A Methodological and Survival Comparison of NCDB and SEER Database for Colon Cancer Research.

Abstract

Background: The National Cancer Database (NCDB) and Surveillance, Epidemiology, and End Results (SEER) database are widely used in colon cancer research, particularly for analyzing overall survival (OS) rates. However, differences in demographics, treatment patterns, and survival outcomes across colon cancer stages and locations between these databases remain incompletely understood. Addressing these disparities is crucial for researchers when selecting the most appropriate registry for survival analysis.

Objectives: This study aims to systematically compare patient characteristics, oncologic outcomes, and OS rates between NCDB and SEER across various tumor locations and disease stages in colon cancer.

Methods: We analyzed patients undergoing surgery for Stages I-IV primary colon cancer (2004-2019), comparing patient characteristics, oncologic outcomes, and OS rates across distinct tumor locations and cancer stages in NCDB and SEER. Our objective was to assess how differences in database structure and sampling methodologies influence reported survival outcomes.

Results: The study included 777 827 patients (NCDB: 572 196; SEER: 205 631). Proximal colon cancers were more common in older and female patients in both databases, whereas distal colon cancers were more prevalent in younger patients. NCDB contained a slightly higher proportion of Caucasian patients, while SEER had a greater representation of Asian patients. Segmental resections were more frequent in SEER, with the highest weighted difference observed in sigmoid colon cancer (6.16%; 95% CI: 5.78%-6.54%). OS rates were generally comparable across databases, though minor variations were observed at different colonic locations and stages.

Conclusion: Despite differences in sampling techniques and follow-up reporting, NCDB and SEER demonstrated remarkable consistency in survival trends across colon cancer stages and locations. Recognizing these database-specific variations is essential for researchers conducting population-based survival analyses and selecting the most suitable registry for their studies. WHAT DOES THIS PAPER ADD TO THE LITERATURE?: This study provides a comprehensive comparison of NCDB and SEER, highlighting how differences in sampling methodologies, follow-up reporting, and patient representation influence overall survival estimates in colon cancer. It clarifies why prior studies report conflicting survival trends and offers a methodological framework for researchers selecting the most appropriate database for their analysis.