Japanese Journal of Clinical Oncology Advance Access originally published online on July 26, 2006
Japanese Journal of Clinical Oncology 2006 36(9):602-607; doi:10.1093/jjco/hyl068
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
© 2006 Foundation for Promotion of Cancer Research
Epidemiology Note |
Survival of Cancer Patients Diagnosed between 1993 and 1996: a Collaborative Study of Population-Based Cancer Registries in Japan
Department of Cancer Control and Statistics, Osaka Medical Center for Cancer and Cardiovascular Diseases, Osaka, Japan
For reprints and all correspondence: Hideaki Tsukuma, Department of Cancer Control and Statistics, Osaka Medical Center for Cancer and Cardiovascular Diseases, 3-3, Nakamichi 1-chome, Higashinari-ku, Osaka 537-8511, Japan. E-mail: tukuma-hi{at}mc.pref.osaka.jp
Received March 9, 2006; accepted May 26, 2006
 |
Abstract |
|---|
 TOP Abstract INTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Background: Survival of cancer patients has been measured only in some limited areas in Japan until recently. The purpose of the present study was to collect data of fairly high quality on the population-based cancer registries and to estimate relative 5-year survival of cancer patients in Japan.
Methods: We requested 11 population-based cancer registries within the research group to submit individual data of the patients diagnosed from 1993 to 1996, together with the prognosis after 5 years, to the collaborative study secretariat. Ten population-based cancer registries (Miyagi, Yamagata, Niigata, Chiba, Kanagawa, Fukui, Aichi, Osaka, Tottori and Nagasaki) then accepted data submission (373 000 data). Among 10 registries, only 7 registries met the required standards for the quality of registration data and prognosis investigation. The relative 5-year survival calculated by pooling 279 000 data from seven registries was taken as the national estimate of that of cancer patients in Japan.
Results: The relative 5-year survival was 53.6% for all cancers (males: 49.2%, females: 59.4%); the survivals of stomach, large bowel, prostate and kidney cancer patients were from 62 to 68%; those of breast, uterus, larynx, skin, testis, bladder and thyroid cancer patients were from 74 to 92%; those of liver, gall bladder and bile duct, pancreas and lung cancer patients ranged from 6 to 23%.
Conclusion: On the basis of the data from seven population-based cancer registries in Japan, we calculated the relative 5-year survival of cancer patients diagnosed from 1993 to 1996 for the first time.
Key Words: survival • cancer • registry
|
INTRODUCTION |
|---|
|
TOPAbstractINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONReferences |
|---|
The survival based on population-based cancer registries, being different from the survivals of the cancer patients diagnosed and treated at specific facilities and departments, is a prerequisite for the designing of projects and the evaluation of measures and treatments for cancers, because only the former is capable of providing patients' data without bias. However, in Japan, the survival has been calculated only in limited areas until now. In addition, regarding the survival calculation, study subjects and methods have not yet been standardized enough. Referring to the EUROCARE study (1), the authors tried to develop the standard methods of calculating survival in Japanese registries through the collaborative study of population-based cancer registries since 1996. In 1998, we proposed the standard methods that required that the vital status of patients be confirmed by inquiring to the resident registration at the time of 5 years after diagnosis, and reported the results of relative 5-year survival based on the data (stomach, lung and breast cancer diagnosed from 1985 to 1989) from cancer registries of Yamagata, Fukui and Osaka prefectures (2), which had already started to collect data satisfying the above criteria (3). In 2001, we collected, from 12 registries belonging to the study group, individual data of all cancer patients (for all sites) diagnosed in 1993 with prognosis after 5 years, and tried a nationwide estimate of relative 5-year survival according to the standard methods (4). A nationwide estimate, however, has not been completed because there were differences in the quality of registration and prognosis investigation among 12 registries.
In this study, we requested population-based cancer registries to submit the data of the patients (for all sites) diagnosed from 1993 to 1996 with those on prognosis after 5 years, and pooled the data of the cancer registries that achieved a standard of quality of data in terms of both registration and prognosis investigation in order to estimate relative 5-year survival of cancer patients in Japan.
|
SUBJECTS AND METHODS |
|---|
|
TOPAbstractINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Ten of 11 registries belonging to the research group traced the prognosis after 5 years of patients diagnosed from 1993 to 1996 and submitted their individual data (a total of 373 000 cases) to the research group secretariat. Among 10 registries, 7 registries (Miyagi, Yamagata, Niigata, Fukui, Osaka, Tottori and Nagasaki) met the required standards for the quality of registration and prognosis investigation. According to a total of 279 000 data provided by these registries, we calculated survivals and considered them as nationwide estimates. Standard for the quality of registration was based on the standard adopted in the nationwide estimates of incidence: DCO (death certificate only cases) was <25%, or DCN (death certificate notification) was <30%, and ID ratio (incidence to death ratio) was not <1.5 (5). As far as the standard for the quality of prognoses investigation was concerned, we set two kinds of criteria according to the follow-up methods. For the registries checking survival of patients by referring to inhabitant's registry, the proportion of prognosis-unknown cases 5 years after diagnosis was <5% (Yamagata, Fukui and Osaka). For the registries having no confirmation of survival 5 years after diagnosis, information on personal identification including names would be magnetized in order to collate their registered patients with death information in high accuracy. Therefore, using these criteria, it was guaranteed that they had rather accurate information about death (Miyagi, Niigata, Tottori and Nagasaki).
The method of calculating survival is based on the EUROCARE study (1) in which 12 countries of the European Union had participated in 1990, and our study group also followed this method. In other words, we excluded DCO cases, in situ cancer cases and mucosal cancer cases of large bowel from the analysis. In the case of multiple cancers, only the first-diagnosed tumour was analysed. In calculating survival, cumulative 5-year survivals were calculated starting from the date of diagnosis. Expected survivals were calculated using the cohort survival table based on life tables of the Japanese population and afterwards using the survival probability in the general population similar to the patients in sex, birth-year and age. The former was divided by the latter to obtain relative 5-year survivals. Besides, the cases that were unknown as of 5 years after diagnoses were dealt with as alive as of the last date of living (2). However, regarding four registries that had not yet started to check the survival of patients by referring to the resident registry, we regarded all of the cases whose death was not confirmed as alive until 5 years, and survivals were calculated.
|
RESULTS |
|---|
|
TOPAbstractINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Table 1 shows the number of incidence, validity indices of the registration and the number of study subjects for survival analysis, according to the registry. Total number of incidence was 279 469, and the following cases were excluded from the survival analysis: DCO (49 278 cases, 17.6% of the total incidence), subsequent primary tumours (18 596 cases, 6.7% of the total), not malignant tumours (487 cases, 0.2% of the total), in situ cancers (3955 cases, 1.4% of the total). In addition, excluding the cases of unknown age at diagnosis and the cases over 100 years old, we analysed the rest of all (209 373 cases, 74.9% of the total) (Subjects 1). Moreover, for DCN cases, complementary cancer reports were requested in Yamagata, Fukui and Osaka prefecture, and the registry records of cases originating from death information were distinguished in Miyagi prefecture. The number of cases in which we traced the death information to incidence were 17 556 (8.4% of the total). The analysis subjects (Subjects 2) excluding these cases were 191 817 cases (68.6% of the total).
|
Table 2 shows the vital status as of 5 years from diagnosis. In Yamagata, Fukui and Osaka Cancer Registries, where the vital status of patients was checked after 5 years by referring to inhabitant's registry, the proportion of unknown cases for vital status was <2–3%, which indicates that the prognosis investigation was highly accurate.
|
Table 3 shows the number of study subjects, relative 5-year survival and its standard error according to the primary site. In addition, when selecting the study subjects, we distinguished between the occasion including the cases regarded as incidence according to death information (Subjects 1) and the occasion excluding the former case. Moreover, regarding the sites covering relatively many subjects, we showed the results of analysis by sex; regarding the sites covering relatively small subjects, we showed only the total results of males and females. Usually, mucosal cancer of large bowel cases should be excluded from the survival analysis (2,6), but some registries (Fukui, Niigata and Tottori) submitted the data including these cases undistinguished. Therefore, regarding all sites, large bowel, colon and rectum, each case was divided between their 1 and 2 on the basis of whether or not mucosal cancer of large bowel was included.
|
Relative 5-year survival of cancer patients of all sites (all sites 2) was 45.1, 54.8 and 49.2% for males, females and total, respectively, on Subjects 1, and 49.2, 59.4 and 53.6% on Subjects 2; the former is higher than the latter. This is because Subjects 1 includes the cases regarded as incidence according to death information. Same as many registries in Japan, under the situation of high proportion of the cases not reported and the cases registered on the basis of death certificates, it will be possible for the survival calculated on the basis of Subjects 1 to be estimated lower than it is. In contrast, it is also possible for survival to be estimated higher than it is on Subjects 2. In Japan, each population-based registry decides whether or not to request medical institutions to submit information on cancer diagnosis and treatment for DCN. Consequently, if we need domestic comparison of survival, the survival of Subjects 2 would be better than that of Subjects 1 in terms of comparability. Therefore, in this study, we will regard the survival calculated on the basis of Subjects 2 as that of cancer patients in Japan.
The relative 5-year survivals by site on Subjects 2 were calculated as follows: the survivals of cancers of stomach, large bowel, prostate, and kidney were 62–68%; those of cancers of breast, uterus, larynx, skin, testis, bladder and thyroid were 74–92%; those of cancers of liver, gallbladder and bile duct, pancreas and lung were 6–23%.
Table 4 shows observed and relative 1- to 5-year survival by sites. Observed 1- to 5-year survival in all sites 2 were 72.5, 61.1, 54.9, 50.6 and 47.5% in order; relative survival were 74.4, 64.2, 59.0, 55.7 and 53.6% in order.
|
Table 5 shows relative 5-year survivals for major sites of cancer (all sites 2, stomach, large bowel 2, liver, lung, breast and uteri) by sex and age at diagnosis. The relative 5-year survivals for cancers of stomach, liver, lung and uterus decreased markedly in old age; however, the difference was not pronounced in age regarding those for cancers of large bowel 2 and breast. Regarding lung cancer, females had a higher survival than males in all age groups. Besides, the same difference was also observed for the survival of liver cancer, where there was a marked difference in sex. In contrast, females had a lower survival than males in the following sites and ages: the young patients of stomach cancer aged from 15 to 44 years old, and the old patients of large bowel cancer aged over 75 years old.
|
|
DISCUSSION |
|---|
|
TOPAbstractINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONReferences |
|---|
On the basis of the data from seven population-based cancer registries in Japan that achieved a standard of quality of data in terms of both registration and prognosis investigation, we calculated relative 5-year survival of cancer patients for the first time. On the basis of the standard methods of calculating survival employed in EUROCARE study (1), we estimated the Japanese representative survival for major sites of cancer as well as childhood cancers diagnosed before the age of 15 years. This study estimated the survival not only for the cancers including the followed-back cases from DCN (Subjects 1) but also the cancers excluding them (Subjects 2). The former used the same method as EUROCARE study employed, and it is the estimate that should be utilized for international comparison of survival based on population-based cancer registries. In contrast, the latter is the estimate that should be utilized for domestic comparison of survival in Japan where some registries do not conduct follow-back inquiries according to death information.
Now we would like to discuss the issues for the study in the future. First of all, it is important to improve the quality of registration, because the high proportion of patients not registered will degrade the accuracy of survival estimate. In this study, we required each registry to meet the necessary standards for participating in the nationwide estimates of incidence (5). Therefore, it would be reasonable to assume that the survival in Japan has been calculated on the basis of the fairly accurate data of population-based cancer registries for the first time. However, from the viewpoint of international standards, where the registry data showing DCO >10% for all sites are regarded as poor completeness of registration, the cases not registered are still not negligible in Japan (7). Thus, we have to admit that the validity of survival has not improved to the desired extent. Moreover, it should be taken into consideration that the survival calculated in this study was based on the data submitted by the very limited areas of seven prefectures. If we assume the difference among prefectures, it will be desirable to utilize wider range of data from more prefectural regions.
In the three prefectures where the vital status of patients was checked after 5 years from diagnosis, the proportion of unknown cases for vital status was 3% or less, which implies that the prognosis investigation was highly accurate. However, the other four prefectures did not check the vital status of patients. The fact that they do not check the survival of patients would have a relatively small effect on the overestimation of the survival, because it is estimated that collating with death information can be done with high accuracy in these four prefectures, and that frequency of moving out to different prefectures is relatively low. However, for collecting more accurate data of survival, it is necessary from now on to investigate prognosis of patients by referring to resident registry.
Regarding mucosal cancers of large bowel, we should have excluded them from the survival analysis, since they were regarded as in situ cancers according to the agreement of UICC (6). However, some population-based cancer registries did not distinguish them in this study. Therefore, we calculated the following two cases: the case including mucosal cancer of large bowel (all sites 1, large bowel 1, colon 1, rectum 1) and the case excluding distinguished mucosal cancer (all sites 2, large bowel 2, colon 2, rectum 2). Moreover, it was not easy for some population-cancer registries to distinguish multiple cancers. For more reliable results of survival, it is necessary to distinguish them from other cancers. In this study, however, it seems that the proportion of mucosal cancer of the large bowel and that of multiple cancers except the first-diagnosed tumour were not very large; therefore, it is reasonable to think that they did not affect the result of survival that much.
The EUROCARE study started as a collaborative study of the European Union (1), and it currently involves 67 population-based cancer registries operating in 22 European countries (8). Furthermore, the CONCORD study (9) extends the EUROCARE study to include North America (the USA and Canada), Australia and Japan. In Japan, the cancer registries of Yamagata, Fukui and Osaka prefecture currently participate in this study, all of which have already started to check the vital status of patients after 5 years from diagnosis. It is desirable that more registries will take part in this study. In the CONCORD study, the following studies are developed: Phase 1 study based on existing data of registries; Phase 2 study to investigate clinical data retrospectively by taking samplings from databases of population-based cancer registries; and Phase 3 study on the central review of pathological specimens. Currently, it is only the Phase 1 study that Japan is participating in; however, it will be necessary to build up a framework for our participation in the high-resolution study of Phase 2 and 3. Through our participation in these types of international collaborative studies, it is expected that the reliability of the study on the survival based on population-based cancer registries in Japan will be improved.
|
Acknowledgments |
|---|
In 2005, the Research Group conducted collaborative study on population-based cancer survival with contribution from 11 cancer registries: Miyagi (D. Shibuya), Yamagata (T. Matsuda), Chiba (H. Mikami), Kanagawa (N. Okamoto), Niigata (K. Ogoshi), Fukui (M. Fujita), Aichi (H. Ito), Osaka (H.T.), Tottori (T. Kishimoto), Hiroshima City (N. Nishi) and Nagasaki (M. Soda). The study was supported by a Grant-in-Aid for Cancer Research from the Japanese Ministry of Health, Labor and Welfare (14-2).
|
References |
|---|
|
TOPAbstractINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONReferences |
|---|
1 Berrino F, Sant M, Verdecchia A, et al. Survival of Cancer Patients in Europe: The EUROCARE Study. IARC Scientific Publications No. 132, Lyon: IARC 1995.
2 Ajiki W, Matsuda T, Sato Y, Fujita M, Yamazaki S, Murakami R, et al. Standard method of calculating relative survival rates in population-based cancer registries—an investigation using stomach cancer patients. Jpn J Cancer Clin 1997;43:1005–14 (in Japanese with English abstract).
3 Ajiki W, Matsuda T, Sato Y, Fujita M, Yamazaki S, Murakami R, et al. A standard method of calculating survival rates in population-based cancer registries. Jpn J Cancer Clin 1998;44:981–93 (in Japanese with English abstract).
4 Oshima A, Ajiki W, Tsukuma H. Estimation of survival of cancer patients in Japan (preliminary report). In: Tajima K, Kuroishi T, Oshima A, editors. Cancer Mortality and Morbidity Statistics: Japan and the World. Gann Monograph on Cancer Research. 51. Tokyo: Japan Scientific Societies Press 2004;131–5.
5 Research Group for Population-Based Cancer Registration in Japan. Cancer incidence and incidence rates in Japan in 1999: estimates based on data from 11 population-based cancer registries. Jpn J Clin Oncol 2003;33:241–5.
6 Sobin LH, Wittekind CH. TNM Classification of Malignant Tumours, 6th edn. New Jersey: John Wiley & Sons 2002.
7 Parkin DM, Whelan SL, Ferlay J, Teppo L, Thomas DB. Cancer Incidence in Five Continents, Vol. VIII. IARC Scientific Publications No.155, Lyon: IARC 2002.
8 Berrino F. The EUROCARE Study: strengths, limitations and perspectives of population-based, comparative survival studies. Ann Oncol 2003;14(Suppl 5):v9–13.[CrossRef][Medline]
9 CONCORD. Available from: http://www.lshtm.ac.uk/ncdeu/cancersurvival/concord/protocol.htm
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  N. Sato, Y. Ito, A. Ioka, M. Tanaka, and H. Tsukuma Gender Differences in Stomach Cancer Survival in Osaka, Japan: Analyses Using Relative Survival Model Jpn. J. Clin. Oncol., October 1, 2009; 39(10): 690 - 694. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Shibata, T. Matsuda, W. Ajiki, and T. Sobue Trend in Incidence of Adenocarcinoma of the Esophagus in Japan, 1993-2001 Jpn. J. Clin. Oncol., July 1, 2008; 38(7): 464 - 468. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. Tabata, Y. Ohno, R. Matsui, H. Sugiyama, Y. Ito, H. Tsukuma, and A. Oshima Partial Cancer Prevalence in Japan up to 2020: Estimates Based on Incidence and Survival Data from Population-based Cancer Registries Jpn. J. Clin. Oncol., February 1, 2008; 38(2): 146 - 157. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Yoshino, N. Boku, Y. Onozawa, S. Hironaka, A. Fukutomi, Y. Yamaguchi, N. Hasuike, K. Yamazaki, N. Machida, and H. Ono Efficacy and Safety of an Irinotecan plus Bolus 5-Fluorouracil and L-Leucovorin Regimen for Metastatic Colorectal Cancer in Japanese Patients: Experience in a Single Institution in Japan Jpn. J. Clin. Oncol., September 1, 2007; 37(9): 686 - 691. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Ito, Y. Ohno, B. Rachet, M. P Coleman, H. Tsukuma, and A. Oshima Cancer Survival Trends in Osaka, Japan: the Influence of Age and Stage at Diagnosis Jpn. J. Clin. Oncol., July 16, 2007; (2007) hym047v1. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||