Akademik Veri Yönetim Sistemi
ASCRS Baltimore, Annual Scientific Meeting 2024, Maryland, Amerika Birleşik Devletleri, 1 - 04 Haziran 2024, ss.1
The current study aimed to assess and demonstrate whether variations in the data sampling process between the National Cancer Database (NCDB) and the Surveillance, Epidemiology, and End Results (SEER) database lead to differences in patient characteristics and oncologic outcomes. This approach sought to investigate the factors contributing to variations in overall survival rates observed across studies that utilize these databases, and to assist researchers in making informed decisions when choosing the most appropriate database for their particular research questions.Methods/Interventions: Patients who underwent curative surgery for stage I-IV primary colon cancer between 2004 and 2019 were identified from the NCDB (a hospital-based cancer registry) and SEER (a population-based cancer registry) database according to the International Classification of Diseases for Oncology, Third Edition (ICD-O-3) topographic codes. Patients were categorized into eight anatomical subsegments. Our primary outcome was the assessment of the 5-year overall survival (OS) rates in distinct colon cancer locations between the NCDB and the SEER by conducting Kaplan-Meier survival analysis. Subsequently, subgroup analyses were executed for stages I, II, III, and IV. The intent of these sub-analyses was to explore potential discrepancies in OS rates across different stages of colon cancer within the two databases. Our secondary outcome was the comparison of patient demographics, tumor characteristics, and treatment variables for each distinct colonic cancer location by using weighted proportional differences for categorical variables and standardized difference for continuous variables.Results/Outcomes: A total of 1.523.212 primary colon cancer patients (NCDB, 1.116.041; SEER, 407.171) were identified. Overall, 777.827 patients (NCDB, 572.196; SEER, 205.631) were eligible for our study. The largest difference in tumor stage distribution between the databases was the frequency of sigmoid colon cancer excluding the appendix (weighted proportional difference, 1.63%; 95% confidence interval, 1.21%-2.05%). In most colonic locations, 5-year OS rates were slightly lower for the NCDB cases compared with those in SEER (ascending colon: NCDB 62.7%, SEER 63.1%; transverse colon: NCDB 61.4%, SEER 62.1%; descending colon: NCDB 65.0%, SEER 65.2%; sigmoid colon: NCDB 67.4%, SEER 67.7%) Conversely in some colonic locations, the 5-year OS rates were slightly higher for the NCDB cases compared with those in SEER. (cecum: NCDB 59.8 %, SEER 59.5%; appendix: NCDB 61.1%, SEER 61.0%; hepatic flexure of colon: NCDB 62.1%, SEER 61.7%; splenic flexure of colon: NCDB 62.7%, SEER 62.6%). Cecum cancer displayed the worst 5-year OS rates in all stages, while sigmoid colon cancer exhibited the highest 5-year OS rates in stages 1, 2 and 3; and the appendiceal cancer had the highest 5-year OS rates in stage 4 in both datasets.Conclusion/
Discussion: The analysis of stage I-IV primary colon cancer patients in each database reveals that despite variances in sampling techniques, the NCDB and the SEER database demonstrate remarkable consistency in 5-year overall survival rates across different stages of distinct colonic cancer locations."Stage I Overall Survival Rates of NCDB and SEER Database.jpg)
NCDB: National Cancer Database, SEER: Surveillance, Epidemiology, and End Results
Stage II Overall Survival Rates of NCDB and SEER Database
NCDB: National Cancer Database, SEER: Surveillance, Epidemiology, and End Results
Stage III Overall Survival Rates of NCDB and SEER Database
NCDB: National Cancer Database, SEER: Surveillance, Epidemiology, and End Results
Stage IV Overall Survival Rates of NCDB and SEER Database
NCDB: National Cancer Database, SEER: Surveillance, Epidemiology, and End Results"