Childhood Cancer Incidence and Survival 1985-2009 , Khon Kaen , Thailand

In Thailand, 21% of the population is under 15 (National Statistics Office, 2003) and although cancers among this age group are rare, ~1,000 were diagnosed with cancer based on the expected incidence (Vattanasapt, 1999; Wiangnon et al., 2003; Thai Pediatric Oncology Group, 2011). Between 2007-8, the respective cancer related mortality rates in Thai children 6-12 and 13-18 was 2.16 and 2.13 per 100,000 (Sutra et al., 2009). The previous population-based incidences of childhood cancer in Thailand were calculated by extrapolating the incidences from the data, which was limited because of the small number of cancer registries. Recently, the Thai Pediatric Oncology Group (ThaiPOG) reported on the country incidence between 2003 and 2005; but survival was thus limited to children registered during this period.


Introduction
In Thailand, 21% of the population is under 15 (National Statistics Office, 2003) and although cancers among this age group are rare, ~1,000 were diagnosed with cancer based on the expected incidence (Vattanasapt, Surapon Wiangnon 1 *, Arunee Jetsrisuparb 1 , Patcharee Komvilaisak 1 , Krittika Suwanrungruang 2 In addition, the previously reported survival of childhood cancer patients was described with respect to specialized hospitals and/or institutions. The aim of the current investigation was to perform a long-term population-based retrospective study of the incidence of childhood cancer in a homogeneous population in Khon Kaen province. Registration area and registry procedures: Khon Kaen province is located in northeastern Thailand. The province covers an area of 10,886 km 2 with a population of 1.7 million (Figure 1). The Khon Kaen Cancer Registry (KKCR) was established in 1984 at the Faculty of Medicine, Khon Kaen University. The KKCR has been conducting both hospital and population-based registration.
The data on cancer patients was collected using both active and passive methods from all hospitals and private clinics in the province. The registrars regularly visit all registered sites. The medical charts of patients diagnosed with cancer were reviewed and the cancer data registered by registry personnel. Any questionable cases were traced to the original resource of information for clarification. Multiple primary cases were checked by physicians using their original records and with any other physicians concerned. Death certificates with a mention of cancer were collected from the population statistics of the office of the Ministry of Interior. The vital status of registered cases has been regularly updated by checking with the population statistics of the office of the Ministry of Interior. To prevent duplication, all data were reviewed and checked with existing registry files before data entry. All the data were verified, checked for duplication coded and entered into the CanReg4 software.

Materials and Methods
Data on all childhood cancers were retrieved from the KKCR data set for the period 1985-2009. Cancer sites, morphology and behavior were coded according to the International Classification of Diseases for Oncology-3rd edition (ICD-O-3) (WHO, 2000). Tumor types were grouped according to the International Childhood Cancer Classification (ICCC) (Kramarova et al., 1996). Both histologically-verified and non-verified patients were included. All patients were observed until the end of December 2012. The aged-standardization rate (ASR) was performed using a direct method with groups 0-14 years (0-4, 5-9 and 10-14 years) (Parkin et al., 1998). ASRs were calculated by the direct method, using the world standard population and expressed per million persons. ASRs for five-year age groupings (viz., 1985-1989, 1990-1994, 1995-1999, 2000-2004,2005-2009) were calculated for both boys and girls (Jensen et al., 1991). Five-year cumulative survival rates were calculated using the Kaplan-Meier method. The Generalized Linear Model (GLM) method was applied for analysis based on incidence rates, corresponding p-values.

Population denominators
The population denominators used for calculation of the incidence rate were (a) the numbers of children from the population censuses in 1990 and 2000 and (b) the estimated projected numbers based on data of 2010 (National Statistic Office, 2003). This provided an annual estimate by age group and sex. Notably, the child population has been continuously decreasing since1980s.
Most of the pediatric cancers were histologicallyconfirmed (83.7%) but the proportion varied between 52.4% for cancer of the central nervous system (ICCC group III) and 99.7% for leukemia (ICCC group I).
The age-standardized rate  was 83 per million. The incidence varied between 58.1 and 99.3 per million according to the 5-year interval. The time trend analysis revealed that childhood cancers increased in boys and girls (9%, p=0.004 and 7% p=0.05) (Table 3, Figure 2). Leukemia was the most common cancer (N=360, ASR 33.8) followed by neoplasms of the central nervous system (CNS) (N=150, ASR 12.8) and lymphoma (N=79, ASR 7.0). (Table 3) Among the leukemias, acute lymphoblastic leukemia (ALL) was more common than acute non-lymphoblastic leukemia (ANLL) (ratio, 3.9:1). ALL was more common among boys than girls (ASRs 31.1 vs 20.3) while the incidence of ANLL between sexes was not different in both sexes. Non-Hodgkin lymphoma (NHL) is more common than Hodgkin (HD) and Burkitt's lymphoma (ASRs 3.08, 0.82, 1.34, respectively). All types of lymphomas are more common among boys than girls.   .2014.15.18.7989 Childhood Cancer Incidence and Survival 1985-2009 ICCC group 1985ICCC group -1989ICCC group 1990ICCC group -1994ICCC group 1995ICCC group -1999ICCC group 2000ICCC group -2004ICCC group 2005ICCC group -2009ICCC group 1985ICCC group -2009 Leukemias, myeloproliferative diseases and others 28 27.

Survival
The probability of 5-year overall survival (OS) was 52% (95%CI: 0.48-0.55). Survival differed by type of cancer. The best outcome with highest overall survival was for germ cell tumors (74%) while the worst outcome was for sympathetic tumors (neuroblastoma) (25%) ( Table 4).
The respective 5-year overall survival for (1) germ

Discussion
This is the first study in Thailand to survey the incidence of all types of childhood cancer over an extended period from a unique population in Khon Kaen province. The microscopic verification (MV) of all cancers was satisfactorily high (83.7%) even though the MV for CNS tumors was limited (52.4%). The age-standardized rate (ASR, 83 per million) was in the range of those extrapolated from the population-based incidences of 4 registries between 1988-1994 (ASR 70.7) (Vattanasapt, 1999) and 1995-1997 (Wiangnon et al., 2003). The incidence reported by the ThaiPOG for the whole country survey between 2003-5 was 74.9 (Wiangnon et al., 2011). The respective incidence of childhood cancer in Khon Kaen between 1988-1991and 1985-2000 and 73 per million (Sriamporn et al., 1996;. These incidences were lower than the incidence in this study, which may reflect under-coverage; especially during the early registration era. Notwithstanding the longterm registration, the incidence was still lower than rates reported for western countries albeit comparable to other Southeast Asian countries (Parkin et al., 1998).
As in childhood cancer elsewhere, leukemia is the most common cancer, for which the incidence in Khon Kaen is similar to those of resource-rich countries (Ries et al., 1999). Notably, the incidence of brain tumor was low even though the neuroimaging equipment in Thailand has not been a limiting factor during the last 20 years. Some patients, however, may not have been fully investigated using this modality, resulting in under diagnosis of CNS tumors. Microscopic verification was also rather low for this group of tumors.
Bone tumor incidences were low compared to those of western countries and resource-rich countries in Asia. Generally, bone tumors are more common in adolescents and young adults and we limited the age of children up to 15. Lymphoma incidence was also low without a clear explanation.
Childhood cancer is becoming an important cause of death in Thailand as poverty-related causes are declining (Sutra et al., 2009). Over the past two decades, childhood cancer mortality has declined dramatically in resource rich countries (Ries et al., 1999;Gatta et al., 2005). The decrease over this period is partially due to the effects of (a) understanding tumor biology and (b) improvements in diagnosis and therapy (Redner, 2011).
Overall survival (OS) of all childhood cancers in the current analysis was lower than those of developed countries; however, it was comparable to population-based survival in Osaka (Ajiki et al., 2004), Thailand (52 vs 54.9) (Wiangnon et al., 2011) and in Khon Kaen at a university hospital . The study of survival of childhood cancers in a population-based in Thailand was unable to identify any trends because of the short duration of the study (i.e., between 2003-5).
Mortality from childhood cancer between 1975-1995 in the USA decreased by 40% (Ries et al., 1999). Indeed, the outcome has been improving for all types of cancer, especially for leukemia and lymphoma. The five-year survival of lymphoid leukemia has surpassed 80% in the Nordic countries, Western and Southern Europe, the United Kingdom, and West Germany (Gatta et al., 2005). Our outcomes for leukemia remain low, especially for ANLL. The survival of ALL patients has been gradually increasing from 35% (1985-95) to 70% (2006-9). The survival of ANLL patients remains inferior (21-45%).
Abandonment of chemotherapy was the most common cause of treatment failure, which was strongly related to poor socioeconomic status and financial support. In 2003, the National Security Office (NHSO) implemented universal health coverage nationwide and since then adherence to intensive cancer treatment has significantly risen. In 2006, the NHSO provided and supported a disease management program for the treatment of childhood leukemia and lymphoma. Common treatment protocols were used nationwide. In principle, risk stratification and monitoring of leukemia diagnosis and treatment was obligatory using flow cytometry. In addition, the use of chemotherapy also increased over the period together with overall improvement of supportive care, which resulted in the improvement of ALL survival to 73% (Seksarn et al., 2011). Notwithstanding, the survival of ANLL was not satisfactorily improved at the national level (Seksarn et al., 2011). Nowadays and importantly, the quality of life is also evaluated for these cancer survivors (Pakakasama et al., 2010).
The two referral-hospitals in Khon Kaen province have pediatric oncologists treating childhood cancers. According to the Thai national referral system, all cases Asian Pacific Journal of Cancer Prevention, Vol 15, 2014 7993 DOI:http://dx.doi.org/10.7314/APJCP.2014.15.18.7989 Childhood Cancer Incidence and Survival 1985-2009 residing in Khon Kaen must be referred to one of these 2 centers, in which institutional protocols are initially used. In 2000, many treatment protocols in our institution were revised and developed by modifying the standard international chemotherapy protocol. These were used until the national protocol was implemented in 2006. Survival of ALL was significantly improved during this change.
Recently, the Thai Pediatric Oncology Group proposed common protocols for nationwide use. Improvement of survival and quality of life for children with cancers including survivors are expected (ThaiPOG, 2014).
In conclusion, the incidence and survival rates of childhood cancer in Khon Kaen are lower than those reported for western countries. Survival is, nevertheless, increasing over time after implementation of national protocol and improvement of supportive care facilities. Quality of life of childhood survivors needs to be studied systematically..