Survival analysis of 2003-2005 data from the population-based Cancer Registry in Macao.

AIM: Macao Cancer Registry was established in 2003. It is population-based and has been collecting cancer reports from all possible settings where pathological and management services are available. To get a better idea over the prognosis and survival of all and major cancer sites, a survival analysis was here performed to estimate the relative survival rates of cancers diagnosed and registered during 2003 to 2005 with a follow-up of vital status till 31 Dec, 2008. METHODS: 3,244 cancer cases diagnosed and registered during 2003-2005 in Macao Cancer Registry were considered for analysis. Cases of in-situ carcinoma, extreme age and poor data quality were deliberately excluded, leaving 2,623 newly diagnosed cancers eligible. Vital status of registered cases through 31 December 2008 was confirmed by matching with death certificates and review from the Hospital Information System (HIS) of the only public hospital. Observed survival rates were calculated using a Life Table method, and relative survival rates were examined using an algorithm written in SAS by Paul Dickman with minor adaptations. Apart from general relative survival rates, specific rates by sex and age strata were also estimated. RESULTS: 3-year and 5-year relative survival rates of all cancers were 61% and 56% respectively for both sexes; (54% and 47%, respectively, for males and 68% and 64%for females). The 3-year relative survival rates for major cancer sites ranged from 21% to 90%, with lung cancer showing the lowest and female breast cancer the highest. 5-year relative survival rates for major cancer sites ranged from 18% to 85%, with liver cancer showing the lowest and again female breast cancer the highest. Female cancer patients had higher relative survival than males across the 5-year follow up period, with a sex difference of nearly 15%. CONCLUSION: Comparison of survival rates from this first trial in Macao, deriving survival statistics from population-based cancer registration, with other Asian countries/cities, like Taiwan, Singapore and Japan, showed Macao and Taiwan to have the closest estimates for 3-year relative survival. Random variation was found to exist in the stratification of sex and age in certain cancer sites due to scarce case numbers in the subgroups. It is important to note that the 3-year survival rates are relatively more consistent and reliable than 4-year or 5-year ones. Promotion of reporting cancer stage by physicians as well as improvement in data quality of cancer registration are essential to allow further informative statistics derived from the cancer registry with reference to cancer prevention.