Venn diagram of the distribution of clinical features at the initial examination. RD indicates retinal detachment; CNV, choroidal neovascularization.
Right eye of patient 36 at the initial examination (left) and 2 years later (right), showing even growth at all margins.
Right eye of patient 33 at the initial examination (left) and 2 years later (right) showing growth of a pseudopodium.
Kaplan-Meier survival curve for a visual acuity of 20/200 or worse.
Choroidal neovascularization involving the fovea in patient 7.
Kaplan-Meier survival curve for the development of choroidal neovascularization.
Aylward GW, Chang TS, Pautler SE, Gass JDM. A Long-term Follow-up of Choroidal Osteoma. Arch Ophthalmol. 1998;116(10):1337-1341. doi:10.1001/archopht.116.10.1337
To provide long-term follow-up information on a large series of patients with choroidal osteoma.
Review of patients with a diagnosis of choroidal osteoma who had been examined at the Bascom Palmer Eye Institute, Miami, Fla, or known to one of us (J.D.M.G.). Information was obtained from hospital medical records or by a questionnaire sent to referring ophthalmologists. Life-table analysis was used to study the loss of vision and development of choroidal neovascularization.
We followed up 36 patients, 31 (89%) were female, mean age, 21 years (range, 5-54 years) for a mean of 10 years (range, 2-22 years). Growth was observed for 9 (41%) of 22 well-documented osteomas. The probability of loss of visual acuity to 20/200 or worse was 58% by 10 years and 62% by 20 years. The probability of developing choroidal neovascularization was 47% by 10 years and 56% by 20 years. Successful treatment of the choroidal neovascularization with laser photocoagulation was performed for 5 (25%) of 20 patients.
Most patients with choroidal osteomas maintain good vision in at least 1 eye, but they have a high risk of developing choroidal neovascularization. When this occurs, only a minority can be successfully treated with laser photocoagulation.
CHOROIDAL OSTEOMA is a rare benign tumor characterized by the presence of cancellous bone within the peripapillary choroid. At the meeting of the Verhoeff Society in 1975, during a discussion of false-positive phosphorus 32 tests, a case of an osseous tumor of the choroid contributed to the Armed Forces Institute of Pathology by Henry Van Dyk, MD, was reported. This case and 3 others were later reported by Gass et al,1 who pointed out that histopathologic features of a choroidal osteoma had been described in Reese's textbook2 but had been misdiagnosed as an ossified choroidal hemangioma. The same case also was reported by Williams et al.3 Since then, the clinical features of choroidal osteoma have been well established.4- 7 The tumor predominantly affects young females but can occur in males.4,8 A serious complication is the formation of choroidal neovascularization (CNV) that may lead to the loss of vision.1,4,9 Limited treatment success with laser photocoagulation has been reported.10- 14 Growth of the tumor has been documented, and involution may occur spontaneously15,16 after laser photocoagulation17,18 or after treatment with corticosteroids.19(pp152-153),20 We studied a large series of patients with choroidal osteoma to determine growth patterns, occurrence of complications, effects of treatment, and the long-term prognosis for vision.
We reviewed the clinical and photographic records for all patients at the Bascom Palmer Eye Institute, Miami, Fla, with a diagnosis of choroidal osteoma. Osteoma was defined as an orange, irregular, subretinal mass, with spiderlike vascular tufts on its surface, and bone density was demonstrated by radiography, echography, or computed tomography. Patients had been examined at the Bascom Palmer Eye Institute, or their photographs had been sent to one of us (J.D.M.G.). Information was extracted from the hospital chart or from a questionnaire sent to the referring ophthalmologist. The questionnaire was standardized and inquired about systemic disease, including diabetes. If the patient had moved or changed physicians, we contacted the current ophthalmologist. For all patients, postal and telephone inquiry continued until follow-up details were obtained or no further information was available. Informed consent was obtained from all patients contacted.
We collected details of history, visual acuity, and ocular findings. The location, appearance, and size of the tumor were determined by reviewing clinical photographs, or for 1 patient, carefully constructed retinal drawings. Size was expressed in units of disc areas. The percentage of growth was calculated by comparing photographs from the initial examination with those from the final examination. A subset analysis of growth was performed for patients with at least 5 years of follow-up and good quality stereophotographic documentation.
The onset of poor vision and the development of CNV were analyzed by using life-table methods. In the eyes that had received laser photocoagulation for CNV, follow-up information obtained after treatment was excluded to eliminate the potential modifying effect of laser on the behavior of the tumor. For the life-table analyses, the data were censored after the time point of laser treatment.
There were 50 patients who had been given a diagnosis of choroidal osteoma. Eight had been examined only once and were subsequently lost to or unavailable for follow-up, 1 patient had died, and no information was forthcoming for 5 patients. These 14 patients (28%) were excluded from the study. The remaining 36 patients (72%) comprised 15 (42%) who had been examined within the previous 12 months and 21 (58%) who had been followed up for at least 2 years but had been subsequently lost to or unavailable for follow-up.
The baseline details, initial and final visual acuity measurements, length of follow-up, and state of the macula are given in Table 1. There were 31 females (86%) and 5 males (14%). Thirty-two (89%) were white, 1 (3%) was black, 2 (6%) were Hispanic, and 1 (3%) was Asian. The mean age at the initial examination was 21 years (range, 5 to 54 years). The right eye was affected in 11 patients (30%) and the left eye in 16 (44%); 9 (25%) had bilateral osteoma at the initial examination. For patients with bilateral osteoma, only data from the symptomatic eye were included in further analysis.
A history of trauma was elicited from 4 patients (11%). Most patients, 32 (89%), had no evidence of systemic disease at the initial examination. One patient was pregnant but healthy, 1 had a prolactin-producing adenoma, 1 had diabetes, and 1 had multiple intracranial meningiomas. A history of previous ocular or periocular inflammation was obtained from 3 patients (8%). Of the 3 patients, 1 had experienced an episode of uveitis, 1 had experienced corneal inflammation due to injury by a cactus quill, and 1 had Harada disease. The family history was negative for choroidal osteoma and CNV for all patients except 3 (patients 2-4), who were all siblings.
The presenting symptoms were as follows: blurred vision, 29 (80%); metamorphopsia, 7 (19%); photophobia, 1 (3%); and strabismus, 1 (3%). For 3 patients (8%), no symptoms were present, and the choroidal osteoma was noted during a routine examination. Subnormal visual acuity at the initial examination was due to the presence of subfoveal osteoma, serous retinal detachment, CNV, or a combination of these. The numbers of eyes with these findings are shown in Figure 1. The osteomas were juxtapapillary in 33 eyes (92%) and extended beneath the center of the fovea in 14 eyes (39%). The average size was 14 disc areas (range, 2-30 disc areas).
Patients were followed up for a mean of 10 years (range, 2-22 years). Thirty patients (83%) were followed up for 5 years or more, and of these, 22 (73%) had sufficiently good photographic documentation to allow analysis of growth of the osteoma. An increase in size was documented for 9 osteomas (41%), including 2 that grew beneath the fovea. Growth occurred by gradual enlargement at all margins (Figure 2) or by formation of a pseudopodium (Figure 3). In 2 patients (9%), the osteoma decreased in size, presumably owing to osteoclastic activity; one of these has been described.17 In 2 patients (9%), new osteomas developed in the previously healthy contralateral eye.
The results of life-table analysis performed for preservation of visual acuity better than 20/200 were as follows:
The results also are shown in Figure 4. The reasons for visual loss (acuity 20/200 or worse) at the final examination were as follows: CNV, 12 (33%); chronic retinal detachment, 4 (11%); and atrophy of the retinal pigment epithelium, 4 (11%). Only 1 patient with bilateral osteoma sustained visual loss to an acuity of 20/200 or worse in both eyes.
The major complication during follow-up and the major cause of visual loss was the development of CNV (Figure 5). The probability of developing CNV was determined by using life-table analysis, the results of which are as follows:
Results also are shown in Figure 6. In 20 eyes, CNV, which was classic in all patients, developed. In 11 eyes, the CNV was extrafoveal and was treated with laser photocoagulation. In 5 eyes, the membrane was considered to be obliterated. Argon laser was used for treatment of all patients. In 1 patient with persistent hyperfluorescence, 2 further treatments with krypton laser were performed, but without success.
Of the 22 eyes with serous retinal detachments at the initial examination, the detachment had resolved in 14 (64%) by the time of the final examination. Of the remaining 8 eyes, 7 (88%) had a final visual acuity of 20/200 or worse, although 3 of these eyes also had CNV.
The results of this study confirm the results of previous work about the manifestations and clinical features of choroidal osteoma.4- 8,24 Most patients were young healthy females with a single affected eye. The proportion with bilateral osteoma (25%) is the same as has been previously reported.24 Several familial cases are reported in the literature,21,25,26 but we could find no further cases in our series other than the 3 siblings whose cases have been reported.21 In early publications, an association with trauma was suspected,4,27 but we have been unable to confirm this. The proportion of patients giving a history of trauma (11%) is no greater than would be expected in a series of young people.
The pattern of growth of the osteomas in our study seemed quite variable. Overall, 9 (40%) of 22 well-documented osteomas showed evidence of growth, which is less than the 64% reported by Gass4 for a previous series. The majority (11/22 [50%]) of tumors remained unchanged for the duration of follow-up. When growth occurred, it was uniform with gradual enlargement from all margins or was uneven, with formation of pseudopodia. We could detect no association between growth and female sex, puberty, or any other factor. The tumors in 2 patients in our series showed a reduction in size. Involution or decalcification of choroidal osteomas has been reported and is believed to be due to a predominance of osteoclastic activity. The tumor in 1 patient grew slowly for 5 years before becoming thin and decalcified.15 Another tumor developed a deep depression in its center.16 Laser photocoagulation may lead to local resorption of bone17 or complete disappearance of the tumor.18
The probability of visual loss to an acuity of 20/200 or worse was 58% by 10 years, and 62% by 20 years. This is better than might be expected, given the limited follow-up in the literature. For example, 5 (36%) of 14 eyes with choroidal osteoma that were followed up for between 1 and 7 years lost vision to an acuity of 20/200 or worse.4 The major cause of visual loss in the present series was the development of CNV. The likely stimulus to neovascularization is the compromise of the Bruch membrane overlying the tumor.
Laser photocoagulation of CNV associated with osteoma has been described.10- 13 Argon laser was used for the treatment of all but 1 of the patients in our series in which the CNV was considered treatable. The rate of success in closing the membrane was 5 (45%) of 11 eyes, which compares poorly with what might be expected for CNV associated with age-related macular degeneration. The depigmentation of the retinal pigment epithelium that often is seen in osteomas reduces the absorption of laser energy, which might make treatment less effective. It would seem more logical to use argon laser, which is absorbed by hemoglobin, in these situations rather than krypton laser.
The cause of choroidal osteoma remains unknown. It has been suggested that it is an osseous choristoma.3 This suggestion is supported by the peripapillary location, a site favored by other developmental tumors, and by occurrence of the osteoma in the absence of any other disease process. Evidence against a choristomatous origin is sex predilection and potential for growth. An alternative cause is secondary ossification after some other disease process, such as inflammation or trauma. Although osteomas occur in these settings, the majority occur in patients with no evidence of previous systemic or ocular disease.
A case of multiple osteomas developing in association with bilateral pseudotumors of the orbit raised the possibility that inflammation may have a part in the cause.20 Other cases have occurred in association with ocular inflammation19(pp152-153),23 or systemic illness.28 It is of interest that 3 of the patients we studied had a definite history of inflammatory disease, but this may be simple coincidence, because most of the patients had no such history. Similarly, the few patients in our series with systemic disease do not provide strong evidence of a systemic contribution to the cause of the osteoma.
It is tempting to speculate that there is some endocrine influence moderating growth to explain the predilection for young females and the occurrence during pregnancy.19(pp178-179),29 However, we can offer no further evidence to support this theory. Noble21 suggested that there may by a congenital, perhaps inherited, abnormality of the peripapillary choroid that is undetectable clinically until calcification begins to occur under the influence of other factors.
We present long-term follow-up data on the largest series of choroidal osteoma yet reported; patients were followed up for up to 22 years. The probability of visual loss to an acuity of 20/200 or worse was more than half by 20 years. Only 1 patient with bilateral osteoma experienced visual loss in both eyes. The major complication leading to visual loss was the development of CNV, which had a poor response to treatment.
Accepted for publication May 19, 1998.
Dr Aylward was supported by a grant from the TFC Frost charitable trust, London, England. Dr Chang was an E. A. Baker Fellow (1993-1994).
We thank the many referring ophthalmologists who took time to complete detailed questionnaires on their patients, without whom this study would not have been possible.
Reprints: G. W. Aylward, MD, FRCS, FRCOphth, Moorfields Eye Hospital, City Road, London EC1V 2PD, England.