Article Text
PDF
Abstract
Background/aims To assess spectral domain-optical coherence tomography (SD-OCT) contribution to choroidal osteoma characterisation.
Methods A retrospective chart review of a series of patients diagnosed with choroidal osteoma, which included patient, clinical, ultrasonographic, photographic and SD-OCT imaging.
Results 11 patients were included in this series. Their mean age was 42.5 years (median=43.0; range, 14–73). Using statistical analysis, the mean basal diameters of tumours as derived from fundus photographs (5.2 mm) and ultrasound images (6.4 mm) were significantly different (paired t-test, p=0.03). Tumours were SD-OCT hyporeflective in two cases, isoreflective in seven cases and hyper-reflective in two cases. Intrinsic reflectivity of the tumour was inhomogeneous in four cases. The overlying choroid was compressed by the tumour in eight cases and the retina exhibited degenerative changes in five cases.
Conclusion This study revealed that SD-OCT provided deeper and higher resolution images of choroidal osteoma when compared with previous studies using time domain-OCT. These findings offer new insights into the pathophysiology and diagnosis of choroidal osteoma.
- Osteoma
- choroid
- spectral
- tomography
- imaging
Statistics from Altmetric.com
Introduction
Choroidal osteoma (CO) is a rare, benign osseous tumour that predominantly affects young females. COs are typically unilateral, well-demarcated, yellow-white, peripapillary tumours located in the inner third of the choroid.1 2 COs can change over time. They can both grow and decalcify. COs can affect the surrounding tissues and alter retinal function, with choroidal neovascularisation (CNV) and tumour decalcification being identified as main risk factors for CO-related vision loss.1 3
Time domain-optical coherence tomography (TD-OCT) studies have led to several descriptions of CO structure and that of the overlying retina.4–6 However, small series have been hampered by insufficient resolution provided by TD-OCT.4–6 Because of its higher scan acquisition rate and deeper penetration, spectral domain (SD)-OCT enables better definition of the structures located below the retinal pigment epithelium (RPE).7
Herein, we present a series of patients with CO examined with SD-OCT. This study reveals unique, previously unobserved characteristics that can be related to different stages of tumour evolution.
Materials and methods
A retrospective, non-comparative, observational case series study was performed. The clinical records of all patients with a diagnosis of CO for which SD-OCT was performed between January 2008 and December 2010 were reviewed. This study adhered to the tenets of the Declaration of Helsinki of 1975, as revised in 2000, the Health Insurance Portability and Accountability Act of 1996, and was approved by the Institutional Review Board of The New York Eye Cancer Center.
Entry criteria
The definition of a CO was derived from the existing literature.8 9 These characteristics included that the COs were located posterior to the equator, usually close to the optic disc or involving the macula. Tumour colour ranged from white to orange-red, and shape ranged from round-oval to bilobed with geographic, scalloped margins. Tumours were strongly hyperechoic on ultrasonography with evidence of posterior signal attenuation (orbital shadowing). None of the patients had demonstrable active CNV. Tumours presenting as calcifications located outside the vascular arcades occurring in patients over 60 years of age were considered sclerochoroidal calcifications and excluded from this study. Of note, most of these sclerochoroidal calcifications were multiple or bilateral.
Tumour location and measurement
Macula and fovea were defined as central circular areas of 1.5 mm and 350 μm in diameter, respectively.10 Tumour location was assessed on review of fundus photographs (TRC-50IX, Topcon Medical Systems, Oakland, New Jersey, USA). Measurements of two perpendicular basal diameters were taken (ImageNet 2000 2.55, Topcon Medical Systems). Tumour measurements (basal diameter, thickness) were also established in a masked fashion from B-scan images (A-2000; OPKO—OTI, Ophthalmic Technologies Inc., Miami, Florida, USA) by a different operator. To verify whether measurements of basal diameters were accurate, a paired T-test using SPSS 17.0 statistical software (SPSS Inc.) was performed to determine whether a difference existed between the two methods.
SD-OCT imaging
Two different OCT devices were used during this study: a Cirrus HD-OCT (Carl Zeiss Meditec, Dublin, California, USA, software version 5.1.0.96, optical source wavelength of 840 nm) and a Spectralis OCT (Heidelberg Engineering, Heidelberg, Germany, software versions 5.1 and 5.3, optical source wavelength of 870 nm). Examination was performed after pupil dilation. A raster horizontal and/or vertical raster protocol was used, depending on the tumour shape. It used up to 49 lines, depending on tumour size. Data obtained from review of the OCT images included choroidal changes (thinning), tumour reflectivity (hyporeflective, isoreflective, hyper-reflective) and retinal changes (eg, retinal atrophy, retinal oedema, presence of subretinal fluid, RPE alterations). Because of variable signal attenuation induced by changes in the RPE, tumour reflectivity was primarily assessed in locations surrounded by healthy RPE.
Results
Demographics
Patients' data are listed in table 1. Eleven patients were included in the study, 8 of whom were female. Mean age at the time of OCT was 42.5 years (median=43.0; range, 14–73). The left eye was involved in five cases, whereas the right eye was involved in six cases. Mean visual acuity was 20/60 (range, 20/16—counting fingers).
Left: Patient's data (F, female; M, male; CF, counting fingers)
Tumour location, size and ultrasonography
The macula was involved in six cases and the fovea in 4. Tumours were unilateral in all cases. Mean basal diameter as measured on the fundus photographs was 5.2 mm (range, 1.4–15.2). On ultrasound examination, the mean basal diameter was statistically larger (mean: 6.4 mm; range: 3.9–15.8; p=0.03). Mean thickness was 1.2 mm (range, 0.5–2.8).
Tumours were hyperechoic with significant posterior shadowing in all cases. In small lesions, shadowing was faint and limited to the central part of the tumour. In one case, no posterior shadowing could be demonstrated, due to complete tumour decalcification (figure 1).
(A) Choroidal osteoma with a history of laser treatment for choroidal neovascularisation showing decalcification changes. The green line indicates the section through the retina in which the optical coherence tomographic (OCT) scan was performed. (B) The tumour is highly reflective on ultrasound (arrowhead) with no evidence of posterior shadowing. (C) OCT imaging demonstrates an isoreflective tumour surrounded by atrophic retina (arrowhead). Choroidal vessels posterior to the tumour are visualised and appear unremarkable.
SD-OCT findings
Choroidal findings
The majority of tumours (n=6) had isoreflective stroma. Nevertheless, reflectivity ranged from hyporeflective (n=2) to hyper-reflective (n=3). Although most tumours exhibited a homogeneous internal reflectivity, intratumoral variations in reflectivity were noticed in four cases. Some of these variations appeared similar to those previously reported on the histopathological analysis (figure 2).2
(A) Large bilobed choroidal osteoma with areas of retinal pigment epithelium hyperplasia and metaplasia. (B) Posterior attenuation on ultrasound prevents accurate demarcation of the posterior edge of the tumour (arrowhead). (C) The tumour exhibits variations in intrinsic reflectivity (arrowheads) similar to intratumoral channels described by Gass et al. (D) Specimen photograph reproduced from the original article.2 Permission granted by the Archives of Ophthalmology.
In eight cases, tumours both extended posteriorly to the sclera and created a localised elevation of the retina over the tumour. In these cases, the choroid between the tumour and the RPE was thinner (figure 3). In the three remaining cases, tumours were flat enough to cause neither disturbance of the choroid nor elevation of the retina (figure 4).
(A, C) COs that demonstrate choroidal displacement and thinning consistent with compression. (B) Compression is evidenced by localised thinning of the choroid with paucity of vessels (arrowhead). SD-OCT imaging enables accurate measurement of the tumour when located below the choroid (white line). (D) CO compressive effects on the choriocapillaris are associated with atrophy of the overlying retinal pigment epithelium, breakdown of the outer blood-retinal barrier with accumulation of fluid within outer retinal layers (arrowhead). A window defect is created (arrow), leading to artefactual change in light transmission through the tumour.
(A) A fundus photograph reveals the amelanotic CO in the macula. It measures 3.6×4.1 mm from the fundus photograph. (B) Ultrasonography reveals a 0.9 mm thick tumour with posterior shadowing (arrowhead). (C) This tumour is hyporeflective with intrinsic hyper-reflective dots (arrowhead). The posterior edge of the tumour is visible allowing for more accurate tumour thickness measurement. The corresponding (white line) measurement on SD-OCT is 320 μm.
Retinal findings
In five cases, the tumour induced changes in the overlying retina. These changes were associated with either tumour decalcification (n=1) or choriocapillaris compression (by the tumour (n=4)). Changes ranged from alterations of the inner segment/outer segment photoreceptor junction (n=1), retinal oedema (n=1), presence of subretinal fluid (n=3), RPE hyperplasia (n=2), to atrophy of the retinal outer layers (n=4) (figure 3).
Discussion
This study suggests that SD-OCT offers better resolution of the internal structure of CO compared with TD-OCT and ultrasonography analysis. The former offers less penetration and the latter appears restricted by the highly ultrasound-reflective and deeply located nature of the tumour. SD-OCT analysis reveals previously obscured characteristic of CO. For example, in this study we found a wider range of intra-tumour reflectivity as well as changes produced by the compression of the choroid and choriocapillaris in the overlying retina. Such physiologic processes revealed by OCT may be used to differentiate between osteomas and alternative amelanotic choroidal tumors. For example, COs may be confused with non-pigmented choroidal nevi. Herein, we describe original non-nevoid characteristics in CO, such as hyporeflectivity and intrinsic changes in reflectivity. Therefore, our study supports the use of SD-OCT to help differentiate these tumours.11
Underlying neovascularisation is a sight-threatening complication of CO, and can be treated with photodynamic therapy, transpupillary therapy, argon laser photocoagulation and intravitreal anti-VEGF agents.12–15 This study suggests that SD-OCT will aid in the detection and surveillance of CNV, by demonstrating changes in the reflectivity of CNV and decalcification that can occur after laser,16 as well as monitoring for resolution of subretinal and intraretinal fluid.
TD-OCT versus SD-OCT
With respect to tumour infrared reflectivity, various profiles (ie, isoreflective and hyper-reflective) were previously described with TD-OCT.16 17 CO reflectivity is subject to changes induced by varying wavelengths of the light source of the OCT devices, media opacities, changes in the overlying retinal tissue and observer interpretation. However, we report the first cases, to our knowledge, of CO that demonstrate SD-OCT hyporeflectivity to infrared. In our series of patients, tumour reflectivity ranged from markedly hyporeflective to hyper-reflective. Moreover, there was a significant trend towards a positive correlation between tumour reflectivity and tumour thickness, leading us to hypothesise that tumour reflectivity can be correlated to thickness and may change (over time) along with bone maturation (table 1).6
Previous reports have recognised the existence of ‘tracks’ of hyper-reflectivity, or of a ‘cavernous’ structure within the tumour.4 5 In our study, SD-OCT better visualised these hyporeflective structures in three cases. Further, we have noted them to be remarkably similar in appearance to the cancellous regions within CO pathology described by Gass et al.2
In our study, retinal changes induced by tumour decalcification were consistent with TD-OCT findings described by Shields et al.12 However, our study adds additional findings of retinal injury apparently induced by a mechanical effect of tumour on the vascular supply to the photoreceptors (figure 3). This has been observed with choroidal nevi.18
SD-OCT offers a new method to measure lateral tumour dimensions as well as (when light is transmitted through the whole tumour) thickness of CO. This is important in that the assessment of tumour dimensions on ultrasonography is typically hampered by resolution and posterior tumour shadowing that prevents exact delimitation of their posterior edge. Further, ultrasonographic image acquisition is complex and operator dependent, and resolution of devices are usually limited to 200 μm.
In this study, we found that ultrasound-measured tumour diameters were larger than those derived from fundus photographs. Though this could be related to differences in methods or software design, it suggests that photographic measurements underestimated the size of CO because its extension into the deeper layers of the choroid may not be visible (figure 3). The ability of OCT (both TD and SD) is also limited in measuring the tumour diameter. For example, in our study, the posterior aspect of only four CO could be visualised on SD-OCT (figure 4). The posterior aspect of CO is often variable due to shadowing created by the tumour's internal reflectivity. In addition, the lateral dimensions of CO are often too wide to fit the OCT's scanning field. More peripheral COs may not be reachable with most OCT machines that are designed to image the macula and optic nerve. Therefore, only small COs at the posterior pole are ideal for OCT-based measurements with current machines.
However, our study found that SD-OCT was able to improve the delineation of CO edges, and therefore with newer software and improvements in the flexibility of the scan angle, SD-OCT has the potential to outperform ultrasound and photography in measuring tumour dimensions. Further studies using enhanced depth imaging SD-OCT or SD-OCT with light sources of longer wavelengths may also enable a more accurate ascertainment of the choroidal thickness parameters important for tumour growth monitoring.19 20
The mean age of patients in this series is notably high.1 3 21 This probably reflects the fact that most of our patients had been diagnosed with CO for several years prior to their SD-OCT examination. In consideration of this reported age difference, we considered the alternative diagnosis of sclerochoroidal calcification (the latter being known to occur in older patients).9 We did not consider the tumours in this series to be sclerochoroidal calcifications because they were not bilateral nor did they have a sex distribution of 50:50. In addition, all patients were derived from an established ophthalmic oncology referral centre.
As we lack guidelines for optimal management of CO, future studies will determine how SD-OCT, TD-OCT and ultrasonography can define an appropriate time where intervention to prevent CO growth or decalcification is required. However, we herein revealed unique images of CO, the overlying retina and adjacent structures that led to insights and correlations to the existing literature. Our findings support the continued investigation of SD-OCT for diagnosis and follow-up of CO.
Acknowledgments
The authors thank Drs Julian Garcia, Pratima Kathil and Peter Gold for providing assistance during data acquisition.
References
Footnotes
Funding This research work was supported by The Eye Cancer Foundation, Inc. (http://eyecancerfoundation.net/). Dr Freton received a Fellowship grant from The Eye Cancer Foundation, New York. The sponsor had no role in the design or conduct of this research.
Competing interests None.
Ethics approval This study was conducted with the approval of the Institutional Review Board of The New York Eye Cancer Center, New York, New York, USA.
Provenance and peer review Not commissioned; externally peer reviewed.