Nationwide Incidence of Ocular Melanoma in South Korea by Using the National Cancer Registry Database (1999–2011)

Abstract

Purpose: The purpose of this study was to determine the age-standardized incidence rate of ocular melanoma and its subtypes in South Korea from 1999 to 2011.

Methods: The national cancer registry in South Korea was reviewed to ascertain all ocular melanoma incidents between 1999 and 2011, which covered the entire South Korean population. Age-standardized incidence rates of ocular melanoma and its subtypes were calculated. Trends in incidence were analyzed by the use of Joinpoint regression. Analyses were stratified by tumor type, sex, age, and year of cancer diagnosis.

Results: A total of 464 ocular melanoma cases (227 men [48.9%]) were identified from 1999 to 2011. The age-standardized incidence rate for 1999 to 2011 was 0.60 (95% confidence interval [CI]: 0.55–0.66) per million people. Among ocular melanomas, uveal melanoma had the highest age-standardized incidence rate for 1999 to 2011 (0.42 [95% CI: 0.38–0.47] per million people), followed by conjunctival melanoma (0.12 [95% CI: 0.09–0.14] per million people). The age-standardized incidence rate was higher in 2006 to 2011 than in 1999 to 2005, and increasing incidence trends were observed throughout the study period.

Conclusions: Incidence rate of ocular melanoma for 1999 to 2011 is relatively lower in South Korean population than in that reported in white population. Although ocular melanoma is still considered a rare cancer in South Korea, our findings suggest increasing incidence trends. This may warrant future in-depth study.

Uveal melanoma is the most common primary intraocular malignancy in adults, and ocular melanoma, which includes uveal melanoma, is a common type of noncutaneous melanoma.^1–4 Studies have reported the incidence of ocular melanoma and its subtypes in white populations, and recent studies have suggested that incidence of ocular melanoma might vary across ethnicities and regions.^4–12 However, there are few data addressing the ocular melanoma incidence in Asian populations, which include only a few population-based incidence studies.^6,7,10 The uncertainty surrounding the ocular melanoma incidence in Asian populations led us to conduct this study based on the Korean national cancer registry. As ocular melanoma is quite rare in Asian populations,^6,7,10 a representative national registry system with a sufficient population at risk was needed to provide reliable incidence estimates.¹³ South Korea is a promising place in which to investigate the ocular melanoma incidence in an Asian population because Korea, which is situated in Northeast Asia, is among the most populous countries worldwide (approximately 50 million people) and has provided universal health care coverage since 1988.¹⁴ In addition, the national cancer registry, the Korea Central Cancer Registry (KCCR), has recorded all cancer incidents in South Korea since 1999.^13,15 Using the KCCR database, we estimated the numbers and age-standardized incidence rates of ocular melanoma in South Korea during a 13-year period from 1999 to 2011 and report the changes that occurred during this period. 

Methods

Data Sources

Data from 1999 to 2011 were obtained from the KCCR database. The KCCR was established in 1980 by the Korean Ministry of Health and Welfare as a national cancer registry, and in 1999, the KCCR began to collect nationwide cancer incidence data by covering the entire population in South Korea under a population-based regional cancer registry program.^13,15 Detailed descriptions of the KCCR database have been provided elsewhere.^13,15,16 Briefly, for each cancer patient, the database includes information such as age, sex, region, birthdate, histological type, date of diagnosis, and primary cancer site. Information regarding staging was collected using Surveillance, Epidemiology and End Results (SEER) summary staging¹⁷ and reliable information available since 2006.¹⁸ Population data were also obtained from Statistics Korea, using the mid-year population on July 1 of the specified years (available at http://kosis.kr). The research protocol for the present study was approved by the institutional review board of the National Cancer Center (IRB No: NCC2014-0156, Goyang, Republic of Korea). 

Definition and Classification of Ocular Melanoma

Ocular melanoma was defined as site “C69.0–C69.9” and morphology “8720-8790” according to the International Classification of Diseases for Oncology (3rd edition).¹⁹ We classified ocular melanoma as uveal melanoma (codes C69.2, C69.3, C69.4), conjunctival melanoma (codes C69.0, C69.1), adnexal melanoma (codes C69.5, C69.6, C69.8), or unspecified melanoma (code C69.9) according to anatomical site.^5,11,20 Uveal melanoma was subdivided into choroidal melanoma (codes C69.2, C69.3) or iris and ciliary body melanoma (code C69.4). Retinal melanoma was considered a choroidal melanoma, because retinal melanoma cases were assumed to be cases likely to be miscoded as uveal melanoma.^5,11 

Statistical Analysis

We assessed characteristics (i.e., sex, age at diagnosis [patients diagnosed before 54 years of age and after 55 years of age], ocular melanoma subtypes, diagnostic method, and SEER summary staging) of ocular melanoma patients identified during the 13-year study period from 1999 to 2011 and also described the characteristics of those identified in each of two subdivided study periods corresponding to the first 7 years (1999–2005) and last 6 years (2006–2011). 

We estimated crude incidence rates of ocular melanoma in each age group by the use of the number of identified ocular melanoma cases as the numerator and the sum of the mid-year populations during the study period as the denominator. All estimated crude incidence rates were applied by direct age standardization, using Segi's world standard population²¹; expected cases of each age group were obtained by multiplying the age-specific crude incidence rate by proportion of population in the corresponding age groups of Segi's world standard population.²¹ Then, age-standardized incidence rate was calculated as the sum of expected cases of each age group per 1,000,000 person-years; the age-standardized incidence rate was presented as numbers of new cases per 1,000,000 person-years. Specific age-standardized incidence rates according to the defined characteristics were also calculated. In addition, trends in incidence of ocular melanoma and its subtypes throughout the study periods were assessed with the Joinpoint regression model.²² Joinpoint regression model is a segmented linear regression analysis to investigate trends in rates and identify statistically significant changes over time.²² Furthermore, the model is used to detect time points where there are significant changes in trends. In the analysis of population-based cancer registry data, the model is often used to characterize trends in incidence and mortality rates. In our analysis, trends in incidence rates were modeled. The model can be fitted using the Joinpoint software developed by the US National Cancer Institute (http://surveillance.cancer.gov/joinpoint). We carried out the Joinpoint regression model by using the biennial age-standardized incidence rates from 2002 to 2011 and the age-standardized incidence rate of the first 3 years (1999–2001) to reduce instability due to the small sample size in the present study. Incidence trends were investigated separately in men and women. The result of Joinpoint analysis is represented as annual percent changes (APC), which is calculated as summary measurements.²² P values of <0.05 were considered statistically significant. Statistical analyses were performed using Stata version 12.0 (StataCorp LP, College Station, TX, USA), SAS version 9.3 (SAS Institute, Cary, NC, USA), and Joinpoint regression version 4.1.1 (Surveillance Research Program, National Cancer Institute, Bethesda, MD, USA) software. 

Results

A total of 464 patients (227 men [48.9%]) with ocular melanoma were identified in South Korea from 1999 to 2011. Table 1 provides baseline characteristics and detailed information for all ocular melanoma patients identified in the present study. Briefly, the mean and median ages of the ocular melanoma patients were 54.0 and 54.5 years, respectively. Uveal melanoma was most common, accounting for 70.3% of all cases (326 cases), followed by conjunctival melanoma (90 cases [19.4%]). Of 326 uveal melanoma cases, a total of 258 choroidal melanoma cases were ascertained, which included 5 cases having the retinal melanoma diagnostic code. Most cases (367 cases [79.1%]) were confirmed by microscopic histology, whereas the remaining 97 cases (20.9%) were diagnosed by clinical examination or other methods. The proportion of microscopically confirmed cases decreased from 92.9% in 1999 to 2005 to 68.9% in 2006 to 2011. 

Table 1

 

Baseline Characteristics of Ocular Melanoma Patients in South Korea During the Study Period (1999–2011) According to the KCCR Database

 

The age-standardized incidence rate of ocular melanoma in 1999 to 2011 was 0.60 (95% confidence interval [CI]: 0.55–0.66) per 1,000,000 person-years; in men and women, those in 1999 to 2011 were 0.63 (95% CI: 0.55–0.72) per 1,000,000 person-years and 0.58 (95% CI: 0.50–0.65) per 1,000,000 person-years, respectively. The age-standardized incidence rate among patients ≥55 years of age in 1999 to 2011 was 2.09 (95% CI: 1.82–2.36) per 1,000,000 person-years, which was significantly higher than that among patients <55 years of age in 1999 to 2011 (0.34 [95% CI: 0.29–0.38] per 1,000,000 person-years). Among ocular melanoma subtypes, uveal melanoma had the highest age-standardized incidence rate at 0.42 (95% CI: 0.38–0.47) per 1,000,000 person-years in 1999 to 2011, followed by conjunctival melanoma (0.12 [95% CI: 0.09–0.14] per 1,000,000 person-years). Regarding the SEER summary staging used since 2006, the localized stage was most common in Korea (except for unknown cases), followed by regional and distant stages. The age-standardized incidence rate was significantly higher among patients with ocular melanoma diagnosed in 2006 to 2011 (0.69 [95% CI: 0.61–0.77] per 1,000,000 person-years) than among those with melanoma diagnosed in 1999 to 2005 (0.52 [95% CI: 0.44–0.59] per 1,000,000 person-years). This increasing incidence was also observed for the subtype of uveal melanoma (as well as for choroidal melanoma). In addition, age-standardized incidence rate of ocular melanoma diagnosed by clinical examination or other methods increased considerably from 0.03 (95% CI: 0.02–0.05) per 1,000,000 person-years in 1999 to 2005 to 0.22 (95% CI: 0.17–0.26) per 1,000,000 person-years in 2006 to 2011, whereas that diagnosed by microscopic histopathology remained stable from 0.48 (95% CI: 0.41–0.55) per 1,000,000 person-years in 1999 to 2005 to 0.47 (0.40–0.54) per 1,000,000 person-years in 2006 to 2011. Detailed results are provided in Table 2. The crude incidence rates of ocular melanoma before applying the direct age-standardization are also provided in Supplementary Table S1. 

Table 2

 

Age-Standardized Incidence Rate of Ocular Melanoma (per 1,000,000 person-years) in South Korea During the Study Period (1999–2011) According to the KCCR Database

 

The age-standardized incidence rates of both ocular melanoma (Fig. A) and the subtype uveal melanoma (Fig. B) exhibited increasing trends in both men and women throughout the study period, whereas those of the subtypes conjunctival melanoma and adnexa melanoma did not exhibit increasing trends in either sex during the study period. The APCs of ocular melanoma were 6.2 (95% CI: 0.1–12.6) in men and 15.3 (95% CI: 5.0–26.7) in women, and those of uveal melanoma were 9.4 (95% CI: 5.2–13.6) in men and 19.0 (95% CI: 11.4–27.2) in women. Data are not provided for the other ocular melanoma subtypes because of the small numbers of available cases. 

Discussion

We set out to determine the incidence rate and trend among ocular melanoma patients in South Korea from 1999 to 2011 according to the national cancer registry, which covers a large Asian population (approximately 50 million). The present study identified 464 ocular melanoma cases, which corresponds to the age-standardized incidence rate of 0.60 per 1,000,000 person-years during the study period. In addition, the present study also provided detailed incidence estimates according to sex and ocular melanoma subtype and subdivided study periods. 

Uveal melanoma accounts for a large portion of ocular melanoma cases in the present study, similar to the results reported for white populations, but the corresponding incidence rate in our study was quite low compared to rates reported for white populations.^{3,5,7,9,11,23} Similarly, Cheng and Hsu⁶ reported an ocular melanoma incidence rate of 0.39 (95% CI: 0.30–0.40) per 1,000,000 people in Taiwan according to the national cancer registry, although they identified only 128 cases in an 18-year period and did not provide specific incidence rates for ocular melanoma subtypes. As virtually nothing is known about the incidence rates of ocular melanoma subtypes in Asian populations, Hu et al.^7,10 investigated racial/ethnic differences in the development of uveal and conjunctival melanoma in the United States using the SEER database. These authors reported incidence rates of uveal melanoma and conjunctival melanoma among US residents of Asian ancestry as 0.38 (95% CI: 0.2–0.8) per 1,000,000 people and 0.15 (95% CI: 0.05–0.35) per 1,000,000 people, respectively, which are similar to the rates obtained in the present study (0.42 and 0.12 per 1,000,000 people, respectively). The incidence rates of both uveal melanoma and conjunctival melanoma are lower in Asian populations than in white populations. However, although conjunctival melanoma is far less frequent than uveal melanoma in both Asian and white populations, it is interesting that conjunctival melanoma accounted for a larger percentage of ocular melanoma in the present study relative to that of white populations.^1,7,10 The effect of ethnicity on ocular melanoma development might therefore vary in accordance with the tumor site. 

Historically, uveal melanoma occurs more frequently in men than in women.^{1,2,8,9,11,23} However, in the present study, no significant differences between the sexes were observed. This might be because of inherent characteristics of uveal melanoma in Asian populations or because the number of ocular melanoma cases was insufficient to reveal a difference. In addition, in the present study, the incidence rate among patients ≥55 years of age was higher than that among patients <55 years of age (2.09 and 0.34 per 1,000,000 person-years, respectively), which was consistent with the results of previous studies.^{1,2,4,23–26} 

Interestingly, the ocular melanoma incidence rate has increased in South Korea; the incidence rate in 2006 to 2011 was significantly higher than that in 1999 to 2005, and the results of the Joinpoint regression analysis also revealed an increasing trend in ocular melanoma incidence in both men and women. The increase in uveal melanoma incidence or, to be exact, the increase in choroidal melanoma incidence accounted for most of the increased ocular melanoma incidence observed in the present study, whereas the incidence rates of other subtypes of ocular melanoma remained largely stable. Although some European registries, including those of France and Italy, have reported increased incidence rates of uveal melanoma,⁸ the incidence of uveal melanoma has remained stable in most white populations.^2,5,12,27 In fact, some studies from the United States and Sweden have reported decreases in the uveal melanoma incidence.^3,8,28 The present study observed a large increase in incidence rates of ocular melanoma diagnosed by clinical examination but not in cases diagnosed by microscopic histopathology during the study period, suggesting that the observed increase might not indicate a real increase of ocular melanoma occurrence. Rather, the increase in screening medical check-ups in South Korea during the study period might have affected the increased incidence of choroidal melanoma. In 1999, the South Korean government initiated a national screening program for cancer and other common diseases, which was available free of charge or with a small copayment.²⁹ Although ophthalmic examinations were not included in this program, health care providers frequently offer fundus screening with nondilated fundus photography, with an additional low payment of approximately $20 to $30.^29,30 Consequently, numerous hospitals have provided a “health check-up program” that includes fundus photography. In addition, the number of certificated ophthalmologists doubled during the study period (1456 doctors in 1999 and 2976 doctors in 2011 were affiliated with the Korean Ophthalmological Society). As the frequency of clinically diagnosed ocular melanoma increased considerably from 7.1% in 1999 to 2005 to 31.1% in 2006 to 2011, these factors may explain the observed increase in choroidal melanoma incidence in South Korea during the study period at least in partial. Further investigation is warranted to elaborate the observed increases in incidence rates. 

The present study has several limitations. The KCCR database did not include tumor characteristics such as tumor size, intraocular localization of tumor, histopathologic features, and cytogenetic features. The database also lacked proper information regarding the treatments and metastatic status for each ocular melanoma patient. Caution is advised when interpreting the results for ocular melanoma subtypes and trends according to the year of diagnosis because the reported estimates were based on relatively small numbers of patients. In addition, although the number of retinal melanoma cases were quite rare (5 cases), the incidence rate of uveal melanoma (and choroidal melanoma) might be overestimated, because the analyses for uveal melanoma (and choroidal melanoma) estimates included cases having retinal melanoma codes. 

In conclusion, we have reported the nationwide, population-based incidence rate of ocular melanoma in South Korea. As the epidemiologic characteristics of ocular melanoma incidence in Asian populations have not been fully elucidated, further investigations regarding incidence rates and trends in Asian populations are warranted.1552 

Figure

 

Age-standard incidence rate (per 1,000,000 person-years) of ocular melanoma and uveal melanoma in Korea from 1999 to 2011. The incidence rate is presented as numbers of new cases per million people. The incidence rates were rounded to two decimal points. (A) Age-standardized incidence rates of ocular melanoma. The annual percent changes in ocular melanoma were 6.2 (95% CI: 0.1–12.6) in men and 15.3 (95% CI: 5.0–26.7) in women. (B) Age-standardized incidence rate of uveal melanoma. The annual percent changes in uveal melanoma were 9.4 (95% CI: 5.2–13.6) in men and 19.0 (95% CI: 11.4–27.2) in women.

 

Segi's world standard population was used for age-standardization. Joinpoint regression analysis was conducted to determine whether ocular melanoma and uveal melanoma had increased significantly relative to sex during every 2-year period except for the initial period of 1999–2001. Numbers in boxes are the age-standardized incidence of ocular melanoma (A) and uveal melanoma (B). Circles indicate the age-standardized incidence rate of ocular melanoma (A) and uveal melanoma (B) in men, and solid lines present the projected incidence trends of ocular melanoma and uveal melanoma in men according to the Joinpoint regression model.

 

Large X-shapes indicate the age-standardized incidence rate of ocular melanoma (A) and uveal melanoma (B) in women, and dashed lines present the projected incidence trends of ocular melanoma and uveal melanoma in women according to the Joinpoint regression model.

Acknowledgments

Supported by the National Cancer Center, Korea, Grant NCC-1410880-2 and National Research Foundation of Korea Grant NRF-2013R1A2A2A04015829. 

Disclosure: S.J. Park, None; C.-M. Oh, None; B.W. Kim, None; S.J. Woo, None; H. Cho, None; K.H. Park, None