Trends in Size and Treatment of Recently Diagnosed Choroidal Melanoma, 1987-1997: Findings From Patients Examined at Collaborative Ocular Melanoma Study (COMS) Centers: COMS Report No. 20

Objective  To describe time trends, from 1987 through 1997, (1) in size of choroidal melanoma among patients with recent diagnosis confirmed at a clinical center that participated in the Collaborative Ocular Melanoma Study (COMS) and (2) in choice of treatment by patients who did not enroll in one of the COMS randomized trials.

Methods  Investigators at all COMS clinical centers (41 in the United States and 2 in Canada) agreed to report, in a masked fashion that did not include personal identifiers, all patients diagnosed as having choroidal melanoma during the accrual period for COMS randomized trials of radiotherapy. Information reported for patients who did not enroll in a COMS randomized trial included tumor dimensions, date of diagnosis, and initial treatment selected. Patients reported by centers that continued to report cases until 1997 and diagnosed as having choroidal melanoma no more than 1 year before evaluation at a participating COMS center contributed the data analyzed.

Outcome Measures  Time trends in tumor size among patients reported and in elective treatment of patients not enrolled in COMS randomized trials.

Results  Of 8712 patients with choroidal melanoma examined, 6703 met criteria for analysis of time trend in tumor size and 4077 were analyzed for treatment trends over time. The number of cases with longest tumor basal diameter greater than 15.0 mm declined over time from 393 (30%) of 1330 cases reported in 1987 through 1989 to 345 (25%) of 1397 cases reported in 1996 or 1997. The proportion of patients eligible for COMS randomized trials who did not enroll and who elected enucleation remained stable over time for tumors of all sizes; the proportion of these patients who elected eye-conserving radiotherapy increased over time. Juxtapapillary tumors accounted for nearly half of the enucleations among ineligible patients who had tumors no larger than 15.0 mm in longest basal diameter.

Conclusions  Among patients examined at COMS centers during 1987 through 1997, the trends observed for patients with recently diagnosed choroidal melanoma included increasing presentation with tumors of smaller size, decreasing use of enucleation for choroidal melanoma 15.0 mm or less in longest basal diameter, and increasing use of such strategies for larger tumors.

CHOROIDAL MELANOMA is the most common primary intraocular cancer in adults.¹ Nevertheless, only about 6 cases per million population are believed to be diagnosed annually in the United States.^2,3 Because this cancer is often asymptomatic unless or until central vision is affected, diagnosis frequently is made after a routine eye examination. Until about 1970, enucleation was the standard treatment. Beginning in the early 1970s, radiotherapy and other treatment approaches with the goal of preserving the eye became increasingly popular, particularly for smaller choroidal melanomas. In addition, diagnostic aids and skills improved so that misdiagnosis rates declined, as summarized elsewhere.⁴

In the United States, from the early 1970s until the mid 1980s, many patients with suspected choroidal melanoma were referred to a few centers that specialized in the care of patients with ocular cancer, particularly when the tumor was of a size that nonenucleation treatment was judged to be an option. Thus, reports regarding the size of these tumors depended on the source of information, with local ophthalmologists typically treating larger choroidal melanoma with enucleation and large referral centers using eye-conserving methods to treat many smaller tumors. Consequently, information regarding tumor size at the time of initial diagnosis and treatment, whether based on cases seen at individual referral centers^5-7 or on cases treated with enucleation and registered centrally, ⁸ may be biased or outdated.

In 1986, ophthalmologists participating in the Collaborative Ocular Melanoma Study (COMS) began to solicit referral of patients with choroidal melanoma of all sizes to be screened for eligibility for randomized trials(large and medium-sized tumors) and for a nonrandomized prospective study of small choroidal melanomas. The goal of each randomized trial was to compare survival after enucleation alone with survival after radiotherapy, delivered either as an adjunct to enucleation for large choroidal melanomas⁹ or as an alternative to enucleation for medium-sized choroidal melanomas.¹⁰ Before initiation of the COMS, terms referring to tumor size had been used in various ways by researchers and clinicians.¹¹ Thus, it was difficult to predict with confidence the number of patients who would be diagnosed as having choroidal melanomas of eligible sizes for the randomized trials. In an effort to determine the feasibility of achieving the target sample sizes for COMS randomized trials and to aid in the design of future randomized trials of treatments for choroidal melanoma, participating ophthalmologists agreed to report all patients diagnosed to have this cancer who were evaluated during the period of patient accrual to the trials. Although only limited information was collected centrally for patients who did not enroll in COMS trials, tumor dimensions and initial management were among the data reported. The highly accurate diagnostic ability of COMS ophthalmologists was documented among patients assigned to enucleation in the randomized trials: all patients enrolled in the trial of preenucleation radiation and half the patients who enrolled in the trial of iodine 125 brachytherapy.⁴

The purpose of this article is to report trends over time in the size of recently diagnosed choroidal melanomas during the COMS reporting period and, for patients who did not enroll in COMS randomized trials, trends in treatment choices by tumor size.

Descriptions of many aspects of the COMS design and methods have been published.^4,12-17 The COMS Manual of Procedures¹⁸ and the COMS Study Forms Book¹⁹ are available. The design and methods of the COMS were reviewed and approved by an independent Data and Safety Monitoring Committee in August 1986 before patient screening began. The study was reviewed and approved by the local institutional review board of each participating center before examination of patients for COMS clinical trials and data reporting were initiated at that center.

Patients with suspected choroidal melanoma were examined at 1 of 43 COMS clinical centers for confirmation of the diagnosis and evaluation of eligibility for one of the COMS randomized trials. Clinical centers initiated reporting of cases after their participation was approved by the COMS leadership, by the National Eye Institute, which sponsored the study, and by the local institutional review board and after all necessary local preparations required to implement the COMS protocol had been completed. All patients for whom the diagnosis of choroidal melanoma was confirmed clinically by a COMS ophthalmologist during the enrollment period were reported to the COMS Coordinating Center(Baltimore, Md), but were identified to the Coordinating Center only by study numbers and code names assigned at the local clinical centers. Each clinical center maintained a log of all patients reported to the COMS. Every 6 months, a list of patients reported to the Coordinating Center was sent for review by COMS personnel at each clinical center and confirmation that all new cases of choroidal melanoma examined at the center by COMS personnel had been reported. To avoid duplicate counting of patients, twice-yearly checks of the database were performed to identify patients who had the same birth date, sex, and other sociodemographic information. Apparent matches were referred to the reporting centers for confirmation. Only one set of records was retained in the database for each patient. On the basis of criteria for eligibility for COMS trials, each patient reported was classified as eligible or ineligible for enrollment and random treatment assignment.

As part of the initial ophthalmic evaluation, standardized A-scan and contact B-scan echography was performed.¹⁶ The ophthalmologist provided 3 dimensions of the choroidal melanoma based on examination findings that included indirect ophthalmoscopy and transillumination: height at the tumor apex, longest diameter of the tumor base, and diameter of the tumor base perpendicular to the longest diameter. Typically, apical height was based on measurements from A-scan echograms and examination findings.

Dimensions of a standard eye were provided in the COMS Manual of Procedures¹⁸ for reference when assessing basal tumor dimensions. Tumor dimensions were recorded on standard forms¹⁹ that were sent to the Coordinating Center for data entry and integration with other study data. As of March 1994, reporting of the diameter of the tumor base perpendicular to the longest diameter was discontinued for patients who did not enroll in one of the randomized trials.

For patients judged eligible who enrolled in a COMS randomized trial, treatment of the choroidal melanoma was assigned randomly at time of enrollment. The examining ophthalmologist reported the initial treatment, when known, for patients who did not enroll in one of the trials, either because they were ineligible or because, although eligible, they elected not to enroll. Treatment decisions of patients not enrolled were made independently of treatment assignments possible in COMS randomized trials. Treatment may have been implemented at a COMS center or elsewhere, depending on patient preferences. Reporting of initial management decisions was initiated in 1986 for patients eligible for the COMS randomized trials who did not enroll and in 1989 for patients ineligible for COMS trials.

Data available as of December 31, 2001, were used for this analysis. All data from patient evaluations had been entered into the database, final checks for duplicate reporting of patients had been completed, and information from initial evaluations had been checked against clinical center records for most patients.

Analysis was restricted to the period January 1, 1987, through December 31, 1997, in an effort to minimize the effect on patient referral and treatment that could have resulted from publication of mortality rates by treatment arm from the COMS randomized trial for large choroidal melanoma in June 1998.⁹ This 11-year period was divided into 5 time intervals, during each of which 1300 to 1400 recently diagnosed patients were reported: January 1987 through December 1989, January 1990 through December 1991, January 1992 through December 1993, January 1994 through December 1995, and January 1996 through December 1997. Data from centers that halted reporting before 1997 were omitted from analysis. Only patients whose reported date of diagnosis of choroidal melanoma was no more than 1 year before evaluation at a COMS clinical center were included.

Longest tumor basal diameter, apical height of the tumor, and basal diameter perpendicular to the longest basal diameter were strongly correlated, with Pearson correlation coefficients²⁰ of 0.87 (P<.001) and 0.64 (P<.001) for perpendicular tumor diameter and apical height, respectively, with longest tumor basal diameter. Because of these strong correlations, because longest basal tumor diameter was one of the most important prognostic variables for survival, ^9,10 and because treatment decisions are often guided in part by longest basal diameter, that dimension was used to classify tumor size for this analysis.

Observed time trends in the proportion of tumors in a given size category have been evaluated by the χ² test for trend in proportions among quantitatively ordered categories.²¹ Similarly, trends in the proportion of tumors within each size category that were managed initially in a specified manner were confirmed by means of the same test. SAS software (SAS Institute Inc, Cary, NC) was used for data analysis.

During the entire reporting period, ie, November 1986 through July 1998, 8712 patients diagnosed as having choroidal melanoma were reported by the 43 participating clinical centers. Six centers halted reporting before 1997; the 801 patients reported by those centers have been excluded from this analysis. Table 1 lists the 30 cities in which the 37 centers were located and the date that the first patient was reported for those who reported the patients included in this analysis. Of the 7911 patients reported by these 37 centers, 771 were reported to have had choroidal melanoma diagnosed more than 1 year before evaluation at a COMS center; the date of diagnosis was not reported for 5 other patients who also were excluded. Remaining patients reported before January 1987 (n = 1) or after December 1997 (n = 360) also were excluded from analysis. Finally, patients for whom longest basal diameter was not available (n = 71) were excluded. Thus, analysis of tumor size is based on data from 6703 patients (Figure 1).

As shown in Figure 2, the numbers of cases reported in each 2-year interval from 1990 through 1997 remained nearly constant; a mean of 672 recently diagnosed cases were reported per year during this 8-year period. During the entire 11-year period, the number of cases with longest basal diameter greater than 15.0 mm declined (P = .002).

Of the 6703 patients included in analysis of tumor size, 2032 had treatment assigned randomly at time of enrollment in one of the COMS clinical trials, leaving 4671 patients not enrolled for whom treatment was chosen by the patient in consultation with physicians (Figure 1). Treatment selection was reported from 1987 through 1997 for 1727 patients who were eligible for COMS randomized trials but who did not enroll. Because reporting of treatment elected by patients ineligible for the randomized trials did not begin until 1989, 564 ineligible patients reported before 1990 were excluded from analysis of treatment trends; 30 more ineligible patients were excluded because they had been treated before evaluation at a COMS center. Thus, 2350 ineligible patients were available for assessing 8-year trends in choice of treatment for choroidal melanoma, 1990 through 1997.

Treatments elected for patients who were eligible for COMS trials but who did not enroll are summarized in Table 2 by tumor size and time. The proportion of patients who elected enucleation or eye-conserving radiotherapy varied by tumor size and reporting period. For patients whose tumors were 15.0 mm or less in longest basal diameter, there was no time trend in the proportion who elected enucleation (P = .15 for those 12.0 mm or less in diameter and P = .61 for those 12.1 to 15.0 mm in diameter).

For patients whose tumors were greater than 15.0 mm in longest basal diameter, the proportion who elected enucleation increased over time (P = .005). The proportion of patients who elected eye-conserving radiotherapy increased over time for those whose tumors were no greater than 12.0 mm in longest basal diameter (P = .02) and decreased over time for those whose tumors were greater than 15.0 mm in diameter (P = .002). No time trend in choice of eye-conserving radiotherapy was observed for eligible patients who did not enroll who had tumors of intermediate sizes.

Initial treatments elected by patients ineligible for COMS randomized trials, reported from 1990 through 1997, are summarized in Table 3. Of the 144 ineligible patients with tumors no more than 9.0 mm in longest basal diameter who were treated with enucleation, 72 (50%) had juxtapapillary tumors. Similarly, 125 (52%) of the 242 patients and 84(41%) of 204 patients in the next 2 size categories who were treated with enucleation had juxtapapillary tumors. Only 28 (14%) of 203 patients and 4(3%) of 119 patients in the 2 largest tumor size categories whose eyes were enucleated had juxtapapillary tumors.

Observation was chosen as the initial management strategy for a substantial number of ineligible patients with tumors 9.0 mm or smaller in longest basal diameter: 110 (57%) of 194 ineligible patients reported in 1990 or 1991 and 103 (47%) of 217 ineligible patients reported in 1996 or 1997. In other tumor size categories, the proportion of patients who elected observation was smaller. Reasons for ineligibility for the COMS randomized trials were reported elsewhere^13,22; metastatic melanoma, other primary cancers, and life-threatening coexisting conditions accounted for the majority of ineligible patients and for nearly all patients who were observed after diagnosis of choroidal melanoma at a COMS center.

Because of the multicenter design of the COMS and the commitment of COMS personnel at participating centers to report all patients examined with choroidal melanoma during the period of accrual to COMS randomized trials, the COMS Group is able to provide valuable data regarding recent time trends in the size and management of recently diagnosed tumors of this type. Prospective reporting of cases and standard, prospective data collection strengthen the findings. Because accrual to COMS trials and reporting of cases halted early at some centers, data from those centers were eliminated to avoid potential biases. For patients eligible for COMS trials who did not enroll, trends observed during an 11-year period included increasing use of eye-conserving radiotherapy modalities and decreasing use of enucleation for smaller tumors. For patients whose tumors were in the largest size categories, the observed trend was for increasing use of enucleation over time. Delayed implementation of reporting of treatment elections for patients ineligible for COMS trials restricted the length of the period during which treatment choices could be analyzed for this subgroup of patients with choroidal melanomas evaluated at COMS centers. Although there were variations in treatment choices by ineligible patients from one period to another, there were no clear trends during the 8-year period for which information was available. Factors other than tumor size and location undoubtedly influenced treatment choices by these patients.

Some of the observed changes in distributions of tumor sizes over time may have been related to events in the COMS. Accrual of patients with small tumors for an observational study halted at most centers in 1990. However, as shown in Figure 2, the proportion of tumors in the smallest size category during 1990 and 1991 was almost identical to the proportion of that size in 1987 through 1989. The decrease in the proportion of tumors in this size category during 1992 through 1995 had disappeared by the 1996 to 1997 period. Accrual of patients to the COMS randomized trial of enucleation with and without preenucleation radiation ended in 1994. Referrals of patients with tumors of sizes appropriate for that trial may have declined thereafter. This possibility is consistent with the time trend in reporting of larger tumors observed in Figure 2.Nevertheless, the total number of patients recently diagnosed as having choroidal melanoma at COMS centers remained nearly constant over time. Thus, referral of patients with tumors of a size eligible for the trial of iodine 125 brachytherapy would have had to increase to offset decreases in referrals of tumors in the smallest and largest size categories.

Although it would have been ideal to have been able to identify all cases of choroidal melanoma diagnosed or treated at a participating institution, such case finding was not feasible with the available resources. Physicians who were participating in the COMS were conscientious about reporting all cases of choroidal melanoma with whom they were involved during diagnosis or treatment. However, some of them were aware that local colleagues treated patients with choroidal melanoma who were not examined and reported by COMS personnel. Nevertheless, the cases reported to the COMS are believed to be representative of all cases of choroidal melanoma diagnosed and treated in the United States and Canada during the 11-year period with respect to tumor size, ^13,22 but there are few other large collections of cases for comparison. Uveal melanoma accounted for 3846(85%) of the 4522 cases of ocular melanoma registered in the National Cancer Database and diagnosed between 1985 and 1994.²³ Cases in that database were compared with the COMS cases included in this analysis with respect to age, sex, race or ethnicity, and treatment given. Although cases with ocular melanoma reported to the National Cancer Database had sociodemographic characteristics that were similar to COMS cases of choroidal melanoma analyzed herein, initial treatment of the melanoma differed. Among cases in the National Cancer Database, 67.5% of 4522 were treated with enucleation, either alone or in combination with radiotherapy, chemotherapy, or both, compared with 44.7% of 4077 COMS cases of choroidal melanoma not enrolled in randomized trials. Conversely, 25.9% of cases in the National Cancer Database and 34.3% of COMS cases were treated with radiation alone or radiation and chemotherapy. Because of the overlapping periods covered by the 2 databases and because of reporting by multiple centers, it is possible that some cases are included in both databases. The National Cancer Database also may include some patients who were enrolled in COMS clinical trials for whom treatment was assigned randomly.

It is not possible to estimate the number of patients who were diagnosed as having choroidal melanoma and treated by local ophthalmologists without referral to a COMS center or to a large non-COMS ocular oncology referral center during the period when cases were evaluated for COMS eligibility. Furthermore, large non-COMS referral centers have not reported the sizes and methods of treating patients with choroidal melanoma by time. A report from one such center stated that approximately 2500 patients with choroidal melanoma were managed annually during 1995 through 1999, ²⁴ a period that overlaps the interval during which cases were reported by COMS centers. The annual average of 500 cases managed at that single center suggests that large referral centers that did not participate in the COMS managed at least as many cases of choroidal melanoma as were evaluated at COMS centers. Methods of treatment of patients with choroidal melanoma at those centers may have differed significantly from treatment choices by patients who were examined at COMS centers but did not enroll and from treatments received by patients included in the National Cancer Database.

With this analysis, the COMS Group has provided information about the size of recently diagnosed choroidal melanoma in the largest group of such patients for whom findings have been published. In addition, treatment choices during an 11-year period for eligible patients who did not enroll in COMS trials and during an 8-year period for ineligible patients also have been documented. Patients evaluated for the COMS and reported by the investigators, despite collection of a limited amount of data for each one, provide estimates of tumor size distributions and trends over time and document treatment choices made by both COMS-eligible and -ineligible patients examined at multiple study centers. Data available from multiple clinical centers in the United States and Canada may provide better estimates of the size distribution of choroidal melanoma at time of diagnosis and treatment than data from single centers or from central registries of cases undergoing enucleation. This information may be useful to researchers who design future clinical trials and epidemiologic studies of choroidal melanoma.

Corresponding author and reprints: Barbara S. Hawkins, PhD, COMS Coordinating Center, 550 N Broadway, Ninth Floor, Baltimore, MD 21205-2010.

Submitted for publication May 29, 2002; final revision received March 24, 2003; accepted April 3, 2003.

The COMS Group has received support from the National Eye Institute and the National Cancer Institute, National Institutes of Health, US Department of Health and Human Services, Bethesda, Md, through cooperative agreements EY 06253, EY 06257, EY 06258, EY 06260, EY 06264, EY 06265, EY 06266, EY 06268, EY 06269, EY 06270, EY 06274, EY 06275, EY 06276, EY 06279, EY 06280, EY 06282, EY 06283, EY 06284, EY 06287, EY 06288, EY 06289, EY 06291, EY 06839, EY 06843, EY 06844, EY 06848, EY 06858, and EY 06899.

Some of the material in this article was presented to The Macula Society; June 15, 2002; Barcelona, Spain.

The COMS investigators gratefully acknowledge the contributions of the many referring ophthalmologists.