[Skip to Content]
Access to paid content on this site is currently suspended due to excessive activity being detected from your IP address 54.197.75.176. Please contact the publisher to request reinstatement.
[Skip to Content Landing]
Download PDF
Figure 1.
A, Baseline appearance of thesingle nevus (arrow) that demonstrated growth during 5 years. B, After 5 years,the nevus (arrow) is clearly larger, with a change in diameter from 0.49 to0.73 mm. There are no other features of progression.

A, Baseline appearance of thesingle nevus (arrow) that demonstrated growth during 5 years. B, After 5 years,the nevus (arrow) is clearly larger, with a change in diameter from 0.49 to0.73 mm. There are no other features of progression.

Figure 2.
A, Baseline photograph of thelargest nevus in our study population, with a diameter of 4.04 mm. The nevushas multiple overlying drusen. B, After 5 years, there are no changes otherthan partial depigmentation and an increasing number of drusen.

A, Baseline photograph of thelargest nevus in our study population, with a diameter of 4.04 mm. The nevushas multiple overlying drusen. B, After 5 years, there are no changes otherthan partial depigmentation and an increasing number of drusen.

Figure 3.
A, Scatterplot of largest nevusdiameter at baseline and follow-up. The line of linear regression has a slopeclose to 1.0, which indicates little change in nevus diameter across 5 years(Pearson correlation coefficient = 0.97). B, Scatterplot of nevus surfacearea measurements at baseline and follow-up. The slope of the linear regressionline is again close to 1.0 (Pearson correlation coefficient = 0.99). C, Bland-Altmanplot of nevus diameter measurements displays the level of agreement betweenbaseline and follow-up measurements. Many nevi lie above or below the mean(solid line), which is likely to represent measurement noise. The overallchange in size is minimal (+0.03 mm), as shown by the dotted line.

A, Scatterplot of largest nevusdiameter at baseline and follow-up. The line of linear regression has a slopeclose to 1.0, which indicates little change in nevus diameter across 5 years(Pearson correlation coefficient = 0.97). B, Scatterplot of nevus surfacearea measurements at baseline and follow-up. The slope of the linear regressionline is again close to 1.0 (Pearson correlation coefficient = 0.99). C, Bland-Altmanplot of nevus diameter measurements displays the level of agreement betweenbaseline and follow-up measurements. Many nevi lie above or below the mean(solid line), which is likely to represent measurement noise. The overallchange in size is minimal (+0.03 mm), as shown by the dotted line.

Characteristics of Choroidal Nevi at 5-Year Follow-up
Characteristics of Choroidal Nevi at 5-Year Follow-up
1.
Sumich  PMitchell  PWang  JJ Choroidal nevi in a white population: the Blue Mountains Eye Study. Arch Ophthalmol. 1998;116645- 650
PubMedArticle
2.
Gass  JD Problems in the differential diagnosis of choroidal nevi and malignantmelanoma: XXXIII Edward Jackson Memorial lecture. Trans Am Acad Ophthalmol Otolaryngol. 1977;8319- 48
PubMed
3.
Naumann  GYanoff  MZimmerman  LE Histogenesis of malignant melanomas of the uvea: histopathologic characteristicsof nevi of the choroid and ciliary body. Arch Ophthalmol. 1966;76 (pt 1) 784- 796
PubMedArticle
4.
Hale  PNAllen  RAStraatsma  BR Benign melanomas (nevi) of the choroid and ciliary body. Arch Ophthalmol. 1965;74532- 538
PubMedArticle
5.
Rodriguez-Sains  RS Ocular findings in patients with dysplastic nevus syndrome. Ophthalmology. 1986;93661- 665
PubMedArticle
6.
Smith  REGanley  JP Ophthalmic survey of a community: abnormalities of the ocular fundus. Am J Ophthalmol. 1972;74 ((pt 1)) 1126- 1130
PubMed
7.
Albers  EC Benign melanomas of the choroid and their malignant transformation. Am J Ophthalmol. 1940;23779- 783
8.
Wagener  HPWellbrock  WLA Benign melanoma and melano-epithelioma of the choroid. Arch Ophthalmol. 1930;4509- 514Article
9.
Yanoff  MZimmerman  LE Histogenesis of malignant melanomas of the uvea: relationship of uvealnevi to malignant melanomas. Cancer. 1967;20 ((pt 2)) 493- 507
PubMedArticle
10.
MacIlwaine  WAAnderson  B  JrKlintworth  GK Enlargement of a histologically documented choroidal nevus. Am J Ophthalmol. 1979;87480- 486
PubMed
11.
Ganley  JPComstock  GW Benign nevi and malignant melanomas of the choroid. Am J Ophthalmol. 1973;7619- 25.
PubMed
12.
Sallet  Gde Laey  JJ Follow-up of suspected choroidal naevi. Bull Soc Belge Ophtalmol. 1993;24829- 35
PubMed
13.
Mims  JLShields  JA Follow-up studies of suspicious choroidal nevi. Ophthalmology. 1978;85929- 943
PubMedArticle
14.
Butler  PChar  DHZarbin  MKroll  S Natural history of indeterminate pigmented choroidal tumors. Ophthalmology. 1994;101710- 716
PubMedArticle
15.
Augsburger  JJMcCarthy  EF  JrGonder  JRShields  JA Macular choroidal nevi. Int Ophthalmol Clin. 1981;2199- 106
PubMedArticle
16.
Naumann  GOHHellner  KNaumann  LR Pigmented nevi of the choroid: clinical study of secondary changesin the overlying tissues. Trans Am Acad Ophthalmol Otolaryngol. 1971;75110- 123
17.
Tamler  E A clinical study of choroidal nevi: a follow-up report. Arch Ophthalmol. 1970;84:29- 32.
PubMedArticle
18.
Mitchell  PSmith  WAttebo  KWang  JJ Prevalence of age-related maculopathy in Australia: the Blue MountainsEye Study. Ophthalmology. 1995;102:1450- 1460
PubMedArticle
19.
Mitchell  PWang  JJForan  S Five-year incidence of age-related maculopathy in Australia: the BlueMountains Eye Study. Ophthalmology. 2002;1091092- 1097
PubMedArticle
20.
Klein  RDavis  MDMagli  YLSegal  PKlein  BEHubbard  L The Wisconsin age-related maculopathy grading system. Ophthalmology. 1991;981128- 1134
PubMedArticle
21.
Albert  DMLahav  MPacker  SYimoyines  D Histogenesis of malignant melanomas of the uvea: occurrence of nevus-likestructures in experimental choroidal tumors. Arch Ophthalmol. 1974;92318- 323
PubMedArticle
22.
Albert  DMGaasterland  DECaldwell  JBHoward  ROZimmermann  LE Bilateral metastatic choroidal melanoma, nevi, and cavernous degeneration:involvement of the optic nervehead. Arch Ophthalmol. 1972;8739- 47
PubMedArticle
23.
Smolin  G Malignant change of a benign melanoma: report of a case. Am J Ophthalmol. 1966;61174- 177
PubMed
24.
Augsburger  JJSchroeder  RPTerrito  CGamel  JWShields  JA Clinical parameters predictive of enlargement of melanocytic choroidallesions. Br J Ophthalmol. 1989;73911- 917
PubMedArticle
25.
Sahel  JAAlbert  DM Choroidal nevi. Ryan  SJ,ed.Retina. St Louis, Mo CV Mosby Inc2001;650- 663
26.
Shammas  HFBlodi  FC Prognostic factors in choroidal and ciliary body melanomas. Arch Ophthalmol. 1977;9563- 69
PubMedArticle
27.
McLean  MJFoster  WDZimmerman  LE Prognostic factors in small malignant melanomas of choroid and ciliarybody. Arch Ophthalmol. 1977;9548- 58
PubMedArticle
Epidemiology
January 2004

Absence of Change in Choroidal Nevi Across 5 Years in an Older Population

Author Affiliations

From the Department of Ophthalmology, University of Sydney, and theWestmead Millennium Institute Centre for Vision Research, Westmead Hospital,Westmead, Australia. The authors have no financial interest in this article.

Arch Ophthalmol. 2004;122(1):89-93. doi:10.1001/archopht.122.1.89
Abstract

Objective  To determine the proportion of choroidal nevi that were previously identifiedin a population cross section and that showed evidence of growth or progressionduring a 5-year period.

Methods  The Blue Mountains Eye Study was a cohort study of residents 49 yearsand older living in an area west of Sydney, Australia. Retinal photographswere used to identify choroidal nevi. Repeat photographs were obtained 5 yearslater and graded side-by-side to ascertain clinical growth or progressionof all identified nevi. The greatest diameter and surface area of each nevuswere measured. Nevus growth was defined as an increase in size of at least33%.

Results  There were 160 choroidal nevi identified in the 128 subjects with neviwho participated in both eye examinations. Only 1 nevus (0.6%) exhibited clinicalgrowth during the 5 years. No nevi developed other indicators of progression,such as subretinal fluid or orange pigment accumulation.

Conclusions  Findings from this study indicate that benign nevi in older personsrarely progress. Regular eye examinations may be unnecessary for clearly definedsmall nonsuspicious choroidal nevi. This information could relieve patientanxiety and reduce costs associated with regular monitoring of nevi.

CHOROIDAL NEVI ARE A RELAtively common finding, with reported prevalencerates ranging between 1% and 30%.18 Althoughmost are generally considered benign, the possibility that some nevi may transforminto malignant melanomas has led many ophthalmologists to recommend regularsurveillance of choroidal nevi.911

A number of investigators have reported on the proportions and characteristicsof nevi that across time were confirmed as malignant melanomas, but thesenevi were generally regarded as suspicious in that it was difficult to differentiatethem from melanoma at initial examination.1215 Fewpopulation-based studies, however, have been performed to characterize andquantify the proportion of initially nonsuspicious nevi that show evidenceof malignant transformation. Naumann et al16 followed124 choroidal nevi for 2 years, none of which grew during that time. Tamler17 showed that in 28 patients, no nevi progressed during9.5 years.

Findings in the Blue Mountains Eye Study1 indicatedthat choroidal nevi were present in 6.5% of this older, largely white population.The purpose of the current article is to quantify the number of these nevithat progressed during 5 years. If the growth rate of clearly defined benignnevi was substantially lower than that of suspicious nevi, as suggested bysome study results, there may be no need to recommend surveillance of theselesions, with the benefits of reducing patient anxiety and costs.

METHODS

The Blue Mountains Eye Study18 is a population-basedstudy involving residents of an area west of Sydney, Australia. Noninstitutionalizedpermanent residents 49 years and older were eligible to participate. The firstexaminations, including a detailed medical and ocular history and eye examination,were performed from 1992 through 1993, with a participation rate of 82.4%.After 5 years, of the 3654 initial participants, 543 persons had died (14.9%),and 2335 (63.9%) underwent a second examination from 1997 through 1999. Detailsof the study methods have been described elsewhere.18,19

A fundus camera (Zeiss FF3; Carl Zeiss, Oberkochen, Germany) was usedat both eye examinations to obtain 30° stereoscopic retinal photographsof the macula and optic disc. Nonstereoscopic photographs were obtained inthe upper and lower temporal arcades and nasal retina, with a total fieldof approximately 70°.19 A nonsuspiciousnevus was defined as an unequivocal pigmented choroidal lesion with an elevationup to 1 mm and was slate blue. Partially depigmented and amelanotic nevi werealso included. Other pigmented lesions such as pigment clumps and pigmentedscars were excluded. No melanomas or melanocytomas were identified in thispopulation.

Subjects with a nevus identified in the first study1 formedthe population of interest. Slide transparencies from both examinations wereviewed with a fluorescent viewing box and a Donaldson stereo viewer (totalmagnification, ×15). Nevi were assessed for changes in size, shape,or color; an increase in number of overlying drusen, pigment clumping, orsubretinal fluid accumulation; and the presence of orange pigment. Clear plasticgrids containing 3 concentric circles with radii of 500, 1500, and 3000 µm,initially developed for grading age-related maculopathy,19,20 wereplaced over the nevus baseline and follow-up photographs to enable side-by-sidegrading of clinical change in size and other features. In addition, the photographswere digitally scanned, and the surface area and greatest diameter of eachnevus at baseline and follow-up were measured (Scion Image Beta 4.02; ScionCorp, Frederick, Md). Nevus growth was defined as an increase in size of atleast 33% in the follow-up photograph. All measurements were obtained by 1grader (S.T.). A random selection of 50 nevi were regraded and tested forintragrader reliability.

Descriptive analyses of the nevi and scatterplots of baseline vs follow-upmeasurements, including linear regression lines, were obtained (StatisticsPackage for the Social Sciences version 10.0; SPSS Inc, Chicago, Ill). Confidenceintervals (CIs) for the proportion of nevi showing growth were calculatedby using the exact method (EXACTPCI, a macro function in Statistical AnalysisSystem, version 6.2; SAS Institute Inc, Cary, NC). Intragrader reliabilitywas measured by using Pearson correlation analysis (R2; SPSS Inc) and quadratic weighted κ statistic(Stata, release 6; Stata Corp, College Station, Tex). The level of agreementbetween baseline and follow-up nevus diameter measurements was also assessedwith a Bland-Altman plot (Analyse-it, version 1.65; Analyse-It Software Ltd,Leeds, England).

RESULTS

Of the 232 subjects with nevi identified in the first study, 157 returnedfor the 5-year examinations. Of participants who were not reexamined, 27 haddied (11.6%), 20 refused to participate (8.6%), and 28 (12.1%) had eithermoved from the area or were lost to follow-up. The mortality rate in subjectswith nevi (11.6%) was comparable with that in the whole cohort (14.9%), andno deaths were attributed to choroidal melanoma or other melanoma. Of thosewho returned, 29 (18.5%) did not have adequate 5-year photographs of theirnevi, leaving 128 subjects with gradable photographs from both examinations.

This study subset of 128 subjects had a mean ± SD age of 64.5± 7.9 years (range, 49-83 years), and 61.7% were women. A total of160 nevi was found in the 128 subjects. Nevi were found in the right eye in54 subjects, the left eye in 63 subjects, and both eyes in 11 subjects. Twoor more nevi were found in 27 subjects (21.1%), with 1 participant having4 nevi. The mean ± SD nevus diameter was 1.28 ± 0.55 mm (range,0.34-4.04 mm), and the mean ± SD surface area was 1.09 ± 1.17mm2 (range, 0.07-9.49 mm2). Intragrader reliability for these measurements was high for both nevus diameter(κ = 0.97, R2 =0.99) and surface area (κ = 0.97, R2 = 0.99).

Table 1 outlines the baselinefeatures of nevi in our study. Most nevi were oval (43.8%), had a distinctedge (46.3%), were slate blue (90.0%), and were located in the posterior pole(68.8%), with almost two thirds (60.6%) located within the macula. Most werefound in the temporal region: 42.5% in the upper and 43.8% in the lower temporalregion. Drusen were found overlying 76.9% of the nevi. Pigment clumping andsubretinal fluid were each associated with 2 nevi. None of the nevi studiedhad overlying orange pigment. Characteristics of nevi identified at baselinethat were not reexamined at 5 years, including nevus location, color, shape,and other features, were similar to those that were reexamined.

One nevus (0.6%; 95% CI, 0.01%-3.4%) was clinically noted to have increasedin size, with a diameter of 0.49 mm and a surface area of 0.11 mm2 at the initial examination and a diameter of 0.73mm and a surface area of 0.29 mm2 at follow-up(Figure 1). No other nevi exhibitedgrowth during this period. There was no evidence of orange pigment depositionin any nevi at follow-up. The appearance of 2 nevi that had associated subretinalfluid at the baseline examination was unchanged at the 5-year examination.No new nevi or melanomas were identified.

Figure 2 shows a large nevusthat was unchanged during follow-up. Figure3A and B shows scatterplots of baseline vs follow-upmeasurements of the largest diameter and surface area of each nevus. The linearregression line generated for each of these plots has a slope close to 1.0,which suggests that there was little change in size during the 5 years. Figure 3C displays a Bland-Altman plot ofnevus diameter at baseline and follow-up. It demonstrates the level of agreementbetween the 2 sets of measurements by plotting the difference in nevus sizebetween the examinations against the mean of the 2 values. A similar numberof nevi appeared to have either grown slightly or shrunk slightly during the5 years, which is likely to represent measurement noise; however, on average,there appears to have been little change in nevus size during the 5 years(bias from zero = +0.03 mm; 95% CI, 0.009-0.05).

COMMENT

Choroidal nevi are generally considered a benign finding. A small proportionof nevi may develop into malignant melanomas, which is similar to findingsin cutaneous nevi. Evidence supporting this theory has been obtained froma number of histological reports demonstrating features of nevi in establishedmelanomas.21 Yanoff and Zimmerman9 foundthat 73 of 100 malignant melanomas had benign nevus cells at the peripheryand at the scleral edge. Albert et al22 alsodescribed nevuslike cell morphology at the base of multiple metastatic choroidalmelanomas in 1 patient. Authors of a number of case series have also reportedpossible transformation of a nevus into a malignant melanoma.7,8,10,23 Mostof these authors, however, agree that transformation is rare, given that choroidalnevi are much more frequent than are choroidal melanomas.4

In our study of 160 benign choroidal nevi, only 1 was judged to havegrown during a 5-year interval. The clinical importance of this growth isuncertain given the initial small size of the nevus. Our results support thefindings of Naumann et al16 who followed 124choroidal nevi in 112 patients for an average of 2 years and found that nonegrew during that time. After follow-up for 5 years in patients referred toan oncology service, Augsburger et al,24 foundthat 4.5% of 62 benign nevi grew. Tamler17 alsoobserved 28 patients with nevi (most were between one fourth and 2 disc diameters)for 9.5 years and found that none grew. Results of another study11 witha smaller number of subjects also support this finding.

The current recommended management for all nevi is identical to thatfor small melanomas25 and consists of annualfollow-up eye examinations. This approach is recommended because (in additionto potential malignant transformation), in some instances, it may be difficultto differentiate a small melanoma from a large nevus. Sahel and Albert25 described features of these suspicious nevi, includingthickness greater than 3 mm, presence of overlying orange pigment, retinaldetachment without choroidal neovascularization, and pinpoint leaks at fluoresceinangiography. Mims and Shields13 addressed thisissue by describing suspicious nevi as being 1 to 2 mm thick and having alargest diameter between 2 and 5 disc diameters, and they also described effectson overlying structures, especially orange pigment. Authors of a number ofarticles focus on these lesions and features that might predict progression.12,14,24

None of these features was present in the nevi followed in our study.The low rate of growth of the nevi in our population suggests that nevi clearlydefined as benign on the basis of size (<3-mm diameter and <1 mm thick),with no other suspicious features, may not need any regular surveillance.Patients with small nevi could be reassured that progression is rare. Clinicianscan also assume that although a large number of benign nevi go undetected,it is unlikely that these lesions will produce symptoms or be associated withincreased mortality. In the rare case of progression, evidence suggests thatthe mortality rate from small melanomas is low.26,27

A potential limitation of our study is that neither fluorescein angiographynor ultrasonography was performed to further differentiate these lesions frommelanomas. A strength of our study is that our findings are likely to be morerepresentative of the general older population, as compared with findingsin nevi studies based on referral practices. We were also able to measurenevus size by using multiple techniques to increase accuracy.

In summary, our study of nonsuspicious choroidal nevi in a population-basedcohort of older persons showed that fewer than 1% exhibited growth acrossa 5-year period. These findings suggest that small nonsuspicious nevi rarelyprogress and may not require routine surveillance.

Back to top
Article Information

Corresponding author: Paul Mitchell, MD, PhD, FRANZCO, Centre forVision Research, Dept of Ophthalmology, University of Sydney, Westmead Hospital,Hawkesbury Rd, Westmead, New South Wales, Australia 2145 (e-mail: paul_mitchell@wmi.usyd.edu.au).

Submitted for publication July 2, 2002; final revision received July25, 2003; accepted August 15, 2003.

This study was supported by grant 974159 from the National Health andMedical Research Council, Canberra, Australia.

References
1.
Sumich  PMitchell  PWang  JJ Choroidal nevi in a white population: the Blue Mountains Eye Study. Arch Ophthalmol. 1998;116645- 650
PubMedArticle
2.
Gass  JD Problems in the differential diagnosis of choroidal nevi and malignantmelanoma: XXXIII Edward Jackson Memorial lecture. Trans Am Acad Ophthalmol Otolaryngol. 1977;8319- 48
PubMed
3.
Naumann  GYanoff  MZimmerman  LE Histogenesis of malignant melanomas of the uvea: histopathologic characteristicsof nevi of the choroid and ciliary body. Arch Ophthalmol. 1966;76 (pt 1) 784- 796
PubMedArticle
4.
Hale  PNAllen  RAStraatsma  BR Benign melanomas (nevi) of the choroid and ciliary body. Arch Ophthalmol. 1965;74532- 538
PubMedArticle
5.
Rodriguez-Sains  RS Ocular findings in patients with dysplastic nevus syndrome. Ophthalmology. 1986;93661- 665
PubMedArticle
6.
Smith  REGanley  JP Ophthalmic survey of a community: abnormalities of the ocular fundus. Am J Ophthalmol. 1972;74 ((pt 1)) 1126- 1130
PubMed
7.
Albers  EC Benign melanomas of the choroid and their malignant transformation. Am J Ophthalmol. 1940;23779- 783
8.
Wagener  HPWellbrock  WLA Benign melanoma and melano-epithelioma of the choroid. Arch Ophthalmol. 1930;4509- 514Article
9.
Yanoff  MZimmerman  LE Histogenesis of malignant melanomas of the uvea: relationship of uvealnevi to malignant melanomas. Cancer. 1967;20 ((pt 2)) 493- 507
PubMedArticle
10.
MacIlwaine  WAAnderson  B  JrKlintworth  GK Enlargement of a histologically documented choroidal nevus. Am J Ophthalmol. 1979;87480- 486
PubMed
11.
Ganley  JPComstock  GW Benign nevi and malignant melanomas of the choroid. Am J Ophthalmol. 1973;7619- 25.
PubMed
12.
Sallet  Gde Laey  JJ Follow-up of suspected choroidal naevi. Bull Soc Belge Ophtalmol. 1993;24829- 35
PubMed
13.
Mims  JLShields  JA Follow-up studies of suspicious choroidal nevi. Ophthalmology. 1978;85929- 943
PubMedArticle
14.
Butler  PChar  DHZarbin  MKroll  S Natural history of indeterminate pigmented choroidal tumors. Ophthalmology. 1994;101710- 716
PubMedArticle
15.
Augsburger  JJMcCarthy  EF  JrGonder  JRShields  JA Macular choroidal nevi. Int Ophthalmol Clin. 1981;2199- 106
PubMedArticle
16.
Naumann  GOHHellner  KNaumann  LR Pigmented nevi of the choroid: clinical study of secondary changesin the overlying tissues. Trans Am Acad Ophthalmol Otolaryngol. 1971;75110- 123
17.
Tamler  E A clinical study of choroidal nevi: a follow-up report. Arch Ophthalmol. 1970;84:29- 32.
PubMedArticle
18.
Mitchell  PSmith  WAttebo  KWang  JJ Prevalence of age-related maculopathy in Australia: the Blue MountainsEye Study. Ophthalmology. 1995;102:1450- 1460
PubMedArticle
19.
Mitchell  PWang  JJForan  S Five-year incidence of age-related maculopathy in Australia: the BlueMountains Eye Study. Ophthalmology. 2002;1091092- 1097
PubMedArticle
20.
Klein  RDavis  MDMagli  YLSegal  PKlein  BEHubbard  L The Wisconsin age-related maculopathy grading system. Ophthalmology. 1991;981128- 1134
PubMedArticle
21.
Albert  DMLahav  MPacker  SYimoyines  D Histogenesis of malignant melanomas of the uvea: occurrence of nevus-likestructures in experimental choroidal tumors. Arch Ophthalmol. 1974;92318- 323
PubMedArticle
22.
Albert  DMGaasterland  DECaldwell  JBHoward  ROZimmermann  LE Bilateral metastatic choroidal melanoma, nevi, and cavernous degeneration:involvement of the optic nervehead. Arch Ophthalmol. 1972;8739- 47
PubMedArticle
23.
Smolin  G Malignant change of a benign melanoma: report of a case. Am J Ophthalmol. 1966;61174- 177
PubMed
24.
Augsburger  JJSchroeder  RPTerrito  CGamel  JWShields  JA Clinical parameters predictive of enlargement of melanocytic choroidallesions. Br J Ophthalmol. 1989;73911- 917
PubMedArticle
25.
Sahel  JAAlbert  DM Choroidal nevi. Ryan  SJ,ed.Retina. St Louis, Mo CV Mosby Inc2001;650- 663
26.
Shammas  HFBlodi  FC Prognostic factors in choroidal and ciliary body melanomas. Arch Ophthalmol. 1977;9563- 69
PubMedArticle
27.
McLean  MJFoster  WDZimmerman  LE Prognostic factors in small malignant melanomas of choroid and ciliarybody. Arch Ophthalmol. 1977;9548- 58
PubMedArticle
×