Artificial Intelligence-Guided Analysis of WNT Pathway Alterations: Associations with Genomic Burden and Survival Across African American and Non-Hispanic White Populations, Age Groups, and FOLFOX Treatment in Colorectal Cancer.

Colorectal cancer (CRC) exhibits substantial heterogeneity across ancestry, age at onset, and treatment exposure. Although dysregulation of the WNT signaling pathway is a hallmark of CRC, its associations with genomic burden and survival across diverse clinical contexts remain incompletely understood. We analyzed 2562 CRC cases from AACR Project GENIE and cBioPortal, stratified by ancestry (African American [AA] and non-Hispanic White [NHW]), age at onset (early vs. late), and FOLFOX treatment status. Associations between WNT pathway alterations, genomic burden (mutation count, tumor mutational burden, and fraction of genome altered), and survival were assessed using conventional statistical methods. AI-HOPE and AI-HOPE-WNT conversational artificial intelligence platforms were used to facilitate data integration and exploratory analyses. WNT pathway alterations were highly prevalent across all subgroups and were predominantly driven by APC alterations. Late-onset CRC, particularly among NHW patients not treated with FOLFOX, exhibited higher mutation burden and enrichment of AXIN1 and AXIN2 alterations. Survival analyses demonstrated context-dependent associations between WNT alterations and outcomes. Among early-onset FOLFOX-treated patients, WNT alterations were associated with differential survival. In NHW patients, WNT alterations were linked to improved survival across multiple clinical settings, whereas associations in AA patients were more limited and context-specific. WNT pathway alterations are pervasive in CRC but exhibit ancestry-, age-, and treatment-dependent associations with genomic complexity and survival. AI-guided analyses may accelerate identification of clinically relevant subgroup-specific molecular patterns.