Swept-Source Optical Coherence Tomography Features of Choroidal Nevi JASMINE H. FRANCIS, CLAUDINE E. PANG, DAVID H. ABRAMSON, TATYANA MILMAN, ROBERT FOLBERG, SARAH MREJEN, AND K. BAILEY FREUND PURPOSE: To investigate the morphologic characteristics of choroidal nevi using swept-source optical coherence tomography and compare this with enhanced-depth optical coherence tomography. DESIGN: Retrospective observational case series. METHODS: One choroidal nevus each from 30 eyes of 30 patients was included and received imaging with swept-source OCT (SS-OCT) and enhanced-depth imaging OCT (EDI-OCT). For SS-OCT, a scan acquisition protocol was used involving 12 mm horizontal and vertical scans in the posterior fundus. The main outcome measures were morphologic features of choroidal nevi obtained with SS-OCT imaging. These features were compared to images obtained with EDI-OCT. A 2-tailed Fisher exact test was the statistical method used. RESULTS: SS-OCT allowed for an appreciation of intralesional details: Of the 30 nevi imaged, intralesional vessels were apparent in 30 (100%), intralesional cavities in 6 (20%), intralesional granularity in 14 (47%), abnormal choriocapillaris in 25 (83%), and abnormal choriocapillaris confined to the tumor apex in 17 (58%). Distended bordering vessels were identified in 22 nevi (73%) and were significantly associated with the presence of previous or persistent subretinal fluid. Intrinsic hyperreflectivity with hyporeflective shadowing was significantly (P [ .05) more apparent in 14 of 21 melanotic nevi (67%) compared with 2 of 9 amelanotic nevi (22%). Visualization of the complete nevus-scleral interface was significantly (P [ .02) more apparent in 7 of 9 amelanotic nevi (78%) compared with 6 of 21 melanotic nevi (29%), and was not significantly related to tumor thickness (measured by ultrasound) or to tumor configuration. Tumor diameter (but not tumor height) was statistically significantly associated with secondary retinal changes (P [ .05) and configuration (P [ .01). EDI-OCT was equivalent at determining secondary retinal changes (P [ .29), the presence of distended bordering vessels (P [ 1), visualization of the nevus-scleral interface
Accepted for publication Oct 7, 2014. From Memorial Sloan Kettering Cancer Center (J.H.F., D.H.A.); Weill Cornell Medical College, New York Presbyterian Hospital (J.H.F., D.H.A.); Vitreous Retina Macula Consultants of New York (C.E.P., S.M., K.B.F.); and New York Eye and Ear Infirmary (T.M.), New York, New York; and Oakland University William Beaumont School of Medicine (R.F.), Rochester, Minnesota. Inquiries to Jasmine H. Francis, Memorial Sloan Kettering Cancer Center, 1275 York Ave, New York, NY 10065; e-mail: [email protected] 0002-9394/$36.00 http://dx.doi.org/10.1016/j.ajo.2014.10.011
Ó
2015 BY
(P [ .6), and hyporeflective gradation at the nevusscleral interface (P [ .33). However, in melanotic lesions, SS-OCT was significantly superior at visualizing intralesional vessels (P [ .0002), intralesional granularity (P [ .0005), and abnormal choriocapillaris (P [ .0001). CONCLUSION: Imaging of choroidal nevi with SS-OCT enables visualization of intralesional details such as vessels (present in 100% of tumors imaged), cavities, and granularity. For melanotic lesions, SS-OCT is significantly better at depicting certain intralesional characteristics compared to EDI-OCT. Distended bordering vessels were recognized in over two thirds of the nevi imaged and were significantly associated with previous or persistent subretinal fluid. (Am J Ophthalmol 2015;159:169–176. Ó 2015 by Elsevier Inc. All rights reserved.)
T
HE FIRST REPORTS ON OPTICAL COHERENCE TOMOG-
raphy (OCT) imaging of choroidal nevi described the ‘‘limitations’’ of the technology.1 In 1998, Schaudig1 wrote that ‘‘the more important information of tumor histology is hidden in diffuse and weak subretinal backscattering.’’ Time-domain OCT (TD OCT) could provide information on secondary retinal changes and this became the subject of multiple reports from various groups.2–5 However, owing to shadowing and lack of signal penetration, TD OCT gave poor imaging details deeper to these retinal changes. Literature on the use of spectral-domain OCT (SD OCT) imaging of choroidal nevi is scarce. Sayanagi and associates6 showed that SD OCT allows for better identification of secondary retinal changes, but imaging of the tumor ‘‘was limited only to the anterior aspect.’’ In 2011, the use of enhanced-depth imaging OCT (EDIOCT), a modification of SD OCT image acquisition, for evaluating choroidal nevi was described. Using EDI-OCT, Torres and associates7 further expounded upon the influence of pigmentation on image quality and characteristics, and demonstrated that amelanotic nevi had less shadow and a homogenous, medium reflectivity.7 In the same year, Shah and associates used EDI-OCT to capture images of the anterior tumor sufficient for them to describe that 94% of choroidal nevi had choriocapillaris thinning.8 However, even with the EDI-OCT technique, pigmented lesions continued to demonstrate shadowing, which compromised visualization beyond the anterior tumor or inner choroid.
ELSEVIER INC. ALL
RIGHTS RESERVED.
169
Swept-source OCT (SS-OCT) has recently gained momentum, and aspects of this technology would presumably offer improved images of the choroid and pigmented lesions therein. It uses a wavelength-tunable laser and dualbalanced photodetector offering higher imaging speed. Its adaptability to a longer wavelength (on the order of 1 mm) improves the imaging range to include the choroid and penetrates melanin to a greater extent.9 Besides investigations of vitreous and retinal architecture,10–13 SS-OCT has revealed novel findings in choroidal and scleral characteristics in normal eyes and in those with pathologic myopia,14–16 in addition to choroidal features of diseases such as chronic central serous chorioretinopathy and reticular pseudodrusen.17,18 However, there are still opportunities to expand the utilization of this novel imaging technology in the analysis of additional choroidal pathology. The aim of this present study is to use this long-wavelength SS-OCT system to image the morphologic characteristics of choroidal nevi.
METHODS THE STUDY ADHERED TO THE TENETS OF THE DECLARATION
of Helsinki, complied with the Health Insurance Portability and Accountability Act, and was approved by the Institutional Review Board of Memorial Sloan Kettering Cancer Center. Informed consent was obtained from all patients prior to SS-OCT examination. The study included 30 choroidal nevi in 30 eyes of 30 patients (7 male, 23 female) recruited from the institution of Memorial Sloan Kettering Cancer Center, New York between September 1, 2012 and May 31, 2014. Patients with macular choroidal nevi were consecutively included. Of the 40 patients that were approached for inclusion in the study, 10 patients declined participation because of a scheduling conflict. Patients received imaging with the SS-OCT device at the clinical practice of Vitreous Retina Macula Consultants of New York and EDI-OCT imaging at either Vitreous Retina Macula Consultants of New York or Memorial Sloan Kettering Cancer Center. EXAMINATION: All enrolled patients received an ophthalmologic examination complete with bestcorrected visual acuity, ultrasonography, intraocular pressure, dilated fundus examination, and fundus photography. Patients were imaged with the DRI OCT-1 Atlantis 3D SSOCT device (Topcon Medical Systems, Oakland, New Jersey, USA), centered at 1050 nm for enhanced choroidal penetration. The axial resolution was approximately 6 mm with a scan speed of 100 000 A-scans per second. A scan width of 12 mm was used and 5-line cross-scan patterns with 0.25 mm spacing were chosen in both the horizontal and vertical direction. Thirty-two averaging scans were obtained per line. EDI-OCT images were obtained
170
with the Heidelberg Spectralis HRAþOCT (Heidelberg Engineering, Heidelberg, Germany). A scan of 9 mm was used and 13-line cross-scan patterns were chosen in the both the horizontal and vertical direction. DATA COLLECTION: Demographic data were collected on each patient, including sex and age. Clinical features included diameter of the nevus (mm), height (mm) and configuration (plateau, dome, mushroom) of the nevus determined by standard ultrasonography, color (melanotic vs predominantly amelanotic), and retinal changes (retinal pigment epithelial [RPE] changes, drusen, orange pigment, halo, subretinal fluid). SS-OCT features were recorded and included presence of secondary retinal/RPE changes: drusen, RPE alterations, cystoid macular edema, ellipsoid alterations, visualization of Bruch membrane, and orange pigment correlated with fundus autofluorescence. Previous or persistent subretinal fluid was established through historical OCT images and/or the presence of characteristic areas of hyper-autofluorescence detected with autofluorescence imaging. The main outcome measure included tumor characteristics determined by SS-OCT, which included intrinsic hyperreflectivity with hyporeflective shadowing (optical shadowing), visualization of the nevus-scleral interface, presence of hyporeflective line at the nevusscleral interface, presence of intralesional vessels, cavities or granularity, presence of distended bordering vessels and association with subretinal fluid, and presence of abnormal choriocapillaris and whether confined to the tumor apex. The configuration of the tumor was recorded into 3 groups: plateau (no distention of the retina), dome (distention of retina only), and almond (distention of the retina and sclera). Secondary retinal changes and tumor characteristics obtained by the SS-OCT technology were compared to EDIOCT images, which were available for all patients. STATISTICAL
analysis was ANALYSIS: Statistical performed using the 2-tailed Fisher exact test on GraphPad Software, Inc (La Jolla, California, USA). Tumor characteristics were compared between melanotic and amelanotic tumors, and configuration and secondary retinal changes were compared between tumors less than or equal to 3 mm in diameter and those larger than 3 mm; and between tumors less than or equal to 1 mm in height and those larger than 1 mm. Visualization of the nevus-scleral interface was compared between different tumor configurations and tumors less than or equal to 1.4 mm in height and those greater than 1.4 mm in height (1.4 mm was the median height and was therefore selected as the cutoff). The presence of previous or persistent subretinal fluid was compared between nevi with and without distended bordering vessels. Retinal changes and tumor characteristics were compared between EDI-OCT and SS-OCT. A P value less than or equal to .05 was considered statistically significant.
AMERICAN JOURNAL OF OPHTHALMOLOGY
JANUARY 2015
TABLE 1. Swept-Source Optical Coherence Tomography of 30 Eyes With Choroidal Nevi: Comparison of Morphologic Characteristics Between Melanotic and Amelanotic Tumors Swept-Source Imaging OCT Characteristics
All Nevi (n ¼ 30)
Melanotic (n ¼ 21)
Amelanotic (n ¼ 9)
P Valuea
Intrinsic hyperreflectivity with hyporeflective shadow Visualization of nevus-scleral interface Hyporeflective gradation of nevus-scleral interface Intratumoral vessels Intratumoral cavities Intratumoral granularity Distended bordering vessels Abnormal choriocapillaris Abnormal choriocapillaris at apex only
53% (16) 43% (13) 27% (8) 100% (30) 20% (6) 47% (14) 73% (22) 83% (25) 58% (17)
67% (14) 29% (6) 19% (4) 100% (21) 14% (3) 48% (10) 76% (16) 76% (16) 48% (10)
22% (2) 78% (7) 44% (4) 100% (9) 33% (3) 44% (4) 67% (6) 100% (9) 78% (7)
.05 .02 .2 1 .33 1 1 .29 .23
OCT ¼ optical coherence tomography. a P value compares melanotic and amelanotic.
RESULTS TWENTY-ONE MELANOTIC AND 9 AMELANOTIC CHOROIDAL
nevi were studied with a median height of 1.4 mm (range 0.8–3.2 mm) and diameter of 5 mm (range 1.5–11 mm). The following secondary retinal changes were detected by both SS-OCT and EDI-OCT in the 30 eyes studied: drusen in 22 (73%), RPE alterations in 26 (87%), cystoid retinal edema in 4 (13%), and ellipsoid alterations in 22 (73%). Of 30 eyes, an intact Bruch membrane was visible by SS-OCT in 17 (57%), compared to 12 (40%) by EDIOCT. With SS-OCT it was possible to detect intralesional details of both melanotic and amelanotic nevi, including intralesional vessels (appearing equidistant and equal in size), cavities, granularity, distended bordering vessels, and hyporeflective gradation at the nevus-scleral interface, as well as other details such as the presence of intrinsic hyperreflectivity with hyporeflective shadowing, visualization of the nevus-scleral interface, and status of the choriocapillaris (Table 1). Distended bordering vessels were significantly associated with the presence of previous or persistent subretinal fluid (P ¼ .03): of 22 tumors with distended bordering vessels, 10 had SRF (45%), compared to 0 of 8 tumors without distended bordering vessels (0%). Three distinct anatomic nevi configurations were identified (Figures 1 and 2). These configurations were termed ‘‘plateau’’ (distention into the sclera only), seen in 6 of 30 eyes (20%); ‘‘dome’’ (distention into the retina only), seen in 8 of 30 eyes (27%); and ‘‘almond’’ (distention into both sclera and retina), seen in 10 of 30 eyes (33%). Even in instances where the nevus-scleral interface was not fully visualized along the entire length of the tumor, there was adequate visualization at the sides to determine presence of scleral distention. Six of the 30 nevi (20%) were either dome or almond in configuration, but shadowing VOL. 159, NO. 1
prevented exact determination, owing to inadequate visualization of the nevus-scleral interface. Ultrasonography could identify the tumor configuration in 27 of 30 tumors (90%). The remaining 3 tumors were read as plateau on ultrasound, but were determined to be dome shape on SS-OCT. Visualization of the nevus-scleral interface was statistically more likely in amelanotic vs melanotic nevi (Table 1). However, visualization was not significantly related to tumor thickness when comparing those less than or equal to 1.4 mm and those greater than 1.4 mm: P ¼ .16 for all tumors, P ¼ .18 for melanotic tumors, and P ¼ .44 for amelanotic tumors. Visualization of the nevus-scleral interface was not significantly associated with tumor configuration (P ¼ .66). Tumor diameter was significantly associated with secondary retinal changes (P ¼ .05) and configuration (P ¼ .01). Of 8 tumors less than 3 mm in diameter, 3 had retinal changes (38%), compared with 21 of 22 tumors greater than 3 mm in diameter (95%). Furthermore, of 22 tumors greater than 3 mm in diameter, 21 had a configuration that distended into the retina (dome or almond) (95%), compared with 4 of 8 tumors less than 3 mm in diameter (50%). Tumor height (less than vs greater than 1 mm) was not associated with secondary retinal changes (P ¼ 1) or a configuration that distended into the retina (P ¼ .3). Of 10 tumors less than 1 mm in height, 9 had retinal changes (85%), comparable to 17 of 20 tumors greater than 3 mm in diameter (95%). Furthermore, of 20 tumors greater than 1 mm in height, 18 had a configuration that distended into the retina (dome or almond) (90%), comparable to 7 of 10 tumors less than 3 mm in diameter (70%). Visualization of secondary retinal changes and tumor characteristics by EDI-OCT and comparison with SSOCT is depicted in Table 2. EDI-OCT was equivalent at determining secondary retinal changes, the presence of
SWEPT-SOURCE OCT OF CHOROIDAL NEVI
171
FIGURE 1. Representative swept-source optical coherence tomography (SS-OCT) image of a choroidal nevus with corresponding diagrammatic illustration. (Top left) Color fundus photograph of amelanotic and melanotic nevus. (Top right) SS-OCT image of choroidal nevus with an almond configuration demonstrating intralesional vessels, distended bordering vessels, and gradation of reflectivity from inner to outer portion of nevus, perhaps representing distribution of different cell types. Note relative gradient of hyporeflectivity as the nevus transitions to sclera, possibly a result of the densely packed spindle cells. (Bottom) Diagrammatic schema illustrating histopathologic characteristics of a representative choroidal nevus that are found on SS-OCT.
distended bordering vessels, visualization of the nevusscleral interface, and its hyporeflective gradation. However, SS-OCT was significantly superior at visualizing intralesional vessels, intralesional granularity, and abnormal choriocapillaris in melanotic lesions. Furthermore, the morphology of melanotic nevi could be visualized significantly better with SS-OCT compared with EDI-OCT (P ¼ .0001). Of 21 melanotic tumors, the morphology could be determined in 21 (100%) with SS-OCT, compared to 16 (53%) with EDI-OCT. In amelanotic lesions, visualization of the morphology was equivalent with SS-OCT and EDIOCT (P ¼ 1).
DISCUSSION AS THE TECHNOLOGY OF OPTICAL COHERENCE TOMOGRAPHY
has progressed, so too has our imaging of choroidal nevi and our understanding of their morphologic characteristics. TD-OCT imaging was predominantly limited to the secondary retinal changes associated with choroidal nevi1–5; and although SD OCT allowed for a better appreciation of retinal changes, the images it provided were restricted to the anterior portion of the tumor.6 EDI-OCT began to give us an understanding of the inner portion of the tumor and choriocapillaris.7,8 The swept-source device used in this study has given us an opportunity to explore deeper aspects of nevi and reveal novel OCT observations. 172
Many of our OCT observations confirm previously established morphologic features that are observed on histopathology of choroidal nevi. In 1966, Naumann and associates19 described 3 distention patterns for choroidal nevi: those that ‘‘push’’ into the retina, those into the sclera, and those into both. This study confirmed the 3 configuration patterns (respectively designated by us as dome, plateau, and almond). The almond shape was particularly discernible on SS-OCT owing to visualization of the nevus-scleral interface, which is not always identified on other imaging modalities such as ultrasound. In this study, SS-OCT and EDI-OCT were equivalent at determining the morphology of amelanotic nevi, but SS-OCT was significantly better at visualizing the morphology of melanotic nevi (P ¼ .0001). Although visualization of the nevus-scleral interface was equivalent by both modalities, SS-OCT gave adequate visualization at the sides, allowing for depiction of scleral distention. Furthermore, in this cohort, EDI-OCT and SS-OCT were comparable at distinguishing secondary retinal changes, the presence of distended bordering vessels, and hyporeflective gradation at the nevus-scleral interface. However, in melanotic lesions, SS-OCT was significantly better at showing finer intralesional details such as vessels, granularity, and abnormalities of the choriocapillaris. As suggested by previous studies that used EDI-OCT, identification of the nevus-scleral interface is related to inherent melanin. One study remarks upon the ease of interface visualization in amelanotic nevi compared to
AMERICAN JOURNAL OF OPHTHALMOLOGY
JANUARY 2015
FIGURE 2. Configuration styles of representative choroidal nevi. (Top left) Color fundus photograph of an amelanotic choroidal nevus. (Top right) Corresponding SS-OCT demonstrating a dome configuration and visualization of the nevus-scleral interface. (Bottom left) Color fundus photograph of a melanotic nevus. (Bottom right) Corresponding SS-OCT demonstrating a boat configuration and visualization of the nevus-scleral interface. Note intralesional vessels that are equidistant and equal in size.
TABLE 2. Enhanced-Depth Imaging Optical Coherence Tomography of 30 Eyes With Choroidal Nevi: Comparison of Morphologic Characteristics Between Melanotic and Amelanotic Tumors, and to Swept-Source Optical Coherence Tomography Enhanced-Depth Imaging a
OCT Characteristics
All Nevi (n ¼ 30)
P Value
Melanotic (n ¼ 21)
P Valuea
Amelanotic (n ¼ 9)
P Valuea
Intrinsic hyperreflectivity with hyporeflective shadow Visualization of nevus-scleral interface Hyporeflective gradation of nevus-scleral interface Intratumoral vessels Intratumoral cavities Intratumoral granularity Distended bordering vessels Abnormal choriocapillaris Abnormal choriocapillaris at apex only
70% (21) 33% (10) 13% (4) 53% (16) 7% (2) 3% (1) 73% (22) 7% (2) 3% (1)
.29 .6 .33 .0001 .25 .0002 1 .0001 .0001
76% (16) 19% (4) 10% (2) 48% (10) 5% (1) 0% (0) 76% (16) 0% (0) 0% (0)
.73 .72 .66 .0002 .61 .0005 1 .0001 .0005
56% (5) 67% (6) 22% (2) 67% (6) 11% (1) 11% (1) 67% (6) 22% (2) 11% (1)
.34 1 .62 .21 .58 .29 1 .002 .02
OCT ¼ optical coherence tomography. P value compares enhanced-depth imaging and swept-source OCT.
a
the shadowed choroidal images associated with intensely pigmented nevi.8 In another study using EDI-OCT, a single melanotic nevus out of 9 had an identifiable ‘‘inner sclera.’’7 In our present study, 78% of the amelanotic nevi had a visible nevus-scleral interface. However, over a quarter of the melanotic lesions, which ranged in height from 0.8 to 1.8 mm, also had an identifiable interface, demonstrating that pigmented nevi can be imaged in their entirety with SS-OCT. Nonetheless, 71% of melanotic nevi did not have an identifiable nevus-scleral interface, indicating that even with this improved SS-OCT technology, the interface cannot be determined in the majority of pigmented nevi. In their study of various choroidal tumors, Torres and associates7 determined that the ‘‘inner sclera’’ was only visible in those lesions that were less than VOL. 159, NO. 1
0.9 mm in height. In our series, tumor thickness or configuration had no significant influence on whether the nevusscleral interface was identified. In their 1994 paper, Rummelt and associates20 illustrated the histopathology of choroidal nevi and the intralesional choroidal vessels that are apparent. Owing to the enhanced imaging available with SS-OCT, these intralesional vessels were identified in all choroidal nevi imaged in this study. As established by histopathology assessments, the hallmark of a benign choroidal nevus is the manner in which the intralesional vessels mimic those of the normal choroid: they appear equidistant and equal in size. This pattern was recognized in all the nevi imaged in this series and is demonstrated in Figure 1. In addition, 20% of the nevi had intralesional cavities, which appear larger than and
SWEPT-SOURCE OCT OF CHOROIDAL NEVI
173
FIGURE 3. Intralesional cavities and previous subretinal fluid of representative chorodial nevi. (Top left) Color fundus photograph of choroidal nevus. (Top middle) Corresponding indocyananine green angiography demonstrating no intralesional hyperfluorescence. (Top right) Swept-source optical coherence tomography (SS-OCT) demonstrating intralesional cavities, which appear larger and distinct from the intralesional vessels. (Bottom left) Color fundus photograph of choroidal nevus with retinal pigment epithelial changes in area of previous subretinal fluid. (Bottom right) SS-OCT demonstrating a hyporeflective band at the chorioscleral junction in the location of the previous subretinal fluid, suggestive of a mild suprachoroidal effusion.
FIGURE 4. Characteristics of representative choroidal nevi. (Top left) Color fundus photograph of a melanotic choroidal nevus. (Top right) Corresponding swept-source optical coherence tomography (SS-OCT) demonstrating lack of visualization of the nevus-scleral interface owing to hyperreflectivity and optical shadowing (hyporeflectivity). (Bottom left) SS-OCT of a choroidal nevus distending into the retina with compression of choriocapillaris and secondary retinal changes (pigment epithelial detachment, disturbance of ellipsoid segment) at the apex and relative sparing at the sides of the tumor. (Bottom right) SS-OCT of choroidal nevus with distended bordering vessels appearing larger than adjacent normal choroid.
distinct from the intralesional vessels. In 1 of these cases, there was a lack of indocyanine green (ICG) enhancement corresponding to these cavities (Figure 3, Top). This suggests the cavities are not vascular in nature (ie, not vascular ‘‘lakes’’), and alternatively could be distinct clumps of cells with differing (and homogeneous) reflectivity or fluid-filled cysts inaccessible to the vasculature. Another recurring feature was the distended vessels at the border of the tumor, which were found in more than two thirds of the cases (Figure 4). Distended bordering 174
vessels are found on ICG angiography of choroidal nevi,21 and this is consistent with our SS-OCT findings. This is also a common histopathologic finding in choroidal nevi and we may speculate as to the reason for vessel distention. Perhaps the mass effect of the nevus results in congestion of blood flow, the lack of fenestration of the medium and large choroidal vessels further impedes flow, and the outflow obstruction results in vessel distention. This mechanism may relate to the accumulation of subretinal fluid, which we found to be significantly more likely in nevi
AMERICAN JOURNAL OF OPHTHALMOLOGY
JANUARY 2015
with distended bordering vessels. In fact, for this reason, distended bordering vessels may act as a clinical prognosticator for visual disturbance from subretinal fluid, or future studies may determine these vessels to be a significant clinical feature associated with growth and, possibly, malignant transformation. Secondary retinal changes are detected commonly with choroidal nevi and were described extensively with the use of TD OCT and EDI-OCT.1–5 Based on their histopathologic findings, Naumann and associates19 originally suggested that compression of the choriocapillaris may contribute to the formation of drusen. Shah and associates8 later used their EDI-OCT observations to extend this to other types of retinal alterations. In our series, choriocapillaris compression was more apparent at the apex of the tumor and less often at the sides, and this coincided with the pattern of secondary retinal changes (Figure 4). It is intriguing that secondary retinal changes were significantly less common in tumors with a diameter less than 3 mm (P ¼ .05). Furthermore, these smaller tumors were significantly less likely to assume a configuration that distended into the retina (P ¼ .01), which presumably results in less apical compression of the choriocapillaris. In 1965, Zimmerman22 wrote that choroidal nevi had a distinct distribution of cells. The plump polyhedral cells were on the inner portion of the tumor; these transitioned to spindle cells, which became more densely packed toward the scleral interface.22 Since each cell type may have a unique reflectivity, it could explain the gradation of reflectivity that we observed from the inner to outer portion of the tumor (Figure 1). Furthermore, the increased hyporeflectivity as the nevus transitions to the sclera could be
attributable to the more densely packed spindle cells. On the other hand, it could also be related to pigment shadowing, although this pattern is apparent in 44% of amelanotic nevi compared to 19% of melanotic lesions. Intralesional granularity was apparent in almost half the nevi of this series, and appeared distinct from the normal choroid and retina. This imaging characteristic is previously unreported and future studies will determine whether it is unique to SS-OCT. It is unclear what the granularity signifies, but we posit that it may represent an aspect of the cellular milieu that comprises the nevus. Recently, Yiu and associates23 described a hyporeflective band at the choroid-scleral junction, which was more common in hyperopic eyes. They speculated this line to be the suprachoroidal layer with a mild effusion. We noted an identical hyporeflective band in 2 eyes. This raises the question as to whether accumulation of a mild suprachoroidal effusion occurs for the same reason in both hyperopic eyes and those with choroidal nevi, namely, Gass’24 theory of osmotic fluid retention. Alternatively, compression of the veins by the nevus may raise hydrostatic pressure and result in a transudation of fluid. In summary, the long-wave SS-OCT penetrates melanin better and at a greater depth in the choroid compared to other OCT techniques, offering detailed images of choroidal nevi. SS-OCT allows for in vivo identification of features that are typically only observed on histopathology, thereby lending itself as a useful diagnostic tool. This is particularly relevant to melanotic nevi, where SS-OCT is significantly better at depicting intralesional characteristics compared to EDI-OCT. Longer-term longitudinal studies may reveal if any of these features are predictive of growth and/or malignant transformation.
ALL AUTHORS HAVE COMPLETED AND SUBMITTED THE ICMJE FORM FOR DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST and the following was reported: K.B.F.: consultant for Heidelberg Engineering (Heidelberg, Germany), Genentech (San Francisco, California, USA), Regeneron (Tarrytown, New York, USA), Bayer HealthCare (Leverkusen, Germany). The other authors reported no disclosures. This study was supported by The Fund for Ophthalmic Knowledge (New York, New York). The authors attest to their independence in reporting the study data and interpretation of the data. Contributions of authors: design and conduct of the study (J.H.F., D.H.A.); collection, management, analysis, and interpretation of the data (J.H.F., C.E.P., D.H.A., T.M., R.F., S.M., K.B.F.); and preparation, review, or approval of the manuscript (J.H.F., C.E.P., D.H.A., T.M., R.F., S.M., K.B.F.).
REFERENCES 1. Schaudig U, Hassenstein A, Bernd A, Walter A, Richard G. Limitations of imaging choroidal tumors in vivo by optical coherence tomography. Graefes Arch Clin Exp Ophthalmol 1998;236(8):588–592. 2. Muscat S, Srinivasan S, Sampat V, Kemp E, Parks S, Keating D. Optical coherence tomography in the diagnosis of subclinical serous detachment of the macula secondary to a choroidal nevus. Ophthalmic Surg Lasers 2001;32(6):474–476. 3. Espinoza G, Rosenblatt B, Harbour JW. Optical coherence tomography in the evaluation of retinal changes associated with suspicious choroidal melanocytic tumors. Am J Ophthalmol 2004;137(1):90–95.
VOL. 159, NO. 1
4. Muscat S, Parks S, Kemp E, Keating D. Secondary retinal changes associated with choroidal naevi and melanomas documented by optical coherence tomography. Br J Ophthalmol 2004;88(1):120–124. 5. Shields CL, Mashayekhi A, Materin MA, et al. Optical coherence tomography of choroidal nevus in 120 patients. Retina 2005;25(3):243–252. 6. Sayanagi K, Pelayes DE, Kaiser PK, Singh AD. 3D Spectral domain optical coherence tomography findings in choroidal tumors. Eur J Ophthalmol 2011;21(3): 271–275. 7. Torres VLL, Brugnoni N, Kaiser PK, Singh AD. Optical coherence tomography enhanced depth imaging of choroidal tumors. Am J Ophthalmol 2011;151(4):586–593.e2.
SWEPT-SOURCE OCT OF CHOROIDAL NEVI
175
8. Shah SU, Kaliki S, Shields CL, Ferenczy SR, Harmon SA, Shields JA. Enhanced depth imaging optical coherence tomography of choroidal nevus in 104 cases. Ophthalmology 2012;119(5):1066–1072. 9. Mrejen S, Spaide RF. Optical coherence tomography: imaging of the choroid and beyond. Surv Ophthalmol 2013;58(5): 387–429. 10. Schaal KB, Pang CE, Pozzoni MC, Engelbert M. The premacular bursa’s shape revealed in vivo by swept-source optical coherence tomography. Ophthalmology 2014;121(5):1020–1028. 11. Tanawade RG, Muqit MM, McLeod D, Stanga PE. Sweptsource optical coherence tomography imaging in conservative and surgical management of premacular haemorrhages showing inflammatory response. Clin Experiment Ophthalmol 2014; http://dx.doi.org/10.1111/ceo.12368. 12. Ohsugi H, Ikuno Y, Oshima K, Yamauchi T, Tabuchi H. Morphological characteristics of macular complications of a dome-shaped macula determined by swept-source optical coherence tomography. Am J Ophthalmol 2014;158(1): 162–170. 13. Li D, Kishi S, Itakura H, Ikeda F, Akiyama H. Posterior precortical vitreous pockets and connecting channels in children on swept-source optical coherence tomography. Invest Ophthalmol Vis Sci 2014;55(4):2412–2416. 14. Lim LS, Cheung G, Lee SY. Comparison of spectral domain and swept-source optical coherence tomography in pathological myopia. Eye (Lond) 2014;28(4):488–491. 15. Copete S, Flores-Moreno I, Montero JA, Duker JS, RuizMoreno JM. Direct comparison of spectral-domain and swept-source OCT in the measurement of choroidal thickness in normal eyes. Br J Ophthalmol 2014;98(3):334–338. 16. Adhi M, Liu JJ, Qavi AH, et al. Choroidal analysis in healthy eyes using swept source optical coherence tomography
176
17.
18.
19.
20.
21.
22. 23.
24.
compared to spectral domain optical coherence tomography. Am J Ophthalmol 2014;157(6):1272–1281. Ueda-Arakawa N, Ooto S, Ellabban AA, et al. Macular choroidal thickness and volume of eyes with reticular pseudodrusen using swept-source optical coherence tomography. Am J Ophthalmol 2014;157(5):994–1004. Ferrara D, Mohler KJ, Waheed N, et al. En face enhanceddepth swept-source optical coherence tomography features of chronic central serous chorioretinopathy. Ophthalmology 2014;121(3):719–726. Naumann G, Yanoff M, Zimmerman LE. Histogenesis of malignant melanomas of the uvea. I. Histopathologic characteristics of nevi of the choroid and ciliary body. Arch Ophthalmol 1966;76(6):784–796. Rummelt V, Folberg R, Rummelt C, et al. Microcirculation architecture of melanocytic nevi and malignant melanomas of the ciliary body and choroid. A comparative histopathologic and ultrastructural study. Ophthalmology 1994;101(4): 718–727. Shiraki K, Moriwaki M, Yanagihara N, Kohno T, Miki T. Indocyanine green angiograms of choroidal nevi. Comparison between confocal and nonconfocal scanning laser ophthalmoscope and fundus video camera. Jpn J Ophthalmol 2001; 45(4):368–374. Zimmerman LE. Melanocytes, melanocytic nevi, and melanocytomas. Invest Ophthalmol 1965;4:11–41. Yiu G, Pecen P, Sarin N, et al. Characterization of the choroid-scleral junction and suprachoroidal layer in healthy individuals on enhanced-depth imaging optical coherence tomography. JAMA Ophthalmol 2014;132(2):174–181. Gass JD. Uveal effusion syndrome: a new hypothesis concerning pathogenesis and technique of surgical treatment. Trans Am Ophthalmol Soc 1983;81:246–260.
AMERICAN JOURNAL OF OPHTHALMOLOGY
JANUARY 2015
Biosketch Jasmine H. Francis graduated magna cum laude from Columbia University. She subsequently completed medical school at New York University, pursued an internship at Memorial Sloan Kettering Cancer Center and an ophthalmology residency at the New York Eye and Ear Infirmary. She returned to Memorial Sloan Kettering as a fellow of Ophthalmic Oncology and has since join the faculty. Her research areas within ophthalmic oncology consist predominantly of intraocular tumors such as retinoblastoma and choroidal melanoma.
VOL. 159, NO. 1
SWEPT-SOURCE OCT OF CHOROIDAL NEVI
176.e1
Accepted for publication Oct 7, 2014. From Memorial Sloan Kettering Cancer Center (J.H.F., D.H.A.); Weill Cornell Medical College, New York Presbyterian Hospital (J.H.F., D.H.A.); Vitreous Retina Macula Consultants of New York (C.E.P., S.M., K.B.F.); and New York Eye and Ear Infirmary (T.M.), New York, New York; and Oakland University William Beaumont School of Medicine (R.F.), Rochester, Minnesota. Inquiries to Jasmine H. Francis, Memorial Sloan Kettering Cancer Center, 1275 York Ave, New York, NY 10065; e-mail: [email protected] 0002-9394/$36.00 http://dx.doi.org/10.1016/j.ajo.2014.10.011
Ó
2015 BY
(P [ .6), and hyporeflective gradation at the nevusscleral interface (P [ .33). However, in melanotic lesions, SS-OCT was significantly superior at visualizing intralesional vessels (P [ .0002), intralesional granularity (P [ .0005), and abnormal choriocapillaris (P [ .0001). CONCLUSION: Imaging of choroidal nevi with SS-OCT enables visualization of intralesional details such as vessels (present in 100% of tumors imaged), cavities, and granularity. For melanotic lesions, SS-OCT is significantly better at depicting certain intralesional characteristics compared to EDI-OCT. Distended bordering vessels were recognized in over two thirds of the nevi imaged and were significantly associated with previous or persistent subretinal fluid. (Am J Ophthalmol 2015;159:169–176. Ó 2015 by Elsevier Inc. All rights reserved.)
T
HE FIRST REPORTS ON OPTICAL COHERENCE TOMOG-
raphy (OCT) imaging of choroidal nevi described the ‘‘limitations’’ of the technology.1 In 1998, Schaudig1 wrote that ‘‘the more important information of tumor histology is hidden in diffuse and weak subretinal backscattering.’’ Time-domain OCT (TD OCT) could provide information on secondary retinal changes and this became the subject of multiple reports from various groups.2–5 However, owing to shadowing and lack of signal penetration, TD OCT gave poor imaging details deeper to these retinal changes. Literature on the use of spectral-domain OCT (SD OCT) imaging of choroidal nevi is scarce. Sayanagi and associates6 showed that SD OCT allows for better identification of secondary retinal changes, but imaging of the tumor ‘‘was limited only to the anterior aspect.’’ In 2011, the use of enhanced-depth imaging OCT (EDIOCT), a modification of SD OCT image acquisition, for evaluating choroidal nevi was described. Using EDI-OCT, Torres and associates7 further expounded upon the influence of pigmentation on image quality and characteristics, and demonstrated that amelanotic nevi had less shadow and a homogenous, medium reflectivity.7 In the same year, Shah and associates used EDI-OCT to capture images of the anterior tumor sufficient for them to describe that 94% of choroidal nevi had choriocapillaris thinning.8 However, even with the EDI-OCT technique, pigmented lesions continued to demonstrate shadowing, which compromised visualization beyond the anterior tumor or inner choroid.
ELSEVIER INC. ALL
RIGHTS RESERVED.
169
Swept-source OCT (SS-OCT) has recently gained momentum, and aspects of this technology would presumably offer improved images of the choroid and pigmented lesions therein. It uses a wavelength-tunable laser and dualbalanced photodetector offering higher imaging speed. Its adaptability to a longer wavelength (on the order of 1 mm) improves the imaging range to include the choroid and penetrates melanin to a greater extent.9 Besides investigations of vitreous and retinal architecture,10–13 SS-OCT has revealed novel findings in choroidal and scleral characteristics in normal eyes and in those with pathologic myopia,14–16 in addition to choroidal features of diseases such as chronic central serous chorioretinopathy and reticular pseudodrusen.17,18 However, there are still opportunities to expand the utilization of this novel imaging technology in the analysis of additional choroidal pathology. The aim of this present study is to use this long-wavelength SS-OCT system to image the morphologic characteristics of choroidal nevi.
METHODS THE STUDY ADHERED TO THE TENETS OF THE DECLARATION
of Helsinki, complied with the Health Insurance Portability and Accountability Act, and was approved by the Institutional Review Board of Memorial Sloan Kettering Cancer Center. Informed consent was obtained from all patients prior to SS-OCT examination. The study included 30 choroidal nevi in 30 eyes of 30 patients (7 male, 23 female) recruited from the institution of Memorial Sloan Kettering Cancer Center, New York between September 1, 2012 and May 31, 2014. Patients with macular choroidal nevi were consecutively included. Of the 40 patients that were approached for inclusion in the study, 10 patients declined participation because of a scheduling conflict. Patients received imaging with the SS-OCT device at the clinical practice of Vitreous Retina Macula Consultants of New York and EDI-OCT imaging at either Vitreous Retina Macula Consultants of New York or Memorial Sloan Kettering Cancer Center. EXAMINATION: All enrolled patients received an ophthalmologic examination complete with bestcorrected visual acuity, ultrasonography, intraocular pressure, dilated fundus examination, and fundus photography. Patients were imaged with the DRI OCT-1 Atlantis 3D SSOCT device (Topcon Medical Systems, Oakland, New Jersey, USA), centered at 1050 nm for enhanced choroidal penetration. The axial resolution was approximately 6 mm with a scan speed of 100 000 A-scans per second. A scan width of 12 mm was used and 5-line cross-scan patterns with 0.25 mm spacing were chosen in both the horizontal and vertical direction. Thirty-two averaging scans were obtained per line. EDI-OCT images were obtained
170
with the Heidelberg Spectralis HRAþOCT (Heidelberg Engineering, Heidelberg, Germany). A scan of 9 mm was used and 13-line cross-scan patterns were chosen in the both the horizontal and vertical direction. DATA COLLECTION: Demographic data were collected on each patient, including sex and age. Clinical features included diameter of the nevus (mm), height (mm) and configuration (plateau, dome, mushroom) of the nevus determined by standard ultrasonography, color (melanotic vs predominantly amelanotic), and retinal changes (retinal pigment epithelial [RPE] changes, drusen, orange pigment, halo, subretinal fluid). SS-OCT features were recorded and included presence of secondary retinal/RPE changes: drusen, RPE alterations, cystoid macular edema, ellipsoid alterations, visualization of Bruch membrane, and orange pigment correlated with fundus autofluorescence. Previous or persistent subretinal fluid was established through historical OCT images and/or the presence of characteristic areas of hyper-autofluorescence detected with autofluorescence imaging. The main outcome measure included tumor characteristics determined by SS-OCT, which included intrinsic hyperreflectivity with hyporeflective shadowing (optical shadowing), visualization of the nevus-scleral interface, presence of hyporeflective line at the nevusscleral interface, presence of intralesional vessels, cavities or granularity, presence of distended bordering vessels and association with subretinal fluid, and presence of abnormal choriocapillaris and whether confined to the tumor apex. The configuration of the tumor was recorded into 3 groups: plateau (no distention of the retina), dome (distention of retina only), and almond (distention of the retina and sclera). Secondary retinal changes and tumor characteristics obtained by the SS-OCT technology were compared to EDIOCT images, which were available for all patients. STATISTICAL
analysis was ANALYSIS: Statistical performed using the 2-tailed Fisher exact test on GraphPad Software, Inc (La Jolla, California, USA). Tumor characteristics were compared between melanotic and amelanotic tumors, and configuration and secondary retinal changes were compared between tumors less than or equal to 3 mm in diameter and those larger than 3 mm; and between tumors less than or equal to 1 mm in height and those larger than 1 mm. Visualization of the nevus-scleral interface was compared between different tumor configurations and tumors less than or equal to 1.4 mm in height and those greater than 1.4 mm in height (1.4 mm was the median height and was therefore selected as the cutoff). The presence of previous or persistent subretinal fluid was compared between nevi with and without distended bordering vessels. Retinal changes and tumor characteristics were compared between EDI-OCT and SS-OCT. A P value less than or equal to .05 was considered statistically significant.
AMERICAN JOURNAL OF OPHTHALMOLOGY
JANUARY 2015
TABLE 1. Swept-Source Optical Coherence Tomography of 30 Eyes With Choroidal Nevi: Comparison of Morphologic Characteristics Between Melanotic and Amelanotic Tumors Swept-Source Imaging OCT Characteristics
All Nevi (n ¼ 30)
Melanotic (n ¼ 21)
Amelanotic (n ¼ 9)
P Valuea
Intrinsic hyperreflectivity with hyporeflective shadow Visualization of nevus-scleral interface Hyporeflective gradation of nevus-scleral interface Intratumoral vessels Intratumoral cavities Intratumoral granularity Distended bordering vessels Abnormal choriocapillaris Abnormal choriocapillaris at apex only
53% (16) 43% (13) 27% (8) 100% (30) 20% (6) 47% (14) 73% (22) 83% (25) 58% (17)
67% (14) 29% (6) 19% (4) 100% (21) 14% (3) 48% (10) 76% (16) 76% (16) 48% (10)
22% (2) 78% (7) 44% (4) 100% (9) 33% (3) 44% (4) 67% (6) 100% (9) 78% (7)
.05 .02 .2 1 .33 1 1 .29 .23
OCT ¼ optical coherence tomography. a P value compares melanotic and amelanotic.
RESULTS TWENTY-ONE MELANOTIC AND 9 AMELANOTIC CHOROIDAL
nevi were studied with a median height of 1.4 mm (range 0.8–3.2 mm) and diameter of 5 mm (range 1.5–11 mm). The following secondary retinal changes were detected by both SS-OCT and EDI-OCT in the 30 eyes studied: drusen in 22 (73%), RPE alterations in 26 (87%), cystoid retinal edema in 4 (13%), and ellipsoid alterations in 22 (73%). Of 30 eyes, an intact Bruch membrane was visible by SS-OCT in 17 (57%), compared to 12 (40%) by EDIOCT. With SS-OCT it was possible to detect intralesional details of both melanotic and amelanotic nevi, including intralesional vessels (appearing equidistant and equal in size), cavities, granularity, distended bordering vessels, and hyporeflective gradation at the nevus-scleral interface, as well as other details such as the presence of intrinsic hyperreflectivity with hyporeflective shadowing, visualization of the nevus-scleral interface, and status of the choriocapillaris (Table 1). Distended bordering vessels were significantly associated with the presence of previous or persistent subretinal fluid (P ¼ .03): of 22 tumors with distended bordering vessels, 10 had SRF (45%), compared to 0 of 8 tumors without distended bordering vessels (0%). Three distinct anatomic nevi configurations were identified (Figures 1 and 2). These configurations were termed ‘‘plateau’’ (distention into the sclera only), seen in 6 of 30 eyes (20%); ‘‘dome’’ (distention into the retina only), seen in 8 of 30 eyes (27%); and ‘‘almond’’ (distention into both sclera and retina), seen in 10 of 30 eyes (33%). Even in instances where the nevus-scleral interface was not fully visualized along the entire length of the tumor, there was adequate visualization at the sides to determine presence of scleral distention. Six of the 30 nevi (20%) were either dome or almond in configuration, but shadowing VOL. 159, NO. 1
prevented exact determination, owing to inadequate visualization of the nevus-scleral interface. Ultrasonography could identify the tumor configuration in 27 of 30 tumors (90%). The remaining 3 tumors were read as plateau on ultrasound, but were determined to be dome shape on SS-OCT. Visualization of the nevus-scleral interface was statistically more likely in amelanotic vs melanotic nevi (Table 1). However, visualization was not significantly related to tumor thickness when comparing those less than or equal to 1.4 mm and those greater than 1.4 mm: P ¼ .16 for all tumors, P ¼ .18 for melanotic tumors, and P ¼ .44 for amelanotic tumors. Visualization of the nevus-scleral interface was not significantly associated with tumor configuration (P ¼ .66). Tumor diameter was significantly associated with secondary retinal changes (P ¼ .05) and configuration (P ¼ .01). Of 8 tumors less than 3 mm in diameter, 3 had retinal changes (38%), compared with 21 of 22 tumors greater than 3 mm in diameter (95%). Furthermore, of 22 tumors greater than 3 mm in diameter, 21 had a configuration that distended into the retina (dome or almond) (95%), compared with 4 of 8 tumors less than 3 mm in diameter (50%). Tumor height (less than vs greater than 1 mm) was not associated with secondary retinal changes (P ¼ 1) or a configuration that distended into the retina (P ¼ .3). Of 10 tumors less than 1 mm in height, 9 had retinal changes (85%), comparable to 17 of 20 tumors greater than 3 mm in diameter (95%). Furthermore, of 20 tumors greater than 1 mm in height, 18 had a configuration that distended into the retina (dome or almond) (90%), comparable to 7 of 10 tumors less than 3 mm in diameter (70%). Visualization of secondary retinal changes and tumor characteristics by EDI-OCT and comparison with SSOCT is depicted in Table 2. EDI-OCT was equivalent at determining secondary retinal changes, the presence of
SWEPT-SOURCE OCT OF CHOROIDAL NEVI
171
FIGURE 1. Representative swept-source optical coherence tomography (SS-OCT) image of a choroidal nevus with corresponding diagrammatic illustration. (Top left) Color fundus photograph of amelanotic and melanotic nevus. (Top right) SS-OCT image of choroidal nevus with an almond configuration demonstrating intralesional vessels, distended bordering vessels, and gradation of reflectivity from inner to outer portion of nevus, perhaps representing distribution of different cell types. Note relative gradient of hyporeflectivity as the nevus transitions to sclera, possibly a result of the densely packed spindle cells. (Bottom) Diagrammatic schema illustrating histopathologic characteristics of a representative choroidal nevus that are found on SS-OCT.
distended bordering vessels, visualization of the nevusscleral interface, and its hyporeflective gradation. However, SS-OCT was significantly superior at visualizing intralesional vessels, intralesional granularity, and abnormal choriocapillaris in melanotic lesions. Furthermore, the morphology of melanotic nevi could be visualized significantly better with SS-OCT compared with EDI-OCT (P ¼ .0001). Of 21 melanotic tumors, the morphology could be determined in 21 (100%) with SS-OCT, compared to 16 (53%) with EDI-OCT. In amelanotic lesions, visualization of the morphology was equivalent with SS-OCT and EDIOCT (P ¼ 1).
DISCUSSION AS THE TECHNOLOGY OF OPTICAL COHERENCE TOMOGRAPHY
has progressed, so too has our imaging of choroidal nevi and our understanding of their morphologic characteristics. TD-OCT imaging was predominantly limited to the secondary retinal changes associated with choroidal nevi1–5; and although SD OCT allowed for a better appreciation of retinal changes, the images it provided were restricted to the anterior portion of the tumor.6 EDI-OCT began to give us an understanding of the inner portion of the tumor and choriocapillaris.7,8 The swept-source device used in this study has given us an opportunity to explore deeper aspects of nevi and reveal novel OCT observations. 172
Many of our OCT observations confirm previously established morphologic features that are observed on histopathology of choroidal nevi. In 1966, Naumann and associates19 described 3 distention patterns for choroidal nevi: those that ‘‘push’’ into the retina, those into the sclera, and those into both. This study confirmed the 3 configuration patterns (respectively designated by us as dome, plateau, and almond). The almond shape was particularly discernible on SS-OCT owing to visualization of the nevus-scleral interface, which is not always identified on other imaging modalities such as ultrasound. In this study, SS-OCT and EDI-OCT were equivalent at determining the morphology of amelanotic nevi, but SS-OCT was significantly better at visualizing the morphology of melanotic nevi (P ¼ .0001). Although visualization of the nevus-scleral interface was equivalent by both modalities, SS-OCT gave adequate visualization at the sides, allowing for depiction of scleral distention. Furthermore, in this cohort, EDI-OCT and SS-OCT were comparable at distinguishing secondary retinal changes, the presence of distended bordering vessels, and hyporeflective gradation at the nevus-scleral interface. However, in melanotic lesions, SS-OCT was significantly better at showing finer intralesional details such as vessels, granularity, and abnormalities of the choriocapillaris. As suggested by previous studies that used EDI-OCT, identification of the nevus-scleral interface is related to inherent melanin. One study remarks upon the ease of interface visualization in amelanotic nevi compared to
AMERICAN JOURNAL OF OPHTHALMOLOGY
JANUARY 2015
FIGURE 2. Configuration styles of representative choroidal nevi. (Top left) Color fundus photograph of an amelanotic choroidal nevus. (Top right) Corresponding SS-OCT demonstrating a dome configuration and visualization of the nevus-scleral interface. (Bottom left) Color fundus photograph of a melanotic nevus. (Bottom right) Corresponding SS-OCT demonstrating a boat configuration and visualization of the nevus-scleral interface. Note intralesional vessels that are equidistant and equal in size.
TABLE 2. Enhanced-Depth Imaging Optical Coherence Tomography of 30 Eyes With Choroidal Nevi: Comparison of Morphologic Characteristics Between Melanotic and Amelanotic Tumors, and to Swept-Source Optical Coherence Tomography Enhanced-Depth Imaging a
OCT Characteristics
All Nevi (n ¼ 30)
P Value
Melanotic (n ¼ 21)
P Valuea
Amelanotic (n ¼ 9)
P Valuea
Intrinsic hyperreflectivity with hyporeflective shadow Visualization of nevus-scleral interface Hyporeflective gradation of nevus-scleral interface Intratumoral vessels Intratumoral cavities Intratumoral granularity Distended bordering vessels Abnormal choriocapillaris Abnormal choriocapillaris at apex only
70% (21) 33% (10) 13% (4) 53% (16) 7% (2) 3% (1) 73% (22) 7% (2) 3% (1)
.29 .6 .33 .0001 .25 .0002 1 .0001 .0001
76% (16) 19% (4) 10% (2) 48% (10) 5% (1) 0% (0) 76% (16) 0% (0) 0% (0)
.73 .72 .66 .0002 .61 .0005 1 .0001 .0005
56% (5) 67% (6) 22% (2) 67% (6) 11% (1) 11% (1) 67% (6) 22% (2) 11% (1)
.34 1 .62 .21 .58 .29 1 .002 .02
OCT ¼ optical coherence tomography. P value compares enhanced-depth imaging and swept-source OCT.
a
the shadowed choroidal images associated with intensely pigmented nevi.8 In another study using EDI-OCT, a single melanotic nevus out of 9 had an identifiable ‘‘inner sclera.’’7 In our present study, 78% of the amelanotic nevi had a visible nevus-scleral interface. However, over a quarter of the melanotic lesions, which ranged in height from 0.8 to 1.8 mm, also had an identifiable interface, demonstrating that pigmented nevi can be imaged in their entirety with SS-OCT. Nonetheless, 71% of melanotic nevi did not have an identifiable nevus-scleral interface, indicating that even with this improved SS-OCT technology, the interface cannot be determined in the majority of pigmented nevi. In their study of various choroidal tumors, Torres and associates7 determined that the ‘‘inner sclera’’ was only visible in those lesions that were less than VOL. 159, NO. 1
0.9 mm in height. In our series, tumor thickness or configuration had no significant influence on whether the nevusscleral interface was identified. In their 1994 paper, Rummelt and associates20 illustrated the histopathology of choroidal nevi and the intralesional choroidal vessels that are apparent. Owing to the enhanced imaging available with SS-OCT, these intralesional vessels were identified in all choroidal nevi imaged in this study. As established by histopathology assessments, the hallmark of a benign choroidal nevus is the manner in which the intralesional vessels mimic those of the normal choroid: they appear equidistant and equal in size. This pattern was recognized in all the nevi imaged in this series and is demonstrated in Figure 1. In addition, 20% of the nevi had intralesional cavities, which appear larger than and
SWEPT-SOURCE OCT OF CHOROIDAL NEVI
173
FIGURE 3. Intralesional cavities and previous subretinal fluid of representative chorodial nevi. (Top left) Color fundus photograph of choroidal nevus. (Top middle) Corresponding indocyananine green angiography demonstrating no intralesional hyperfluorescence. (Top right) Swept-source optical coherence tomography (SS-OCT) demonstrating intralesional cavities, which appear larger and distinct from the intralesional vessels. (Bottom left) Color fundus photograph of choroidal nevus with retinal pigment epithelial changes in area of previous subretinal fluid. (Bottom right) SS-OCT demonstrating a hyporeflective band at the chorioscleral junction in the location of the previous subretinal fluid, suggestive of a mild suprachoroidal effusion.
FIGURE 4. Characteristics of representative choroidal nevi. (Top left) Color fundus photograph of a melanotic choroidal nevus. (Top right) Corresponding swept-source optical coherence tomography (SS-OCT) demonstrating lack of visualization of the nevus-scleral interface owing to hyperreflectivity and optical shadowing (hyporeflectivity). (Bottom left) SS-OCT of a choroidal nevus distending into the retina with compression of choriocapillaris and secondary retinal changes (pigment epithelial detachment, disturbance of ellipsoid segment) at the apex and relative sparing at the sides of the tumor. (Bottom right) SS-OCT of choroidal nevus with distended bordering vessels appearing larger than adjacent normal choroid.
distinct from the intralesional vessels. In 1 of these cases, there was a lack of indocyanine green (ICG) enhancement corresponding to these cavities (Figure 3, Top). This suggests the cavities are not vascular in nature (ie, not vascular ‘‘lakes’’), and alternatively could be distinct clumps of cells with differing (and homogeneous) reflectivity or fluid-filled cysts inaccessible to the vasculature. Another recurring feature was the distended vessels at the border of the tumor, which were found in more than two thirds of the cases (Figure 4). Distended bordering 174
vessels are found on ICG angiography of choroidal nevi,21 and this is consistent with our SS-OCT findings. This is also a common histopathologic finding in choroidal nevi and we may speculate as to the reason for vessel distention. Perhaps the mass effect of the nevus results in congestion of blood flow, the lack of fenestration of the medium and large choroidal vessels further impedes flow, and the outflow obstruction results in vessel distention. This mechanism may relate to the accumulation of subretinal fluid, which we found to be significantly more likely in nevi
AMERICAN JOURNAL OF OPHTHALMOLOGY
JANUARY 2015
with distended bordering vessels. In fact, for this reason, distended bordering vessels may act as a clinical prognosticator for visual disturbance from subretinal fluid, or future studies may determine these vessels to be a significant clinical feature associated with growth and, possibly, malignant transformation. Secondary retinal changes are detected commonly with choroidal nevi and were described extensively with the use of TD OCT and EDI-OCT.1–5 Based on their histopathologic findings, Naumann and associates19 originally suggested that compression of the choriocapillaris may contribute to the formation of drusen. Shah and associates8 later used their EDI-OCT observations to extend this to other types of retinal alterations. In our series, choriocapillaris compression was more apparent at the apex of the tumor and less often at the sides, and this coincided with the pattern of secondary retinal changes (Figure 4). It is intriguing that secondary retinal changes were significantly less common in tumors with a diameter less than 3 mm (P ¼ .05). Furthermore, these smaller tumors were significantly less likely to assume a configuration that distended into the retina (P ¼ .01), which presumably results in less apical compression of the choriocapillaris. In 1965, Zimmerman22 wrote that choroidal nevi had a distinct distribution of cells. The plump polyhedral cells were on the inner portion of the tumor; these transitioned to spindle cells, which became more densely packed toward the scleral interface.22 Since each cell type may have a unique reflectivity, it could explain the gradation of reflectivity that we observed from the inner to outer portion of the tumor (Figure 1). Furthermore, the increased hyporeflectivity as the nevus transitions to the sclera could be
attributable to the more densely packed spindle cells. On the other hand, it could also be related to pigment shadowing, although this pattern is apparent in 44% of amelanotic nevi compared to 19% of melanotic lesions. Intralesional granularity was apparent in almost half the nevi of this series, and appeared distinct from the normal choroid and retina. This imaging characteristic is previously unreported and future studies will determine whether it is unique to SS-OCT. It is unclear what the granularity signifies, but we posit that it may represent an aspect of the cellular milieu that comprises the nevus. Recently, Yiu and associates23 described a hyporeflective band at the choroid-scleral junction, which was more common in hyperopic eyes. They speculated this line to be the suprachoroidal layer with a mild effusion. We noted an identical hyporeflective band in 2 eyes. This raises the question as to whether accumulation of a mild suprachoroidal effusion occurs for the same reason in both hyperopic eyes and those with choroidal nevi, namely, Gass’24 theory of osmotic fluid retention. Alternatively, compression of the veins by the nevus may raise hydrostatic pressure and result in a transudation of fluid. In summary, the long-wave SS-OCT penetrates melanin better and at a greater depth in the choroid compared to other OCT techniques, offering detailed images of choroidal nevi. SS-OCT allows for in vivo identification of features that are typically only observed on histopathology, thereby lending itself as a useful diagnostic tool. This is particularly relevant to melanotic nevi, where SS-OCT is significantly better at depicting intralesional characteristics compared to EDI-OCT. Longer-term longitudinal studies may reveal if any of these features are predictive of growth and/or malignant transformation.
ALL AUTHORS HAVE COMPLETED AND SUBMITTED THE ICMJE FORM FOR DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST and the following was reported: K.B.F.: consultant for Heidelberg Engineering (Heidelberg, Germany), Genentech (San Francisco, California, USA), Regeneron (Tarrytown, New York, USA), Bayer HealthCare (Leverkusen, Germany). The other authors reported no disclosures. This study was supported by The Fund for Ophthalmic Knowledge (New York, New York). The authors attest to their independence in reporting the study data and interpretation of the data. Contributions of authors: design and conduct of the study (J.H.F., D.H.A.); collection, management, analysis, and interpretation of the data (J.H.F., C.E.P., D.H.A., T.M., R.F., S.M., K.B.F.); and preparation, review, or approval of the manuscript (J.H.F., C.E.P., D.H.A., T.M., R.F., S.M., K.B.F.).
REFERENCES 1. Schaudig U, Hassenstein A, Bernd A, Walter A, Richard G. Limitations of imaging choroidal tumors in vivo by optical coherence tomography. Graefes Arch Clin Exp Ophthalmol 1998;236(8):588–592. 2. Muscat S, Srinivasan S, Sampat V, Kemp E, Parks S, Keating D. Optical coherence tomography in the diagnosis of subclinical serous detachment of the macula secondary to a choroidal nevus. Ophthalmic Surg Lasers 2001;32(6):474–476. 3. Espinoza G, Rosenblatt B, Harbour JW. Optical coherence tomography in the evaluation of retinal changes associated with suspicious choroidal melanocytic tumors. Am J Ophthalmol 2004;137(1):90–95.
VOL. 159, NO. 1
4. Muscat S, Parks S, Kemp E, Keating D. Secondary retinal changes associated with choroidal naevi and melanomas documented by optical coherence tomography. Br J Ophthalmol 2004;88(1):120–124. 5. Shields CL, Mashayekhi A, Materin MA, et al. Optical coherence tomography of choroidal nevus in 120 patients. Retina 2005;25(3):243–252. 6. Sayanagi K, Pelayes DE, Kaiser PK, Singh AD. 3D Spectral domain optical coherence tomography findings in choroidal tumors. Eur J Ophthalmol 2011;21(3): 271–275. 7. Torres VLL, Brugnoni N, Kaiser PK, Singh AD. Optical coherence tomography enhanced depth imaging of choroidal tumors. Am J Ophthalmol 2011;151(4):586–593.e2.
SWEPT-SOURCE OCT OF CHOROIDAL NEVI
175
8. Shah SU, Kaliki S, Shields CL, Ferenczy SR, Harmon SA, Shields JA. Enhanced depth imaging optical coherence tomography of choroidal nevus in 104 cases. Ophthalmology 2012;119(5):1066–1072. 9. Mrejen S, Spaide RF. Optical coherence tomography: imaging of the choroid and beyond. Surv Ophthalmol 2013;58(5): 387–429. 10. Schaal KB, Pang CE, Pozzoni MC, Engelbert M. The premacular bursa’s shape revealed in vivo by swept-source optical coherence tomography. Ophthalmology 2014;121(5):1020–1028. 11. Tanawade RG, Muqit MM, McLeod D, Stanga PE. Sweptsource optical coherence tomography imaging in conservative and surgical management of premacular haemorrhages showing inflammatory response. Clin Experiment Ophthalmol 2014; http://dx.doi.org/10.1111/ceo.12368. 12. Ohsugi H, Ikuno Y, Oshima K, Yamauchi T, Tabuchi H. Morphological characteristics of macular complications of a dome-shaped macula determined by swept-source optical coherence tomography. Am J Ophthalmol 2014;158(1): 162–170. 13. Li D, Kishi S, Itakura H, Ikeda F, Akiyama H. Posterior precortical vitreous pockets and connecting channels in children on swept-source optical coherence tomography. Invest Ophthalmol Vis Sci 2014;55(4):2412–2416. 14. Lim LS, Cheung G, Lee SY. Comparison of spectral domain and swept-source optical coherence tomography in pathological myopia. Eye (Lond) 2014;28(4):488–491. 15. Copete S, Flores-Moreno I, Montero JA, Duker JS, RuizMoreno JM. Direct comparison of spectral-domain and swept-source OCT in the measurement of choroidal thickness in normal eyes. Br J Ophthalmol 2014;98(3):334–338. 16. Adhi M, Liu JJ, Qavi AH, et al. Choroidal analysis in healthy eyes using swept source optical coherence tomography
176
17.
18.
19.
20.
21.
22. 23.
24.
compared to spectral domain optical coherence tomography. Am J Ophthalmol 2014;157(6):1272–1281. Ueda-Arakawa N, Ooto S, Ellabban AA, et al. Macular choroidal thickness and volume of eyes with reticular pseudodrusen using swept-source optical coherence tomography. Am J Ophthalmol 2014;157(5):994–1004. Ferrara D, Mohler KJ, Waheed N, et al. En face enhanceddepth swept-source optical coherence tomography features of chronic central serous chorioretinopathy. Ophthalmology 2014;121(3):719–726. Naumann G, Yanoff M, Zimmerman LE. Histogenesis of malignant melanomas of the uvea. I. Histopathologic characteristics of nevi of the choroid and ciliary body. Arch Ophthalmol 1966;76(6):784–796. Rummelt V, Folberg R, Rummelt C, et al. Microcirculation architecture of melanocytic nevi and malignant melanomas of the ciliary body and choroid. A comparative histopathologic and ultrastructural study. Ophthalmology 1994;101(4): 718–727. Shiraki K, Moriwaki M, Yanagihara N, Kohno T, Miki T. Indocyanine green angiograms of choroidal nevi. Comparison between confocal and nonconfocal scanning laser ophthalmoscope and fundus video camera. Jpn J Ophthalmol 2001; 45(4):368–374. Zimmerman LE. Melanocytes, melanocytic nevi, and melanocytomas. Invest Ophthalmol 1965;4:11–41. Yiu G, Pecen P, Sarin N, et al. Characterization of the choroid-scleral junction and suprachoroidal layer in healthy individuals on enhanced-depth imaging optical coherence tomography. JAMA Ophthalmol 2014;132(2):174–181. Gass JD. Uveal effusion syndrome: a new hypothesis concerning pathogenesis and technique of surgical treatment. Trans Am Ophthalmol Soc 1983;81:246–260.
AMERICAN JOURNAL OF OPHTHALMOLOGY
JANUARY 2015
Biosketch Jasmine H. Francis graduated magna cum laude from Columbia University. She subsequently completed medical school at New York University, pursued an internship at Memorial Sloan Kettering Cancer Center and an ophthalmology residency at the New York Eye and Ear Infirmary. She returned to Memorial Sloan Kettering as a fellow of Ophthalmic Oncology and has since join the faculty. Her research areas within ophthalmic oncology consist predominantly of intraocular tumors such as retinoblastoma and choroidal melanoma.
VOL. 159, NO. 1
SWEPT-SOURCE OCT OF CHOROIDAL NEVI
176.e1
