images does not necessarily represent photoreceptor cell death. Disturbance of photoreceptor alignment over the lesion may lead to disruption of cone wave-guide properties and lack of reflectance, leading to the dark patches verified in all 3 patients. Furthermore, the cone mosaic represents integrity of part of the cone-cell inner and outer segments.12 The ellipsoid portion of the inner segment and distal portion of the outer segments represented as hyperreflective lines on EDI SD-OCT have been shown to be restored or recovered in some retinal diseases, such as acute retinal pigment epithelitis and acute zonal occult outer retinopathy, and after ocriplasmin use.13 Essentially, part of the cone’s inner and outer segments may be renewed, and disturbances of cone mosaic detected by AO imaging do not necessarily indicate definitive cone loss.
The EDI SD-OCT findings showed an intact ellipsoid band over the nevi, whereas AO imaging demonstrated cone mosaic irregularities, such as a hyperreflectivity or a hyporeflectivity round or oval speck pattern. Investigators14 have hypothesized that rods could actually contribute to the intact signal of the ellipsoid zone when cone population is scant. Moreover, the light focus waist of OCT is 28 μm in diameter, which covers approximately 25 times more photoreceptors than the light focus of 5.6 μm in diameter for AO.
Cone mosaic changes on AO imaging could result from an insidious insult attributed to melanocyte accumulation that potentially disturbs the choriocapillaris flow and consequently disrupts photoreceptor outer segments. Late secondary outer retinal changes, such as retinal pigment epithelial and intraretinal cystic changes, photoreceptor loss, and retinal pigment epithelial detachment, can then be detected by EDI SD-OCT.15 In fact, researchers16 showed that similar melanocyte-driven mechanical effect could cause scleral bowing.
Because AO detected photoreceptor changes, we used microperimetry to check for retinal function. The sensitivity maps of patients 1 (Figure 1D) and 3 (Figure 3C) outlined areas of hyposensitivity (orange- and yellow-dashed squares) near the nevus. However, areas of hyposensitivity are present outside the nevus area, posing the question whether cone mosaic changes and reduced retinal sensitivity on microperimetry are related.
Noteworthy disturbance in photoreceptor arrangement was detected in patients with asymptomatic choroidal nevi and normal EDI SD-OCT findings. These findings suggest that choroidal nevi might cause changes in the overlying retina that reduce visual sensitivity despite their lack of malignant or metastatic potential. Future studies with more cases are necessary to confirm our preliminary findings.
Corresponding Author: Rodrigo Jorge, MD, PhD, Department of Ophthalmology, Ribeirão Preto School of Medicine, University of São Paulo, 3900 Bandeirantes Ave, Ribeirão Preto, SP 14049-900, Brazil (email@example.com).
Accepted for Publication: August 7, 2016.
Published Online: October 6, 2016. doi:10.1001/jamaophthalmol.2016.3633
Author Contributions: Dr Rodrigues had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Correa, Say, Jorge.
Acquisition, analysis, or interpretation of data: Rodrigues, Borges, Siqueira, Cardillo, Jorge.
Drafting of the manuscript: Rodrigues, Borges, Jorge.
Critical revision of the manuscript for important intellectual content: Correa, Say, Siqueira, Cardillo, Jorge.
Statistical analysis: Rodrigues.
Obtained funding: Jorge.
Administrative, technical, or material support: Borges, Jorge.
Study supervision: Say, Siqueira, Cardillo, Jorge.
Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Correa reports serving on the Advisory Board of Castle Biosciences, LLC. No other disclosures were reported.
Funding/Support: This study was supported in part by the National Council for Scientific and Technological Development, the Mary Knight Asbury Endowed Chair of Ocular Oncology and Ophthalmic Pathology (Dr Correa), and Special Scholar Grant 404641/2012-1 from the Ribeirão Preto School of Medicine, University of São Paulo (Dr Jorge).
Role of the Funder/Sponsor: The funding source had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.