A parallel study of quality of life of patients enrolled in the randomized trial of I-125 brachytherapy for medium choroidal melanoma was initiated in 1994.¹⁵ The purpose of the parallel study was to compare treatment arms over time with respect to general health, visual function, and anxiety and depression using scores from several standard interview instruments. The study had two components: (1) a prospective randomized component consisting of 209 patients who enrolled in the trial of I-125 brachytherapy and were interviewed prior to random treatment assignment, at 6 months following treatment, and on annual anniversaries of enrollment for up to 8 years; and (2) a cross-sectional component consisting of 645 additional patients who had enrolled in the randomized trial before the quality of life study was initiated and who were interviewed at least once during scheduled follow-up. These 854 patients represent 90% of patients who were eligible for the quality of life study and 65% of all patients who enrolled in the trial of I-125 brachytherapy.

Observational study

A nonrandomized observational study of small choroidal melanoma was included in the initial COMS design with the goal of providing sufficient information to design a randomized clinical trial for small tumors. Primary objectives were to estimate the number of patients with small choroidal melanoma available for inclusion in a randomized trial, the methods of treating small choroidal melanoma most widely used by COMS investigators, and, to the extent feasible within a small pilot study with short patient follow-up, rates of tumor growth and patient death.

Methods

The COMS design and many of the methods have been published; the COMS Manual of Procedures ¹¹ is available. Patients were evaluated for eligibility, enrolled, and treated at 43 different clinical centers, 41 in the USA and two in Canada. A standard schedule of clinical examinations was followed for data collection purposes. An unusual feature of the COMS design was that the participating ophthalmologists reported de-identified basic demographic information (age, gender, race, or ethnicity) and tumor dimensions for all cases of choroidal melanoma examined during the period of patient accrual, regardless of tumor size, eligibility for the COMS, or willingness of eligible patients to enroll in the COMS.

Random assignment to treatment, oversight of data collection, data management, and data analysis were assigned to the COMS Coordinating Center in Baltimore, Maryland. This center also had major responsibility for monitoring the quality of data provided by the participating centers. Other centers with major quality assurance and monitoring responsibilities included an Echography Center (in Mars Hill, North Carolina; initially in Miami, Florida), where tumor height was measured independently from photoechograms, a Photograph Reading Center (in Iowa City), where tumor characteristics at baseline and postirradiation changes in the posterior retina during follow-up were assessed and recorded, a Pathology Center (in Madison, Wisconsin; initially in Boston, Massachusetts), where tumor size and diagnosis were confirmed from all eyes enucleated, and the Radiological Physics Center in Houston, Texas, where adherence to the radiotherapy protocols was monitored. Overall leadership responsibility for the COMS Group was vested in the COMS Chairman’s Office (in Philadelphia, Pennsylvania; initially in Baltimore, Maryland).

An independent Data and Safety Monitoring Committee, appointed by the Director of the National Eye Institute, was the only group with access to survival data from the randomized clinical trials by treatment arm until this committee judged that the objectives of each individual trial had been met. This group had responsibility for ensuring that COMS trials were conducted in a scientifically valid and ethically sound manner. Scientific leadership of the COMS was provided by the Executive Committee, whose members included representatives of both the resource centers and the participating clinical centers. Three ophthalmic pathologists comprised the Pathology Review Committee who reviewed every enucleated eye from COMS patients to determine whether the clinical diagnosis of choroidal melanoma was correct. Mechanisms for quality assurance and monitoring were developed by the Quality Assurance Committee, which oversaw all aspects of data collection and protocol adherence. Classification of causes of death was the responsibility of the Mortality Coding Committee, whose members did not have responsibility for medical care of COMS patients. The COMS Archives Committee has responsibility for reviewing and approving applications for access to original COMS data by researchers.

Chronology of the COMS

Accrual of patients to the randomized trial of pre-enucleation radiation (PERT) of large choroidal melanoma began in November 1986 and ended in December 1994, with 1003 patients enrolled. Scheduled clinical follow-up of all surviving patients for vital status, incidence of metastasis and second cancers, and complications continued until July 31, 2000. Interim mortality findings and related information that emphasized 5-year outcomes were published in 1998 ^12,^16,¹⁷; mortality findings through 10 years and prognostic factors were published in 2004.¹⁸

Accrual of patients to the randomized trial of I-125 brachytherapy for medium choroidal melanoma began in January 1987 and ended in July 1998, with 1317 patients enrolled, at the recommendation of the Data and Safety Monitoring Committee. Scheduled clinical follow-up of all surviving patients, for clinical and vital status and for quality of life, continued until July 31, 2003, and October 31, 2003, respectively. Interim mortality findings were published in 2001 ¹³; mortality findings through 12 years after enrollment and subgroup findings were published in 2006.¹⁹ Information about complications²⁰^–²² and related information²³ also have been published.

Patient accrual to the nonrandomized study for small choroidal melanoma began in 1987 and ended in 1989, with 204 patients enrolled. Annual follow-up examinations were halted in 1991 as a result of funding constraints. Vital status and treatment status were reassessed in 1993 through 1994, and again in 1995 through 1996, for all patients who had not been lost to follow-up. Findings from this study have been published.^24,²⁵

The COMS database, copies of all publications from the COMS Group, and the “COMS Manual of Procedures” were deposited in the Alan Mason Chesney Medical Archives at the Johns Hopkins University in August 2008. An anonymized public use dataset, containing baseline characteristics and survival outcomes, the COMS Manual, and copies of all COMS publications are available by application to the Medical Archives (http://www.medicalarchives.jhmi.edu). Access to more extensive data, including images of original data forms received at the COMS Coordinating Center, is available to qualified researchers whose application to the COMS Archives Committee (via: schacha@ccf.org or bhawkins@jhmi.edu) is approved by the COMS Archives Committee and by the Medical Archives institutional review board.

Findings from the COMS trial of I-125 brachytherapy for medium choroidal melanoma

Participants

By July 1998, a total of 8712 patients with choroidal melanoma had been reported by COMS investigators; 5046 were classified to be of medium size by COMS criteria (Table 150.1). Among 2882 patients eligible for the randomized trial of I-125 brachytherapy versus standard enucleation, 1317 patients gave signed consent, enrolled, and were assigned randomly to treatment arm: 660 to standard enucleation and 657 to I-125 brachytherapy. Treatment arms were well balanced; adherence to the COMS protocol was excellent.¹³ All but 21 patients, seven in the brachytherapy arm and 14 in the enucleation arm, were treated promptly as assigned. Three patients assigned to brachytherapy crossed over to enucleation; seven enucleation patients crossed over to brachytherapy and two had proton beam radiation as the initial treatment. At the end of clinical follow-up in 2003, the vital status 5 years after enrollment was known for 1313 patients (99.7%), i.e., for all but four patients in the enucleation arm. Of 799 patients eligible for 10 years of follow-up, the vital status at 10 years was known for 791 (99.0%), i.e., for all but one patient in the brachytherapy arm and seven patients in the enucleation arm.

Survival estimates

By September 30, 2000, all patients had been followed for vital status for 2 years or longer, with 1274 patients followed for 3 years or longer and 1072 patients eligible for 5 years of follow-up. Among patients in the enucleation arm, 188 (28%) were known to have died compared with 176 (27%) of those in the brachytherapy arm. Estimated 5-year survival rates and 95% confidence intervals were 81% (95% CI: 77–84%) in the enucleation arm and 82% (95% CI: 79–85%) in the brachytherapy arm. Of the 364 decedents, 159 were judged to have had melanoma metastasis at time of death. Neither all-cause mortality rates nor rates of death with histopathologically confirmed melanoma metastasis differed between treatment arms. Adjustment of mortality rates for independent and statistically significant predictors of time to death (baseline age, tumor dimensions, tumor location, tumor shape, smoking history, and coexisting medical conditions) changed the estimated risk ratio from 0.93 (unadjusted; 95% CI: 0.76–1.14) to 0.99 (adjusted; 95% CI: 0.80–1.22).

Clinical follow-up of patients in the trial of I-125 brachytherapy ended on July 31, 2003, after all patients had been followed for at least 5 years up to a maximum of 15 years. Five-year survival rates were similar to those published in 2001, i.e., 81% in each treatment arm, yielding a pooled 95% CI of 79–83%. Ten-year survival rates also were the same in the two arms: 65% (95% CI: 62–68%). Mortality rates by treatment arm for all causes and with melanoma metastasis are shown in Figure 150.1. Older age and maximum basal tumor diameter were the primary predictors of earlier death.¹⁹ Pooled data were used to summarize mortality, diagnosis of melanoma metastasis,²⁶ and second primary cancers²⁷ through 12 years¹⁹ (Fig. 150.2).

Fig. 150.1 (Top) The cumulative percentage of patients in the COMS trial of I-125 brachytherapy for medium choroidal melanoma who had died by specified times after enrollment. Blue line: patients assigned to brachytherapy. Red line: patients assigned to enucleation. The numbers of patients at risk of death and numbers censored, based on date of enrollment, are given at annual anniversaries of enrollment by treatment assignment. Event is death from any cause. (Bottom) The cumulative percentage of patients in the COMS randomized trial of I-125 brachytherapy who had died with metastatic melanoma by the specified times after enrollment. Blue line: patients assigned to brachytherapy. Red line: patients assigned to enucleation. The numbers of patients at risk of death and numbers censored, based on date of enrollment, are given at annual anniversaries of enrollment by random treatment assignment. Event is death with histologically confirmed metastatic melanoma.

Fig. 150.2 The percentage of patients in the COMS randomized trial of I-125 brachytherapy with the specified status at the end of each year of follow-up, with treatment arms pooled. The number of patients who enrolled early enough to be followed up to the end of the specified year is shown at the top of the corresponding bar.

Complications

As reported in 2002,²⁰ 69 of the 650 patients whose eyes were treated with brachytherapy had the eye enucleated during the first 5 years after initial treatment, yielding a 5-year rate of 12% (95% CI: 10–16%); local treatment failure was reported for 57 eyes in the same time period, with a 5-year cumulative rate of 10% (95% CI: 8–13%). Local treatment failure accounted for 39 of the 69 enucleations. The 3-year cumulative rate of loss of six or more lines of visual acuity from baseline was 49% (95% CI: 44–53%) and the 3-year cumulative rate of loss of visual acuity to 20/200 or worse was 43% (95% CI: 38–48%).²¹ Five-year estimates of cataract among the 532 eyes treated with I-125 brachytherapy that were phakic at baseline and had no history of cataract in the eye, were 83% (95% CI: 79–87%).²²

Quality of life

In the parallel study of quality of life, 206 of 209 patients who enrolled and were treated as assigned (103 each to enucleation or to I-125 brachytherapy) provided 5-year findings regarding quality of life.²⁸ Patients treated with brachytherapy reported significantly better visual function than those treated with enucleation with respect to driving and peripheral vision for up to 2 years following treatment. Differences between treatment arms diminished thereafter, paralleling declining visual acuity in brachytherapy-treated eyes. Patients treated with brachytherapy reported more symptoms of anxiety through 5 years than those treated with enucleation.²⁸

Findings from the COMS trial of pre-enucleation radiation for large choroidal melanoma

Participants

From November 1986 through December 1994, COMS investigators reported 6078 patients with choroidal melanoma. Of these, 1860 had tumors classified as large by COMS criteria (see Table 150.1 for definitions used). Of those classified as large, 1302 were judged eligible for enrollment; 1003 gave signed consent and were enrolled. The two treatment arms were well-balanced with respect to most of the many characteristics of patients, eyes, and tumors considered.¹² Adherence to the COMS protocol was excellent, as was diagnostic accuracy.^12,²⁹ Only nine patients; three assigned to standard enucleation and six assigned to pre-enucleation radiation, were not treated as assigned at time of enrollment. When clinical follow-up ended in 2000, vital status 5 years after enrollment was known for 998 patients (99.5%), i.e., all but three patients in the enucleation alone arm and two patients in the pre-enucleation radiation arm.

Survival estimates

By July 31, 1997, the vital status at 3 years after enrollment was available for all but 26 patients who had enrolled during the last few months of patient accrual. The 5-year vital status was known for 801 (80%) of all patients enrolled; 238 patients assigned to enucleation alone (47%) and 219 patients assigned to pre-enucleation radiation (44%) were known to have died. The estimated 5-year cumulative survival rates and 95% confidence intervals were 57% (95% CI: 52–62%) for patients assigned to enucleation alone and 62% (95% CI: 57–66%) for patients assigned to pre-enucleation radiation. Neither 5-year survival rates nor survival rates over the first 8 years after enrollment differed between treatment arms, to either a statistically or clinically significant degree.¹²

Mortality by treatment arm and by time since enrollment when clinical follow-up ended in July 2000 is summarized in Figure 150.3. Five-year survival rates were identical to those reported earlier; 10-year survival was similar in the two arms and yielded a pooled estimate of 39% (95% CI: 35–42%).¹⁸ Among baseline characteristics of the patients, eyes, and tumors evaluated for their potential as prognostic factors, only patient age at time of treatment and longest tumor basal diameter had statistically significant effects on the length of overall survival.¹⁸

Fig. 150.3 (Top) The cumulative percentage of patients in the COMS trial of pre-enucleation radiation (PERT) for large choroidal melanoma who had died by specified times after enrollment. Blue line: patients assigned to PERT. Red line: patients assigned to enucleation alone. The numbers of patients at risk of death and numbers censored, based on date of enrollment, are given at annual anniversaries of enrollment by treatment assignment. Event is death from any cause. (Bottom) The cumulative percentage of patients in the COMS randomized trial of PERT who had died with metastatic melanoma by the specified times since enrollment. Blue line: patients assigned to PERT; Red line: patients assigned to enucleation alone. The numbers of patients at risk of death and numbers censored based on date of enrollment, are given at annual anniversaries of enrollment by random treatment assignment. Event is death with histologically confirmed metastatic melanoma.

(Modified with permission from Collaborative Ocular Melanoma Study Group. Am J Ophthalmol 2004;138:936–57.)

Among patients assigned to enucleation alone, 130 (26%) died within 5 years of enrollment with histologically confirmed metastatic melanoma, compared with 139 (28%) patients assigned to pre-enucleation radiation, based on review of 435 of the 457 deaths by the Mortality Coding Committee.^30,³¹ Time to death with histologically confirmed metastatic melanoma is displayed in Figure 150.3. The liver was the most common site of melanoma metastasis.³¹ Findings pooled by treatment arm for mortality, diagnosis of melanoma metastases, and second primary cancers are summarized through 10 years in Figure 150.4.

Fig. 150.4 The percentage of patients in the COMS randomized trial of PERT with the specified status at the end of each year of follow-up, with treatment arms pooled. The number of patients who enrolled early enough to be followed up to the end of the specified year is shown at the top of the corresponding bar.

(Reproduced with permission from Collaborative Ocular Melanoma Study Group. Am J Ophthalmol 2004;138:936–57.)

Complications

Only 17 patients treated with enucleation alone and 19 patients treated with pre-enucleation radiation had any surgical or anesthetic complication reported at the time of initial treatment.¹⁷ Orbital tumor recurrence was reported during the first 5 years of follow-up for six patients treated with enucleation alone and for one patient treated with pre-enucleation radiation; patients treated with enucleation alone had nearly twice the 5-year incidence of severe ptosis as reported for patients treated with pre-enucleation radiation.¹⁷

Findings from the coms nonrandomized prospective study of small choroidal melanoma

Of 300 patients with small choroidal melanoma (by COMS criteria) reported from December 1986 through August 1989, 220 were judged eligible for this COMS observational substudy; 204 gave signed consent and enrolled. The majority of the patients enrolled within 1 year of the initial diagnosis of choroidal melanoma.²⁴ No attempt was made to establish uniform criteria for treatment timing or to recommend the type of treatment; the patients and their ophthalmologists made these decisions.

A total of 16 patients were treated shortly after enrollment; 20 additional patients were treated after the melanoma grew to medium or large size (COMS criteria) and the patients were enrolled in one of the COMS randomized trials. As of June 1996, an additional 47 patients had been treated during follow-up.^24,²⁵

By June 30, 1996, 27 patients had died. Survival findings are summarized in Figure 150.5. The estimated 5-year all-cause mortality rate was 6% (95% CI: 3–9%).²⁵ Of 188 patients who were not treated at time of enrollment, 44 had tumors that had grown to medium or large size by February 1997, based on COMS criteria. The estimated 5-year proportion of initially small tumors that grew was 31% (95% CI: 23–39%).²⁴

Fig. 150.5 Cumulative percentage of patients in the COMS nonrandomized, prospective study of small choroidal melanoma who had died by the specified times since enrollment. Thick line: death from any cause. Thin line: death from metastatic melanoma.

Histopathologic findings from enucleated eyes

The COMS Group has one of the highest rates of diagnostic accuracy ever documented.^29,³² Of 1532 eyes enucleated in the COMS by June 1996, 994 from patients enrolled in the randomized trial for large choroidal melanoma and 536 from patients enrolled in the trial for medium choroidal melanoma, 1527 (99.7%) were confirmed histopathologically by the Pathology Review Committee to harbor choroidal melanoma.^29,³² A detailed description of the characteristics of the 1527 confirmed cases has been published.³² Key findings include documentation of local tumor invasion: rupture of Bruch’s membrane in 88% of eyes; invasion of emissary canals in 55%; retinal invasion in 49%; tumor cells in the vitreous in 25%; invasion of tumor vessels in 14%; and vortex vein invasion in 22% of eyes with vortex veins. Scleral invasion was present in 56% of eyes and extrascleral extension in 8%. Histopathologic review confirmed that pre-enucleation radiation significantly reduced mitotic activity.¹²

In other published histopathologic investigations of enucleated eyes of COMS patients, silver-stained nucleolar organizer region scores have been evaluated as predictors of later metastasis, transillumination and histologic measurements of tumor dimensions have been compared,³³ and a clear cell variant of choroidal melanoma has been identified.³⁴

Other published findings

The multicenter organization of the COMS facilitated referral to a COMS clinical center of cases of choroidal melanoma from most of the USA and from a large part of eastern Canada. The large number of patients screened for the COMS and judged to have choroidal melanoma (totaling 8712 as of July 31, 1998, when accrual halted) provides the largest group of patients with this diagnosis for whom data have been collected systematically at time of presentation in accord with a common protocol. Trends in size of choroidal melanoma and treatments over time have been published.³⁵ The COMS Group has reported the first published cases of choroidal melanoma in Native Americans.³⁶ This large database permits comparison of tumor characteristics at time of screening and diagnosis among different racial subgroups. This information may provide clues for future investigation in epidemiologic and genetic studies of choroidal melanoma.

The COMS Group also has published information on surgical and postsurgical complications among the largest group of patients whose eyes have been enucleated because of choroidal melanoma and who have been examined in accord with a common follow-up protocol,¹⁷ documenting the low rates of serious complications following enucleation for this condition. The importance of liver function tests and other tests for metastasis has been evaluated; those findings have been published in the oncology literature.³⁷ Also, 10-year changes in fellow eyes of patients enrolled in COMS randomized trials have been reported.³⁸ Echographic characteristics of melanoma at baseline³⁹ and post-brachytherapy changes observed on fluorescein angiograms and stereoscopic photographs⁴⁰ also have been published. In addition to publications from the COMS Group that have important clinical information, several articles have been published that deal with research methodology.⁴¹^–⁴⁶