Using Period Analysis to Timely Provide Serial Data on Long-Term Survival for Gastric Cancer Patients From Taizhou, Eastern China

Gastric cancer, the fifth most common cancer worldwide, is particularly heavy in China, accounting for 39.5% of global deaths, reported by the International Agency for Research on Cancer (IARC). While timely evaluation of long-term survival of gastric cancer is crucial for guiding early detection screening programs, data on long-term relative survival (RS) for gastric cancer are scarce in China. Traditional methods for estimating 5-year RS require complete 5-year follow-up data, leading to delays of at least 5 years for survival estimates—a challenge that similarly affects 10-, 15-, and 20-year RS. The period analysis approach, introduced by Brenner, overcomes this limitation by enabling more current survival estimates without requiring complete follow-up data [1]. This method is considered the “gold standard” for assessing cancer survival using data from population-based registry and is widely applied in Western countries. In the current study, we aimed to provide, using period analysis, the most recent serial data on long-term survival for gastric cancer patients from a Chinese population, including 5-, 10-, 15-, and 20-year RS.

Our data came from four high-quality cancer registries in Taizhou City, ranked fifth in population size in Zhejiang Province, China. These registries were selected based on the criterion that death certificate-only (DCO) cases should account for less than 13%, as recommended by Brenner [1], which there were nine registries included initially, to ensure data reliability. Aligned with the International Classification of Diseases (ICD), flowchart of gastric cancer patients included in the analyses was shown in Figure 1A. Eventually, we include the following cases for further analyses, that is, 12,270 cases during 2004–2023 for 20-year RS, 10,515 cases (2009–2023) for 15-year RS, 7427 cases (2014–2023) for 10-year RS, and 3534 cases (2019–2023) for 5-year RS. Period analysis was used to estimate age-adjusted 5-, 10-, 15-, and 20-year RS, stratified by gender, region, and age at diagnosis. RS for patients was calculated as the ratio of observed survival divided by the expected survival based on the life tables using the Ederer II method. All survival estimates were age-adjusted RS using world population standardization for age weights. More details on data analyses are provided in the Supplementary Information S1.

The basic characteristics of gastric cancer patients were shown in Figure 1B. Approximately 69.0% of patients were males for all four intervals. Furthermore, the majority of cases stood in ≥ 45 years group (approximately 96%), and patients ≥ 65 years accounted for more than half of all gastric patients during all four intervals. As shown in Figure 1C, the most up-to-date estimates of serial data on long-term survival for gastric cancer patients from Taizhou, eastern China, reaching 58.59% for 5-year RS during 2019–2023, 53.04% for 10-years (2014–2023), 46.93% for 15-year RS (2009–2023), and 43.19% for 20-year RS (2004–2023). We found that women always had higher survival estimates compared to men, urban areas consistently had higher survival than rural areas, which were persistent for 5-, 10-, 15-, and 20-year RS. Additionally, we found a clear age gradient for survival estimates stratified by age at diagnosis, for instance, 5-year RS declined from 72.15% for age at diagnosis < 45 years to 30.52% for age at diagnosis > 74 years. Survival decreased with the years following diagnosis, until reaching a plateau, furthermore, plateaus in trends of 20-year RS were observed at 16–17 years after diagnosis in subgroups of gender, region, and age at diagnosis (Figure 1D), except for rural patients (plateau occurred at 19 years after diagnosis), indicating that long-term cases could be effectively treated in Taizhou.

Our finding of the 5-year RS reaching 58.59% during 2019–2023 in Taizhou City was lower than Japan (60%) during 2010–2014, Korea (69%) during 2013–2019 [2] and Shanghai urban areas (71.0%) during 2012–2015 [3], but much higher than Nantong rural areas (31.48%) during 2012–2017 [3] and the previous report of 35.1% during 2012–2015 using data from 17 Chinese cancer registries [4]. Our finding is plausible due to the following reasons. First, it is common sense that the survival of gastric cancer patients shall be improved along with medical progress and treatment improvement in recent years. Second, the increase in survival is also associated with population-based endoscopic screening programmes aimed at the early detection of gastric cancer. For instance, “Early detection and early treatment for cancers in urban areas” began in 2012, including Zhejiang Province, while opportunistic screening at medical facilities for upper gastrointestinal cancer began in 2019, covering Taizhou city. Third, the previous report of 35.1% during 2012–2015 using data from 17 Chinese cancer registries was actually projected (underestimated) [4] because our group has verified that period approach, for first time for a Chinese population, can provide more timely and precise estimates for long-term survival of cancer patients, compared to traditional cohort approach [5].

Our study has several strengths and limitations. First, this is the first time in China using the period analysis to provide the most recent serial data of long-term survival for patients with gastric cancer from Taizhou, eastern China, not only 5-year RS but also for 10-, 15-, and 20-year RS. What we provided would fill a critical gap and benefits to guiding the second prevention policies. Second, utilizing a population-based design helped prevent selection bias, and, age standardization improved data comparability and value by adjusting for differences in age distributions, ensuring more accurate comparisons between groups. Nevertheless, the lack of detailed histological and pathological staging data, along with medical records and progress notes (e.g., family history, lifestyle factors) that could be matched to individual patients, limits a deeper analysis of prognostic and susceptibility factors. This is a common issue existed in cancer registries worldwide. In addition, we are currently applying for the collection of data from the entire Zhejiang Province to further promote our research to serve for second prevention.

In summary, we provided that, for the first time in China using period analysis, the most up-to-date estimates of serial data on long-term survival for gastric cancer patients from Taizhou, eastern China, reaching 58.59% for 5-year RS during 2019–2023, 53.04% for 10-years RS during 2014–2023, 46.93% for 15-year RS during 2009–2023, and 43.19% for 20-year RS during 2004–2023. Our findings indicate promising survival estimates for gastric cancer patients from eastern China, which is crucial for guiding health policy-makers to timely assess early detection and screening programs for gastric cancer. Nevertheless, further improvements on long-term survival of gastric cancer patients from China are highly warranted, which may require a comprehensive model that integrates data on prevention, screening, diagnosis, treatment, and follow-up care. For instance, systematical collection data on detailed clinical information, including tumor stage, histology, and treatments approaches are highly warranted for stratified survival analyses, which is crucial for guiding health policy-makers to timely assess early detection and screening programs for gastric cancer.

Wei Gu: visualization (equal), writing – original draft (equal), writing – review and editing (equal). Liangyou Wang: data curation (equal), writing – review and editing (equal). Yongran Cheng: data curation (equal), formal analysis (equal), investigation (equal), methodology (equal), software (equal), validation (equal), writing – review and editing (equal). Tianhui Chen: conceptualization (equal), data curation (equal), funding acquisition (equal), methodology (equal), project administration (equal), resources (equal), supervision (equal), visualization (equal), writing – original draft (equal), writing – review and editing (equal). Jinfei Chen: funding acquisition (equal), supervision (equal), writing – review and editing (equal). All authors have read and approved the final manuscript.

Although the data from nine cancer registries in Taizhou City, eastern China, were completely anonymized and their usage did not entail ethical problems, the utilization of these data for this study was approved by the Institutional Review Board of Zhejiang Cancer Hospital, China (IRB-2023-324).

The authors declare no conflicts of interest.