Answers are grounded in this paper and cite their sources.
Background: Colorectal cancer (CRC) exhibits substantial inter-patient heterogeneity in clinical outcomes. Despite this, prognostic stratification tools are limited although multi-omics studies have tried to identify biomarkers for patient discrimination. Alternative splicing represents a major source of transcriptomic diversity in CRC, but its use in biomarker development remains limited. The present study aimed to characterize the prognostic alternative splicing landscape of CRC through an integrated multi-omics framework. Methods: Percent spliced-in (PSI) values for colon and rectal cancers were obtained from TCGASpliceSeq database. Of 1000 high-variance splicing events initially selected, 971 events that met quality filters were evaluated using univariate Cox proportional hazards regression to identify candidate prognostic events followed by construction of a LASSO-Cox multi-feature prognostic signature, which was cross-validated. The prognostic relevance of selected splicing-associated genes was evaluated in three independent cohorts (GSE17538, GSE72970, GSE39582) totaling 875 patients. The biological context of candidate splicing events was characterized through competing risk analysis, mutation-splicing regulatory mapping, and tumor immune microenvironment profiling. Results: Twenty candidate prognostic splicing events were identified based on unadjusted p-values, as no event survived FDR correction; these represent exploratory candidates with a substantial risk of false-positive associations. The exon-skipping event showed the highest mean hazard ratio among the five event types. The LASSO-Cox signature showed moderate time-dependent discrimination in the TCGA discovery cohort (cross-validated C-index 0.586). Gene-level expression of the 20 splicing-associated genes showed consistent prognostic stratification across all three independent cohorts, providing supportive but indirect evidence of prognostic relevance. A total of 322 mutation-splicing associations remained significant after multiple-testing correction (FDR < 0.05). NTMT1 and MAST1 showed a significant correlation with immune infiltration. Conclusions: The study presents a multi-cohort exploratory analysis of alternative splicing as a potential prognostic marker for CRC. Twenty candidate events identified based on unadjusted p-values should be considered as preliminary candidates for further investigation.
No takes yet. Share an insight, caveat, or question.
Lal et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: