Abstract 7413: Establishing a database for the discovery and validation of survival-associated genes in colon cancer

Abstract

Abstract Introduction: Unveiling the prognostic significance of genes as potential biomarkers in colon cancer (CRC) is crucial for the discovery of new therapeutic targets and the effective stratification of patients into clinically relevant groups. The primary objective of this study was to establish a transcriptomic cohort comprising clinical data from colon cancer specimens and to leverage this dataset to pinpoint genes with robust prognostic implications. Methods: Our approach involved a comprehensive search of the Gene Expression Omnibus (GEO) to identify datasets containing both clinical information and raw gene expression measurements for individual colon cancer specimens. Subsequently, we curated and integrated these datasets into a unified database, and linked into our Kaplan-Meier plotter. This integration enabled the identification of genes whose expression changes were associated with altered survival outcomes. The identification of the most influential genes in relation to overall survival in stage 2 and stage 3/4 tumors was accomplished through both uni- and multivariate Cox regression analyses. To mitigate the impact of multiple hypothesis testing, a false discovery rate was computed for each marker. Results: The database encompasses a total of 2,137 tumor samples derived from 17 distinct independent cohorts. Within this dataset, 1,343 samples are equipped with recorded relapse-free survival data, while 1,061 samples provide information on overall survival events along with the corresponding duration. In addition to survival-related metrics, the database encompasses a spectrum of supplementary clinical parameters such as gender, T (tumor size), N (lymph node involvement), M (metastasis), grade, stage, location, and microsatellite instability. The most significant genes associated with relapse-fee survival with FDR<1% in stage 2 CRC include CSF1R (HR=2.86), FLNA (HR=2.88), TPBG (HR=2.65), B2M (HR=2.62) and LOXL2 (HR=2.63). When ordered by the Cox regression p-value in stage 3 and 4 tumors, the five strongest genes were BHLHE40 (HR=2.6), COL4A2 (HR=2.02), TSC22D3 (HR=1.95), NPR3 (HR=1.91), and A2M (HR=1.86). Conclusions: In this study, we successfully developed a robust integrated database dedicated to colon cancer. By leveraging this resource, we pinpointed the most impactful prognostic biomarkers and potential therapy targets specifically within stage 2 and stage 3/4 patients. Beyond its current applications, this database holds promise for future endeavors, serving as a valuable tool for identifying and prioritizing additional promising biomarkers and therapeutic targets in other sub-cohorts of colon cancer. The entire database can be utilized at www.kmplot.com. Citation Format: Balazs Gyorffy. Establishing a database for the discovery and validation of survival-associated genes in colon cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 7413.