Incidence of Retinoblastoma and Survival Rate of Retinoblastoma Patients in Korea Using the Korean National Cancer Registry Database (1993–2010)

Abstract

Purpose.: We determined the incidence of retinoblastoma and the long-term survival rate of retinoblastoma patients in South Korea.

Methods.: We reviewed retrospectively data from the Korea Central Cancer Registry recorded between 1993 and 2010 to identify the incidence of retinoblastoma and the survival rate of retinoblastoma patients in Korea. The incidence of retinoblastoma was estimated by population-based analysis of children aged 0 to 4 years and children aged 0 to 9 years, population-based analysis per live birth, and birth cohort analysis. Survival was analyzed using the Korea Central Cancer Registry database, which was crosschecked with the national death registry.

Results.: The overall incidence was 11.2 for children aged 0 to 4 years and 5.3 for children aged 0 to 9 years per 1,000,000 person-years, 5.9 per 100,000 live births, and 5.3 per 100,000 live births. Birth cohort analysis showed less variable results in incidence rates over 4 calendar-periods compared to the population-based analyses. The all-cause mortality rate was 7.9% at 5 years and 8.4% at 10 years. The rate improved from 12.5% for patients diagnosed in 1993 to 2000 to 4.5% for those diagnosed in 2001 to 2010.

Conclusions.: The incidence of retinoblastoma in Korea was found to be similar to that in the United States, Europe, and Asia. The survival rate of retinoblastoma patients in Korea was significantly better during 2001 to 2010 than during 1993 to 2000.

Introduction

Retinoblastoma is the most frequent primary intraocular malignancy occurring in children in Korea. ^1,2 Although many studies have provided similar results for the incidence of retinoblastoma across the world and races, the incidence of retinoblastoma in Asians including Koreans is, indeed, obscure. ^3–16 Because retinoblastoma is rare, providing a reliable estimate of its incidence requires a representative national registry system or large claim database with a definite population at risk of retinoblastoma. Therefore, most of the well-designed studies of retinoblastoma incidence have been reported from developed countries, and the annual incidence of retinoblastoma for children aged 0 to 4 years in these studies remains constant at 10.0 to 11.8 per 1,000,000 person-years. ^{6–8,10–14} In addition, survival rates for retinoblastoma also vary by race and region, including differences in developed regions versus developing regions. For example, Asia and Africa had a high mortality rate of up to 70%, while developed countries demonstrated a mortality rate of 3% to 5%. ^{3,4,8,10,17,18} Korea is a developing nation with a relatively large population and a medical system that has witnessed rapid advances over the last several decades. Using a national cancer registry database, we investigated the incidence of retinoblastoma and all-cause mortality rate of retinoblastoma patients in Korea for an 18-year period (1993–2010), and report the changes that occurred during this period. 

Methods

Data Source

Data were obtained using the Korea Central Cancer Registry (KCCR) database. Details of the history, objectives, and activities of the KCCR have been reported previously. ^19,20 Briefly, the Korean Ministry of Health and Welfare established a nationwide hospital-based cancer registry in 1980. The registry collected data on cancer patients annually from more than 180 training hospitals throughout the country. As there were only approximately 20 to 30 tertiary hospitals in Korea in those days, the registry covering over 180 hospitals is believed to ascertain all retinoblastoma cases diagnosed in Korea. In 1999, the National Health Insurance (NHI) scheme and its unified claim-review system covering the entire Korean population were established. Using the NHI database, the KCCR set the Population-Based Regional Cancer Registry program. ²⁰ After that, the KCCR ascertains cancer cases using the database of the NHI covering the entire Korean population, the database of the 11 population-based regional cancer registries, the death certificates database, the database of site-specific cancer registries from medical societies, and the database of hospital-based registries. To estimate the survival rate of retinoblastoma patients, the KCCR matched the cases diagnosed during 1993 to 2010 to the vital status of the patient using information from the death registry at Statistics Korea. 

Estimating the Incidence Rate of Retinoblastoma

We identified patients diagnosed with retinoblastoma between 1993 and 2010 from the KCCR database. The population at risk of developing retinoblastoma was defined as the entire population of a defined age group in Korea based on the Population and Housing Census (PHC) available from the Korean Statistical Information Service (KOSIS, available in the public domain at http://kosis.kr). The Korean Statistical Information Service of the Korean central government conducts the PHC every 5 years to obtain information regarding the size, distribution, and structure of the population and housing in Korea. The PHC was conducted in 1990, 1995, 2000, 2005, and 2010, and the next PHC is scheduled for 2015. We divided the study period into the 4 time periods: 1993 to 1995, 1996 to 2000, 2001 to 2005, and 2006 to 2010. The person-time incidence rates were calculated as the number of people who developed retinoblastoma in each time period divided by the total person-time at risk in that period. The population at risk for each period was defined as the mean of the two nearby PHC population data; for example, using the mean values of 2005 PHC and 2010 PHC as the population at risk of the period 2006 to 2010. The number of cases in 1993 to 1995, 1996 to 2000, 2001 to 2005, and 2006 to 2010 was divided by the corresponding population at risk. Therefore, in this analysis, person-years were counted after the incident time. The mean incidence rate of retinoblastoma for the 18-year study period was calculated after age-adjusting for the population of 2010 on the basis of the PHC data. We also estimated the incidence of retinoblastoma per live birth by standard annual analysis and birth cohort analysis. Standard annual analysis is the incidence estimated as the number of people who developed retinoblastoma in each 4-calendar period divided by the total number of live-born children in that period. Birth cohort analysis is the incidence estimated as the number of people who developed retinoblastoma among the people born in each period (birth cohorts) divided by the total number of live-born children in that period. ¹¹ We ascertained the retinoblastoma cases through December 31, 2010. On the basis of the KOSIS birth certificate database, there were 10,162,720 live-born children at risk of developing retinoblastoma during 1993 to 2010, including 2,152,031, 3,243,094, 2,445,341, and 2,322,254 live-born children during 1993 to 1995, 1996 to 2000, 2001 to 2005, and 2006 to 2010, respectively. 

Estimating the Survival Rate of Patients With Retinoblastoma

Survival analysis was conducted in all patients diagnosed as having retinoblastoma during the 18-year study period using the Kaplan-Meier method (product-limit method). We ascertained the data of vital status through December 31, 2011, which had included all cause of deaths. Regretfully, there is no accessible information regarding the cause of death and the development of second malignant neoplasm in the KCCR database. The survival duration of the patients with retinoblastoma was determined as the interval between the date of initial diagnosis of retinoblastoma and the date of death, date of loss to follow-up, or closing date for follow-up. We compared the survival rates of patients diagnosed in 1993 to 2000 with those diagnosed in 2001 to 2010, of patients diagnosed at age ≤ 1 year with those diagnosed at age ≥ 2 years, and between sexes. 

Statistical Analysis and Ethics

We used SAS version 9.2 (SAS Institute, Inc., Cary, NC, USA) for all analyses. A 95% confidence interval (CI) of the incidence rate was estimated based on Poisson distribution. A χ² analysis was used to compare the observed incidence rates between the sexes. The log-rank test was conducted to assess the difference between 2 Kaplan–Meier curves. P values < 0.05 were considered statistically significant. The institutional review board of the Seoul National Bundang Hospital approved the study, which adhered to the tenets of the Declaration of Helsinki. 

Results

Incidence Rate of Retinoblastoma

In total, 600 patients (334 male patients, 55.7%) with retinoblastoma were identified from 1993 to 2010 in Korea. We could not find any retinoblastoma patient aged ≥10 years at the time of diagnosis. The birth cohort from 1993 to 2010 included 537 cases (298 male patients, 55.5%). 

The age- and sex-standardized incidence of retinoblastoma (the unit of incidence is per 1,000,000 person-years) during the entire study period was 11.2 (95% CI, 10.2–12.2) among children aged 0 to 4 years and 5.3 (95% CI, 4.8–5.8) among those aged 0 to 9 years; in the male population, the age-standardized incidence was 11.8 (95% CI, 10.3–13.2) in those aged 0 to 4 years and 5.6 (95% CI, 4.9–6.2) in those aged 0 to 9 years; in the female population, the age-standardized incidence was 10.5 (95% CI, 9.1–11.9) in those aged 0 to 4 years and 5.0 (95% CI, 4.4–5.7) in those aged 0 to 9 years. Difference of retinoblastoma incidence between the sexes was not statistically significant during the 18-year study period (P = 0.13), and those were not statistically significant even in subperiod analyses except in the 1996 to 2000 period (P = 0.92, 0.02, 0.42, and 0.67 in 1993–1995, 1996–2000, 2001–2005, and 2006–2010, respectively). Detailed age- and sex-specific frequencies and incidence rates of retinoblastoma in each 4-time period are provided in Table 1. 

Table 1

 

Number of Retinoblastoma Cases Recorded in the Database of the KCCR and the Incidence Rate of Retinoblastoma in Korea During the Study Period (1993–2010)

Age	Total		Male		Female
	N	Incidence (95% CI)	N	Incidence (95% CI)	N	Incidence (95% CI)
1993–1995
0	25	NA	14	NA	11	NA
1	37	NA	19	NA	18	NA
2	25	NA	14	NA	11	NA
3	8	NA	4	NA	4	NA
4	4	NA	4	NA	0	NA
0–4	99	9.8 (7.9–11.8)	55	10.3 (7.6–13.1)	44	9.3 (6.5–12.0)
5	1	NA	0	NA	1	NA
6	5	NA	1	NA	4	NA
7	0	NA	0	NA	0	NA
8	0	NA	0	NA	0	NA
9	0	NA	0	NA	0	NA
5–9	6	0.6 (0.1–1.0)	1	0.2 (0.0–0.5)	5	1.0 (0.1–1.9)
Total	105	5.1 (4.1–6.1)	56	5.2 (3.8–6.6)	49	5.0 (3.6–6.4)
1996–2000
0	60	NA	36	NA	24	NA
1	47	NA	26	NA	21	NA
2	40	NA	23	NA	17	NA
3	15	NA	12	NA	3	NA
4	5	NA	4	NA	1	NA
0–4	167	10.2 (8.6–11.7)	101	11.7 (9.4–13.9)	66	8.5 (6.5–10.6)
5	3	NA	2	NA	1	NA
6	2	NA	2	NA	0	NA
7	0	NA	0	NA	0	NA
8	1	NA	1	NA	0	NA
9	1	NA	1	NA	0	NA
5–9	7	0.4 (0.1–0.8)	6	0.7 (0.1–1.3)	1	0.1 (0.0–0.4)
Total	174	5.3 (4.5–6.1)	107	6.2 (5.0–7.4)	67	4.3 (3.3–5.4)
2001–2005
0	59	23.3 (17.5–29.3)	30	22.7 (14.6–30.8)	29	24.0 (15.3–32.7)
1	66	25.3 (19.2–31.4)	38	28.0 (19.1–36.9)	28	22.4 (14.1–30.6)
2	36	13.1 (8.82–17.4)	21	14.7 (8.4–20.9)	15	11.4 (5.6–17.2)
3	17	6.0 (3.2–8.9)	10	6.8 (2.6–11.0)	7	5.2 (1.3–9.0)
4	2	0.7 (0.0–1.6)	1	0.6 (0.0–1.9)	1	0.7 (0.0–2.0)
0–4	180	13.1 (11.2–15.0)	100	13.9 (11.2–16.6)	80	12.2 (9.5–14.8)
5	1	0.3 (0.0–0.9)	1	0.6 (0.0–1.7)	0	0.0
6	3	0.9 (0.0–2.00)	2	1.2 (0.0–2.8)	1	0.7 (0.0–2.0)
7	1	0.3 (0.0–0.9)	0	0.0	1	0.6 (0.0–1.9)
8	0	0.0	0	0.0	0	0.0
9	0	0.0	0	0.0	0	0.0
5–9	5	0.3 (0.00–0.6)	3	0.3 (0.00–0.7)	2	0.3(0.0–0.6)
Total	185	6.1 (5.2–7.00)	103	6.5 (5.2–7.7)	82	5.7 (4.5–6.9)
2006–2010
0	52	24.4 (17.8–31.1)	25	22.8 (13.9–31.7)	27	26.2 (16.3–36.1)
1	42	19.1 (13.3–24.8)	26	22.8 (14.1–31.6)	16	15.0 (7.7–22.4)
2	25	10.8 (6.5–15.0)	8	6.7 (2.1–11.3)	17	15.1 (7.9–22.3)
3	7	2.9 (0.8–5.1)	5	4.1 (0.5–7.6)	2	1.7 (0.0–4.2)
4	2	0.8 (0.0–1.9)	0	0.0	2	1.7 (0.0–4.0)
0–4	128	11.1 (9.2–13.1)	64	10.8 (8.1–13.4)	64	11.5 (8.7–14.3)
5	5	1.9 (0.2–3.6)	2	1.5 (0.0–3.5)	3	2.4 (0.0–5.1)
6	3	1.14 (0.00–2.42)	2	1.5 (0.0–3.5)	1	0.8 (0.0–2.3)
7	0	0.0	0	0.0	0	0.0
8	0	0.0	0	0.0	0	0.0
9	0	0.0	0	0.0	0	0.0
5–9	8	0.6 (0.2–1.0)	4	0.6 (0.0–1.1)	4	0.6 (0.0–1.2)
Total	136	5.4 (4.4–6.3)	68	5.2 (3.9–6.4)	68	5.6 (4.2–6.9)

The incidence rate per live birth was 5.9 (95% CI, 5.4–6.4) per 100,000 live births during the study period, corresponding to 1 per 16,938 (95% CI, 15,683–18,411) live births. On the other hand, the birth cohort analysis showed that the incidence rate per 100,000 live births was 5.3 (95% CI, 4.8–5.7) per 100,000 live births during the study period, corresponding to 1 per 18,925 (95% CI, 17,449–20,674) live births. Detailed estimates for incidence rate per live birth and the birth cohort analysis in each 4-time period are provided in Table 2. 

Table 2

 

Incidence Rate per Live Birth and Incidence Rate From Birth Cohort Analysis of Retinoblastoma in Korea During the Study Period (1993–2010)

Period	Incidence per Live Birth (95% CI)	Incidence From Birth Cohort Analysis (95% CI)
1993–1995	4.9 (4.0–5.8) per 100,000 live births or 1 per 20,496 (17,205–25,343) live births	4.8 (3.9–5.7) per 100,000 live births or 1 per 20,894 (17,512–25,894) per live births
1996–2000	5.4 (4.6–6.2) per 100,000 live births or 1 per 18,638 (16,227–21,891) live births	5.3 (4.5–6.1) per 100,000 live births or 1 per 18,746 (16,315–22,029) per live births
2001–2005	7.6 (6.5–8.7) per 100,000 live births or 1 per 13,218 (11,553–15,443) live births	6.7 (5.7–7.7) per 100,000 live births or 1 per 14911 (12,931–17,605) per live births
2006–2010	5.9 (4.9–6.8) per 100,000 live births or 1 per 17,075 (14,618–20,525) live births	4.2 (3.4–5.0) per 100,000 live births or 1 per 23,941 (19,967–29,889) per live births

Survival Rate of Patients With Retinoblastoma

Survival analysis was conducted on all 600 patients diagnosed during the study period. The all-cause mortality rate was 7.9% at 5 years and 8.4% at 10 years (Fig. A). The all-cause mortality rate at 5 years improved from 12.5% for patients diagnosed in 1993 to 2000 to 4.5% for those diagnosed in 2001 to 2010 (log-rank test, P < 0.01, Fig. B). 

Figure

 

(A) The 10-year survival rates including all-cause mortality of patients diagnosed with retinoblastoma from 1993 to 2010 in Korea. The all-cause mortality rate for 10-year period is 8.4%. (B) The 5-year survival rates including all-cause mortality of Korean retinoblastoma patients diagnosed in 1993 to 2000 and 2001 to 2010. The survival rate including all-cause mortality for patients diagnosed in 2001 to 2010 was significantly higher than that for patients diagnosed in 1993 to 2000. (C) The 10-year survival rates including all-cause mortality of male and female retinoblastoma patients diagnosed during 1993 to 2010 in Korea. There was no difference in the survival rates between the sexes. (D) The 10-year survival rates including all-cause mortality of Korean retinoblastoma patients diagnosed at age ≤ 1 and ≥2 years during 1993 to 2010. The age at diagnosis of retinoblastoma was not associated with the survival rate.

There was no difference in the survival rates between sexes; the P value for the log-rank test was 0.25 (Fig. C). In addition, the age at diagnosis of retinoblastoma was not associated with the survival rate. The P value of the log-rank test conducted for patients diagnosed at age ≤ 1 year and those diagnosed at age ≥ 2 years was 0.76 (Fig. D). 

Discussion

The present study evaluated the incidence of retinoblastoma and the survival rate of patients with retinoblastoma from 1993 to 2010 using the Korean population-based national cancer registry. The data from the KCCR is believed to include all cancer cases in Korea; therefore, it should contain reliable data regarding retinoblastoma incidence. Our analysis has revealed conclusive data for retinoblastoma survival because the KCCR crosschecks the vital status of the patients using the death registry from the Statistics Korea and the above-stated registries. 

The annual incidence of retinoblastoma for children aged 0 to 4 years in Korea is similar to that reported by previous studies: 11.0, 10.9, and 11.8 in the United States, ^6,7,12 11.1 in Singapore, ¹⁰ 14.8 in Vietnam, ⁹ 11.6 in Northern Europe (Sweden and Finland), ¹¹ and 10.0 in Britain ¹⁴ per 1,000,000 person-years (Table 3). The incidence rate per live birth in Korea was 1 per 16,938 live births and also was comparable to that of previous studies: 1 per 17,000 live births in the Netherlands, ⁸ 1 per 15,290 live births in Northern Europe, ¹¹ and 1 per 18,000 to 24,000 live births in the United States (Table 2). ^5,12,21,22 Kim and Yu ¹⁵ reported the incidence of retinoblastoma diagnosed in 1991 to 1993 based on a retrospective hospital-based registry in Korea as 5.0 per 100,000 live births, corresponding to 1 per 20,041 live births. We estimated a similar incidence rate value in 1993 to 1995. 

Table 3

 

Reported Annual Incidence of Retinoblastoma in the Literature

Region	Coverage	Source	Study Period	Age < 5 y per Million	Age < 10 y per Million	Live Births
North America
United States	SEER database (10% of US population)	Pendergrass and Davis6	1974–1976	11.0
United States		Tamboli et al.7	1974–1985	10.9
United States		Broaddus et al.12	1975–2004	11.8
Europe
Netherlands	Nationwide	Moll et al.8	1862–1995			1:17,000
Sweden	Nationwide	Seregard et al.11	1958–1998	11.8		1:14,900
Finland	Nationwide	Seregard et al.11	1958–1988	11.2		1:16,100
Britain	Nationwide	MacCarthy et al.14	1963–2002	10.0	3.5
Africa
Kenya	Nationwide	Nyamori et al.4	2006–2007			1:17,000
Asia
Japan	Nagasiki	Takano et al.16	1965–1986			1:16,000
Japan	Nationwide	The Committee for the National Registry of Retnoblastoma17	1975–1985			1:19,800
Singapore	Nationwide	Saw et al.10	1968–1995	11.1
Vietnam	HoChiMinh	Nguyen et al.9	1995–1997	14.8
Korea	Nationwide	Kim and Yu15	1991–1993			1:20,000
Korea	Nationwide	Park et al. (present study)	1993–2010	11.2	5.3	1:16,900

 SEER, Surveillance, Epidemiology and End Results Program of the National Cancer Institute in the United States.

Seregard et al. ¹¹ recommended using the birth cohort analysis for epidemiologic studies of retinoblastoma incidence, because this method produces least variable incidence rates. In our study, retinoblastoma incidence estimated by the birth cohort method showed slightly lower and less variable values compared to that by the standard annual analysis. As the annual number of births in Korea has sharply declined since 1993, from 715,826 live births in 1993 to 470,171 live births in 2010, the birth cohort analysis may reflect the accurate incidence of retinoblastoma better than the standard methods stated above. However, the incidence of retinoblastoma during the most recent study period (2006–2010) by birth cohort analysis may be not fully reliable, because we could ascertain the retinoblastoma patients only until 31 December 2010. 

Historically, the incidence of retinoblastoma was not different between sexes, ^6,8,14 and we reassured the near-equality of retinoblastoma incidence across the 18-year study period. Although the result of the period of 1996 to 2000 showed the significant difference between the sexes in our study, it might be caused by the rapid changes in number of live births during the study period and, probably, the inaccurate population at risk of that period. 

The observed 5-year survival rate for the patients diagnosed in 1993 to 2001 was worse than that reported in developed countries studied during the same period. ^17,18,23 After 2001, the observed 5-year survival rate in Korea significantly increased to 95.5% for those diagnosed in 2001 to 2010 and was comparable to that reported by recent studies. ^17,18,23 The survival rate for retinoblastoma in Korea parallels the overall improvement of the medical system and the infant/pediatric care, both in terms of the advances in treatment modalities and in recognizing retinoblastoma during the examination of infants and children. Systemic chemotherapy and adjuvant treatment were induced widely in late 1990s in Korea, ^15,24 which is the consensus recommendation and current practice. ¹ In addition, a retinoblastoma study group of Korean ophthalmologists was established in 1995, which included most Korean retinal/pediatric clinicians specialized in retinoblastoma. ¹⁵ Their efforts have greatly outpaced improvements in recognizing retinoblastoma, not only by the general public, but also by primary care providers, including pediatricians and ophthalmologists. Although there is no official guideline/regulation regarding the retinoblastoma care in Korea, we could surmise the current practice and trend of retinoblastoma care in Korea from the recent clinical studies conducted in Korea. ^2,25–28  

The present study has several limitations. The KCCR database is restricted to the following information regarding patients with retinoblastoma: birthdate, sex, date of diagnosis, date of follow-ups in hospitals, and date of death. The KCCR does not have any of the following information: unilateral case or bilateral case, heritability and medical records (enucleation, local treatment, chemotherapy, and second malignant neoplasm). Therefore, we could not assess the relationship between retinoblastoma and its well-established risk factors. The KCCR system may undergo improvements in the near future. Another limitation is that we could not estimate the effect of immigration/emigration on the incidence estimate, because of the inherent limitation of the KCCR database and the lack of immigration/emigration data during the study period. However, the immigration/emigration may not have a significant effect on the retinoblastoma incidence estimates in Korea. This is because that the number of immigration and emigration has been quite small compared to the number of Korean population, approximately annually 15,000 to 30,000 persons each. In addition, the KCCR identified the patients having the right of permanent residence in Korea, the foreign retinoblastoma patients who came to Korea for treatment of retinoblastoma, would not be counted. 

In conclusion, the present study demonstrated the incidence of retinoblastoma and the survival rate of patients with retinoblastoma in Korea using data from a national cancer registry. The incidence of retinoblastoma and survival rate of patients with retinoblastoma in Korea was similar to that reported in previous studies. The survival rate of patients with retinoblastoma diagnosed in 2001 to 2010 was significantly better than that in 1993 to 2001. 

Acknowledgments

The authors thank the Korea Central Cancer Registry for providing the database and support for the statistical analyses. 

The authors alone are responsible for the content and writing of the paper. 

Disclosure: S.J. Park, None; S.J. Woo, None; K.H. Park, None