1. PSEUDOMELANOMAS OF THE
POSTERIOR UVEAL TRACT
The 2006 Taylor R. Smith Lecture
JERRY A. SHIELDS, MD, ARMAN MASHAYEKHI, MD, SEONG RA, BS,
CAROL L. SHIELDS, MD
Purpose: To determine the types and frequency of lesions that clinically simulate
choroidal or ciliary body melanoma (posterior uveal melanoma; PUM).
Patients and Methods: A review was conducted on cases of patients referred to the
ocular oncology service from October 1978 through September 2003 with the diagnosis of
possible PUM but who were subsequently diagnosed by the authors to have a simulating
lesion rather than PUM. The type and percent of pseudomelanomas were tabulated and
compared with findings of a similar study from our service on data collected before 1978.
Results: There were 12,000 patients referred because of a lesion believed to be a
PUM during the 25 years included in the data collection. Of these patients, 1,739 (14%)
were found to have a simulating condition. There were 54 different conditions that simu-
lated melanoma. The most frequent condition was choroidal nevus, accounting for 851
cases (49%) of the pseudomelanomas. This was followed by peripheral exudative hem-
orrhagic chorioretinopathy (139 cases; 8%), congenital hypertrophy of the retinal pigment
epithelium (108 cases; 6%), hemorrhagic detachment of the retina or pigment epithelium
(86 cases; 5%), circumscribed choroidal hemangioma (79 cases; 5%) and age-related
macular degeneration (76 cases; 4%). Compared with the 1980 report, the rate of
pseudomelanomas diagnosed as choroidal nevus increased from 26% to 49%.
Conclusion: A variety of lesions can simulate PUM. Suspicious choroidal nevus is still the
lesion most difficult to differentiate from PUM. Most other pseudomelanomas account for a
lower percent compared with findings from the prior study, suggesting that clinicians are now
more familiar with the other pseudomelanomas and less likely to refer them to rule out PUM.
RETINA 25:767–771, 2005
S everal lesions can clinically simulate ciliary body
or choroidal melanoma (posterior uveal melano-
ma; PUM).1– 4 Historically, many eyes with simulating
lesions were enucleated because PUM was highly
suspected.5–7 With increased awareness of the clin-
ical features of these pseudomelanomas combined
with selective ancillary studies and more conserva-
From the Oncology Service, Wills Eye Hospital, Thomas Jef- tive treatments, the problem of erroneous enucle-
ferson University, Philadelphia, Pennsylvania. ation has been alleviated.8 However, there are still
Supported by the Eye Tumor Research Foundation, Philadel-
phia, PA (Drs. C. Shields and J. Shields), the Award of Merit in many patients referred to an ocular oncology center
Retina Research, Houston, TX (Dr. J. Shields), the Macula Foun- with the diagnosis of PUM who are subsequently
dation, New York, NY (Dr. C. Shields), and the Rosenthal Award
of the Macula Society (Dr. C. Shields).
diagnosed to have a simulating condition. We report
To be presented as the 2006 Taylor R. Smith Lecture, Aspen our experience with these simulating lesions over
Retinal Detachment Society, Aspen, Colorado, March 8, 2006. the last 25 years and compare the results with those
Reprint requests: Jerry A. Shields, MD, Ocular Oncology Ser-
vice, Wills Eye Hospital, 840 Walnut Street, Philadelphia, PA of an earlier study of pseudomelanomas from the
19107; e-mail: jerry.shields@shieldsoncology.com same facility.
767

2. 768 RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES ● 2005 ● VOLUME 25 ● NUMBER 6
Patients and Methods Table 1. Demographic Features of Pseudomelanoma in
1,739 Consecutive Patients
Since 1974, we have recorded the referral diagnosis
of all new patients seen at the oncology service. If the Feature Value
patient had a referral diagnosis of ciliary body or Age
choroidal melanoma but proved by examination to Mean 61 y
have a different diagnosis, the case was also coded as Median 64 y
Range 8 wk to 97 y
a pseudomelanoma. We included only those patients
Race, no. (%)*
who were referred because of legitimate concern over White 1,643 (94)
PUM. In this study, we reviewed our files from Oc- African American 52 (3)
tober 1978 through September 2003 to determine the Hispanic 15 (1)
type and frequency of pseudomelanomas. Data were Asian 29 (2)
Sex, no. (%)
collected regarding patient age, race, and sex. The
Male 717 (40)
exact tissue affected and the final diagnoses were Female 1,022 (60)
recorded. We compared the incidence of the specific Follow-up (mo)
pseudomelanomas with those from a prior study from Mean 21
our service in which data collection included the 5 Median 5
Range 6–299
years before October 1978.4
Because we did not have histopathologic confirma- * Information on race was not available for 56 patients.
tion that a small melanocytic choroidal lesion was a
nevus, rather than melanoma, we relied on certain
criteria that we used for years to make a presumptive senting 14% of all patients referred with the diagnosis
diagnosis. We generally coded a melanocytic lesion as of possible PUM. Thus, 86% of patients referred with
a nevus if it were 6 mm in diameter and 2.5 mm a diagnosis of PUM had a correct diagnosis. The most
in thickness and as a melanoma if it were 6 mm in frequent pseudomelanoma was choroidal nevus,
diameter and 2.5 mm in thickness. However, there which accounted for 851 cases or 49% of the total
was some variability in coding such lesions. For ex- number of simulating lesions. Other conditions that
ample, if a lesion were in the nevus size range but had were referred to rule out PUM are shown in Table 2.
surface orange pigment, subretinal fluid, and docu-
mented growth, we upgraded it to a melanoma for Discussion
purposes of diagnostic coding, and it was not included
The conditions that can simulate PUM and the
with the pseudomelanoma tabulations. Likewise, if a
features that help distinguish them from PUM are well
lesion were barely in the melanoma size range but had
described in the literature.1– 4 In our series, choroidal
surface drusen, had no subretinal fluid, and had not
nevus was the most frequent pseudomelanoma, ac-
been documented to, it was coded as a nevus and was
counting for 49% of cases. This represents a sharp
included in the study as a pseudomelanoma. All pa-
increase since the 1980 report,4 in which choroidal
tients were observed by us or by referring physicians,
nevus represented 27% of cases. We realize that by
and in the rare instance where a presumed nevus
using the criteria for differentiating nevus from mel-
showed subsequent growth, it was reclassified as a
anoma as described under Patients and Methods we
melanoma and was not included in the pseudomela-
could have made a few errors in the diagnosis of
noma statistics.
borderline lesions. However, based on our clinical
experience and on publications in the literature, such
Results
categorization should separate nevus from melanoma
During the 25 years included in the data collection, in most cases. A lesion with two or more risk factors
there were 12,000 patients referred to the oncology for metastasis was generally classified as melanoma,
service because of a lesion suspected to be a possible depending on all of the clinical circumstances.9 –11
PUM. General ophthalmologists or retinal specialists There are possible explanations why choroidal ne-
referred most patients. Demographic information is vus accounted for a higher percent of pseudomela-
included in Table 1. Most patients were adult whites. noma in this series as compared with prior studies.
Pseudomelanomas occurred in middle-aged or older First, ophthalmologists have become more familiar
patients, similar to the age distribution for PUM. The with the clinical features of the various lesions that
number and percent of the various pseudomelanomas simulate melanoma, like disciform macular degener-
are shown in Table 2. There were 1,739 patients ation, congenital hypertrophy of the retinal pigment
whose cases were coded as pseudomelanomas, repre- epithelium (RPE), choroidal hemangioma, and choroi-

3. PSEUDOMELANOMAS OF POSTERIOR UVEA • SHIELDS ET AL 769
Table 2. Diagnoses of Pseudomelanoma in 1,739 peripheral exudative hemorrhagic chorioretinopathy
Consecutive Patients (peripheral diskiform degeneration), which accounted
No. (%) of for 139 cases (8%). It accounted for 11% of cases in
Diagnosis Patients our prior series and still represents a lesion that often
prompts referral for a possible PUM.2,4 However, the
Choroidal nevus 851 (49)
Peripheral exudative hemorrhagic 139 (8) classic appearance of subretinal blood in various
chorioretinopathy stages of resolution is different from a comparable-
Congenital hypertrophy of RPE 108 (6) sized peripheral choroidal melanoma that would be
Hemorrhagic detachment retina or 86 (5)
pigment epithelium unlikely to cause appreciable subretinal hemorrhage.
Circumscribed choroidal hemangioma 79 (5) We acknowledge that there could be some overlap
Age-related macular degeneration 76 (4) among the lesions that we called peripheral exudative
Hyperplasia of RPE 42 (2)
Optic disk melanocytoma 37 (2) hemorrhagic chorioretinopathy, hemorrhagic detach-
Choroidal metastasis 34 (2) ment of the retina or RPE, and age-related macular
Hemorrhagic choroidal detachment 29 (2) degeneration. However, we coded a lesion as periph-
Vasoproliferative tumor 20 (1)
Rhegmatogenous retinal detachment 18 (1) eral exudative hemorrhagic chorioretinopathy if it oc-
Choroidal detachment 17 (1) curred as a hemorrhagic and/or exudative lesion in the
Uveal effusion syndrome 17 (1) equatorial region in an older person in the setting of
Choroidal or disk granuloma 14 (1)
Adenoma of RPE or CPE 13 ( 1) peripheral drusen. We coded any other cause of hem-
Sclerochoroidal calcification 12 (1) orrhage besides peripheral exudative hemorrhagic chori-
Staphyloma 12 (1)
Cataract 10 (1) oretinopathy and age-related macular degeneration as
Retinal capillary hemangioma (capillary 10 (1) hemorrhagic detachment of the retina or RPE. Causes of
or cavernous) such bleeding included hemorrhage from retinal macroa-
Adenoma of nonpigmented CPE 10 ( 1)
Leiomyoma, ciliary body 10 ( 1) neurysm, polypoidal choroidopathy, trauma, anticoagu-
Degenerative retinoschisis 8( 1) lant use, and several other conditions.
Retinal cavernous hemangioma 7( 1) The third most frequent pseudomelanoma was sol-
Chorioretinal scar 7( 1)
Vortex vein varix 7( 1) itary congenital hypertrophy of the RPE, accounting
Vitreous hemorrhage 7( 1) for 108 cases (6%).12–14 Larger congenital hypertro-
Choroidal osteoma 5( 1) phy of the RPE lesions that are located in the periph-
Preretinal macular gliosis 5( 1)
Scleritis 5( 1) eral fundus can frequently give the illusion of greater
Combined hamartoma of retina and RPE 4( 1) elevation. However, the sharp border, distinct black or
Ocular melanocytosis (choroidal) 4( 1) gray color, depigmented or pigmented halo, and de-
Subluxated lens 4( 1)
Compression by orbital tumor 3( 1) pigmented lacunae in the lesion should differentiate it
Central retinal vein obstruction 3( 1) from melanoma. Although congenital hypertrophy of
Retinal foreign body 3( 1)
Lens fragments/remnants 3( 1) the RPE was traditionally believed to be a stationary
Neurilemoma 3( 1) lesion, it is now known to gradually enlarge in most
Limited choroidal hemorrhage 3( 1) cases.12,14 In addition, congenital hypertrophy of the
Coloboma 2( 1)
Pars plana cyst 2( 1) RPE can rarely spawn an elevated component that we
Hazy media with suspicious ultrasound 2( 1) believe represents an adenoma of the RPE.15 In one
findings such case in which histopathologic examination was
White without pressure 2( 1)
Familial exudative vitreoretinopathy 1( 1) performed, the lesion proved to be an adenocarcinoma
Lattice degeneration of retina 1( 1) of the RPE.16
Choroidal lymphoma 1( 1) Other lesions that can resemble PUM are listed in
Myelinated nerve fibers 1( 1)
Neurofibroma 1( 1) Table 2. Hemorrhagic detachment of the retina or RPE
Optic disk/retina astrocytic hamartoma 1( 1) accounted for 86 pseudomelanomas (5%). This was
RPE, retinal pigment epithelium; CPE, ciliary body epithelium.
different from typical age-related macular degenera-
tion or peripheral exudative hemorrhagic chorioreti-
nopathy by virtue of its parafoveal or postequatorial
dal metastasis, and are less likely to refer them as location and could have been due to a number of
suspected melanoma. Second, our group in recent unclear causes, as mentioned above. Choroidal metas-
years has propagated treatment of selected borderline tasis can resemble amelanotic choroidal melanoma.
lesions that possess reported risk factors for growth The ophthalmoscopic features that differentiate cho-
and metastasis, thus stimulating more referrals for roidal melanoma from choroidal metastasis have been
borderline lesions. reported.1– 4,5,17 In addition, ancillary studies like flu-
The second most frequent pseudomelanoma was orescein angiography and ultrasonography can be

4. 770 RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES ● 2005 ● VOLUME 25 ● NUMBER 6
helpful in the differentiation.1–5 In some cases, fine- yellow lesion, usually found along the retinal vascular
needle aspiration biopsy of the intraocular mass may arcades, can resemble a amelanotic melanoma. Varix
be necessary to establish the diagnosis.18 Circum- of a vortex vein ampulla can dilate in certain fields of
scribed choroidal hemangioma is another condition gaze and appear as an elevated choroidal mass.42,43
that can mimic choroidal melanoma,2,3,19,20 account- However, it flattens or collapses in other fields of gaze
ing for 5% of pseudomelanomas in our series. It has a or when slight pressure is applied to the globe. Other
typical red orange color, shows early hyperfluores- unusual examples of pseudomelanoma included staph-
cence with angiography, and has high internal reflec- yloma (12 cases), dense cortical cataract that was
tivity with ultrasonography.2,3,20 believed to be a pigmented ciliary body melanoma
Lesions like combined hamartoma of the retina and (10), totally dislocated lens in the posterior fundus (4),
RPE,21,22 reactive hyperplasia of the RPE, optic disk fundus foreign body for which the patient recalled no
melanocytoma,23–26 retinal vasoproliferative tumor,27 ocular trauma (3), and others.
and choroidal osteoma28,29 usually have rather distinc- In summary, we reviewed the diagnoses and fre-
tive features that should serve to differentiate them quency of lesions that prompted referral to our oncol-
from melanoma.21–29 Some intraocular tumors, like ogy service because of suspected PUM and compared
ciliary body leiomyoma,2,3,30 neoplasms of the pig- their frequency with findings from an older study.
ment epithelium,2,3,31,32 and nonpigmented ciliary ep- Choroidal nevus still accounts for most pseudomela-
ithelium,33 were once believed to be indistinguishable nomas, and its differentiation from small PUM re-
from PUM and may still be difficult in certain cases. mains a clinical dilemma. Other lesions that were
However, recent studies have addressed their differ- more difficult to differentiate from melanoma a few
ences from melanoma.30 –33 years ago, like congenital hypertrophy of the RPE,
There are other tumors that may be impossible choroidal hemangioma, and age-related macular de-
clinically to differentiate from melanoma. These in- generation, are now being diagnosed more accurately
clude benign peripheral nerve sheath tumors such as and account for a lower percent of pseudomelanomas.
choroidal neurofibroma and neurilemoma (schwanno- The salient features that differentiate some of these
ma).2,3,34 These rare spindle cell uveal tumors can be lesions from PUM have been elucidated.
identical to amelanotic melanoma with regard to oph-
thalmoscopic features and ancillary studies. We have References
seen two patients with schwannoma who underwent
enucleation after failure of plaque radiotherapy for 1. Gass JDM. Differential Diagnosis of Intraocular Tumors. St.
suspected melanoma, and the diagnosis of neurilem- Louis: CV Mosby; 1974:1–362.
2. Shields JA, Shields CL. Differential diagnosis of posterior
oma was first realized by histopathologic examination. uveal melanoma. In: Shields JA, Shields CL, eds. Intraocular
One case of pigmented schwannoma was impossible Tumors. A Text and Atlas. Philadelphia: Saunders; 1992:
to differentiate from melanoma.35 137–153.
Some inflammatory conditions can also simulate 3. Shields JA, Shields CL. Atlas of Intraocular Tumors. Phila-
posterior uveal melanoma. Solitary choroiditis for delphia: Lippincott Williams & Wilkins; 1999:53–202.
4. Shields JA, Augsburger JJ, Brown GC, Stephens RF. The
which no specific etiology has been determined also differential diagnosis of posterior uveal melanoma. Ophthal-
prompts referral because of suspected amelanotic mology 1980;87:543–548.
PUM.36 In addition, nodular posterior scleritis can be 5. Ferry AP. Lesions mistaken for malignant melanoma of the
similar to amelanotic melanoma but has different fea- posterior uvea. Arch Ophthalmol 1964;72:463– 469.
tures.37,38 In borderline cases, a short course of cortico- 6. Shields JA, Zimmerman LE. Lesions simulating malignant
melanomas of the posterior uvea. Arch Ophthalmol 1973;89:
steroid treatment may help make the correct diagnosis. 466 – 471.
There are several other unusual and fascinating 7. Chang M, Zimmerman LE, McLean I. The persisting pseudo-
lesions that can simulate melanoma and have received melanoma problem. Arch Ophthalmol 1984;102:726 –727.
little or attention in prior reports of pseudomelanoma. 8. Shields JA, McDonald PR. Improvements in the diagnosis of
Limited choroidal hemorrhage that sometimes occurs posterior uveal melanomas. Arch Ophthalmol 1974;91:259 –
264.
after cataract surgery can resemble PUM.39 We have 9. Factors predictive of growth and treatment of small choroidal
seen patients who had fundus photographs and ultra- melanoma: COMS report no. 5. The Collaborative Ocular
sonograms reported to be typical of melanoma, and Melanoma Study Group. Arch Ophthalmol 1997;115:1537–
the transient choroidal hemorrhage had resolved by 1544.
the time the patient arrived for consultation. Sclero- 10. Shields CL, Shields JA, Kiratli H, Cater JR, De Potter P. Risk
factors for metastasis of small choroidal melanocytic lesions.
choroidal calcification has received attention in the Ophthalmology 1995;102:1351–1361.
recent literature, and its association with certain sys- 11. Shields CL, Cater JC, Shields JA, Singh AD, Santos MCM,
temic diseases has been discussed.40,41 This nodular Carvalho C. Combination of clinical factors predictive of

5. PSEUDOMELANOMAS OF POSTERIOR UVEA • SHIELDS ET AL 771
growth of small choroidal melanocytic tumors. Arch Oph- cal manifestations in 103 patients. Arch Ophthalmol 1995;
thalmol 2000;118:360 –364. 113:615– 623.
12. Chamot L, Zografos L, Klainguti G. Fundus changes associ- 28. Gass JDM, Guerry RK, Jack RL, Harris G. Choroidal os-
ated with congenital hypertrophy of the retinal pigment epi- teoma. Arch Ophthalmol 1978;96:428 – 435.
thelium. Am J Ophthalmol 1993;115:154 –161. 29. Shields CL, Shields JA, Augsburger JJ. Review: choroidal
13. Purcell JJ, Shields JA. Hypertrophy with hyperpigmentation osteoma. Surv Ophthalmol 1988;33:17–27.
of the retinal pigment epithelium. Arch Ophthalmol 1975;93: 30. Shields JA, Shields CL, Eagle RC Jr, De Potter P. Observa-
1122–1126. tions on seven cases of intraocular leiomyoma. The 1993
14. Shields CL, Mashayekhi A, Ho T, Cater J, Shields JA. Byron Demorest Lecture. Arch Ophthalmol 1994;112:521–
Solitary congenital hypertrophy of the retinal pigment epi- 528.
thelium: clinical features and frequency of enlargement in 31. Shields JA, Shields CL, Gunduz K, Eagle RC Jr. Adenoma of
330 patients. Ophthalmology 2003;110:1968 –1976. the ciliary body pigment epithelium. The 1998 Albert Ruede-
15. Shields JA, Shields CL, Singh AD. Acquired tumors arising mann Sr. Memorial Lecture. Part 1. Arch Ophthalmol 1999;
from congenital hypertrophy of the retinal pigment epithe- 117:592–597.
lium. Arch Ophthalmol 2000;118:637– 641. 32. Shields JA, Shields CL, Gunduz K, Eagle RC Jr. Neoplasms
16. Shields JA, Shields CL, Eagle RC Jr, Singh AD. Adenocar- of the retinal pigment epithelium. The 1998 Albert Ruede-
cinoma arising from congenital hypertrophy of the retinal mann Sr. Memorial Lecture. Part 2. Arch Ophthalmol 1999;
pigment epithelium. Arch Ophthalmol 2001;119:597– 602. 117:601– 608.
17. Shields CL, Shields JA, Gross N, Schwartz G, Lally S. 33. Shields JA, Eagle RC Jr, Shields CL, De Potter P. Acquired
Survey of 520 eyes with uveal metastases. Ophthalmology neoplasms of the nonpigmented ciliary epithelium (adenoma
1997;104:1265–1276. and adenocarcinoma). The 1995 F. Phinizy Calhoun Jr Lec-
18. Shields JA, Shields CL, Ehya H, Eagle RC Jr, De Potter P.
ture. Ophthalmology 1996;103:2007–2016.
Fine needle aspiration biopsy of suspected intraocular tu-
34. Shields JA, Sanborn GE, Kurz GH, Augsburger JJ. Benign
mors. The 1992 Urwick Lecture. Ophthalmology 1993;100:
peripheral nerve tumor of the choroid. Ophthalmology 1981;
1677–1684.
88:1322–1329.
19. Witschel H, Font RL. Hemangioma of the choroid. A clini-
35. Shields JA, Font RL, Eagle RC Jr, Shields CL, Gass JDM.
copathologic study of 71 cases and a review of the literature.
Melanotic schwannoma of the choroid: immunohistochemis-
Surv Ophthalmol 1976;20:415– 431.
try and electron microscopic observations. Ophthalmology
20. Shields CL, Honavar SG, Shields JA, Cater J, Demirci H.
1994;101:843– 849.
Circumscribed choroidal hemangioma. Clinical manifesta-
tions and factors predictive of visual outcome in 200 consec- 36. Shields JA, Shields CL, Demirci H, Honavar S. Solitary
utive cases. Ophthalmology 2001;108:2237–2248. idiopathic choroiditis. Richard B. Weaver Lecture. Arch
21. Gass JDM. An unusual hamartoma of the pigment epithelium Ophthalmol 2002;120:311–319.
and retina simulating choroidal melanoma and retinoblas- 37. Benson WE, Shields JA, Tasman WS, Crandall AS. Posterior
toma. Trans Am Ophthalmol Soc 1973;71:171–185. scleritis. Arch Ophthalmol 1979;97:1482–1486.
22. Schachat AP, Shields JA, Fine SL, et al. Combined hamar- 38. Benson WE. Posterior scleritis. Surv Ophthalmol 1988;32:
toma of the retina and retinal pigment epithelium. Ophthal- 297–316.
mology 1984;91:1609 –1615. 39. Morgan CM, Gragoudas ES. Limited choroidal hemorrhage
23. Zimmerman LE, Garron LK. Melanocytoma of the optic disc. mistaken for a choroidal melanoma. Ophthalmology 1987;
Int Ophthalmol Clin 1962;2:431– 440. 94:41– 46.
24. Zimmerman LE. Melanocytes, melanocytic nevi and mela- 40. Honavar SG, Shields CL, Demirci H, Shields JA. Sclerocho-
nocytomas. Invest Ophthalmol 1965;4:11– 41. roidal calcification: clinical manifestations and systemic as-
25. Joffe L, Shields JA, Osher R, Gass JDM. Clinical and fol- sociations. Arch Ophthalmol 2001;119:833– 840.
low-up studies of melanocytomas of the optic disc. Ophthal- 41. Shields JA, Shields CL. Sclerochoroidal calcification. Re-
mology 1979;86:1067–1078. view. The 2001 Harold Gifford Lecture. Retina 2002;22:251–
26. Shields JA, Demirci H, Mashayekhi A, Shields CL. Melano- 261.
cytoma of the optic disc in 115 Case. The 2004 Samuel 42. Osher RH, Abrams GW, Yarian D, Armao D. Varix of the
Johnson Memorial Lecture. Ophthalmology 2004;111:1933– vortex ampulla. Am J Ophthalmol 1981;92:653– 660.
1934. 43. Gunduz K, Shields CL, Shields JA. Varix of vortex vein
27. Shields CL, Shields JA, Barrett J, De Potter P. Vasoprolif- simulating choroidal melanoma. Report of four cases. Retina
erative tumors of the ocular fundus. Classification and clini- 1998;18:343–347.