Unbiased transcriptomic analysis of primary early-onset vs late-onset colorectal cancer.

Abstract

228 Background: Colorectal cancer is the 3rd most diagnosed and 2nd leading cause of cancer death. Despite the decline in incidence late-onset colorectal cancer (LOCRC), there is an increase in early onset of colorectal cancer (EORC). Clinicopathologic examinations from single centers revealed left-sided and rectal enrichment, poorly differentiated, and higher stage at presentation. Genomic studies examining the potential germline mutations/variants for EOCRC demonstrated mixed results. Transcriptomic analysis of the EOCRC compared to LOCRC remains limited. We hypothesized that EOCRC present with enrichment in aggressive molecular signatures consisting of CMS4 and iCMS3 signatures. To understand the mechanism behind EOCRC, we performed an unbiased RNA-Seq analysis in age groups of 32-82 in 95 patients. Methods: We utilized RNA-Seq data from the Orien Network, a collaborative between 19 NCI/NCCN cancer centers. We identified 95 patients diagnosed with colorectal cancer ranging in age from 32-82 and undergone RNA-sequencing of their primary tumors. Clinicopathologic assessment were analogously performed. We collected unnormalized raw-count RNA-Seq data from these patients. To account for differences in sequencing depth across samples, raw counts were scaled using the Reads Per Million (RPM) formula. The scaled reads in RPM were then utilized to assess the Consensus Molecular Subtypes (CMS) classification through two distinct approaches: the R package CMScaller and a deconvolution analysis using the R package SpatialDecon with the iCMS gene signature. Differences in age at the sample collection across patients with different CMS classifications were assessed by performing two-tailed unpaired T-tests. Additionally, differential expression analysis between EOCRC and LOCRC patients, defined by the upper and lower quartiles of the patients’ ages at the sample collection, was conducted using the LIMMA R package. Results: Clinicopathologic assessment of the primary tumors revealed left-sided tumors with poorly differentiated phenotype. Differential gene expression analysis between EOCRC and LOCRC demonstrated no significant differences. Stratification of transcriptomic data by decades also did not reveal any significant differences. Furthermore, there were no differences in CMS or iCMS classifications between different age groups. Conclusions: The RNA-Seq analysis comparing EOCRC and LOCRC did not identify any differential gene expression patterns. There was no significant differences were observed in CMS or iCMS classifications. These findings suggest that EOCRC and LOCRC share similar tumor biology based on molecular phenotypes. This implies that the initiating events for the disease may occur earlier in life while triggering a same disease process. As such, the findings underscores the importance of considering preventive measures, including earlier colonoscopy screenings.