Despite its significance, the mechanisms behind therapy resistance in metastatic colorectal cancer (mCRC) remain poorly understood. Circulating tumor DNA (ctDNA) offers a unique opportunity to explore them, as it thoroughly captures intratumor heterogeneity and facilitates longitudinal monitoring. However, most studies have focused on known mutations, mainly related to anti-EGFR resistance. The METACC study aims to go beyond mutation-centric approaches to uncover novel mCRC therapy resistance mechanisms.

We analyzed a retrospective cohort of 30 microsatellite-stable mCRC patients, with plasma samples collected at diagnosis and at progression to first-line therapy (chemotherapy ± anti-VEGF or anti-EGFR). Whole-exome sequencing (WES) was performed on both timepoints and on matched WBCs (to exclude non-tumoral variants). Unique molecular identifiers and high-sensitivity variant calling enabled robust detection of low-frequency variants. McNemar’s test identified mutations enriched at progression, and multivariable Cox regression assessed their impact on survival. Methylation changes were analyzed using Infinium MethylationEPIC v2.0 BeadChips (~930k CpGs) with a plasma-adapted workflow. SeSAMe processing and subsequent DMRCate analysis identified differentially methylated regions (DMRs), prioritizing those overlapping with regulatory regions.

Plasma WES identified ARID1A mutations as the main acquired event (progression vs. diagnosis: 14/30 vs. 5/30; p = 0.022), shortening time to progression (HR = 3.2, p = 0.036). Variant allele frequency (VAF) gains in ARID1A correlated with increases in the tumor mutational burden (TMB) (ρ = 0.37; p = 0.037), in line with the gene's role in maintaining genomic integrity. Progression also concurred with VAF increases in genes associated with therapy resistance via 1) epigenomic dysregulation (e.g.: NF1, KMT2D), and 2) DNA mismatch repair (MSH2, MLH1). The latter strongly correlated with TMB (ρ = 0.72; p < 0.001), further implicating genomic instability. Differential methylation analysis identified 16,026 DMRs (198 with |Δβ| > 0.2), including the hypermethylation of an ARID1A enhancer (Δβ = 0.29), suggesting a dual mechanism of ARID1A inactivation. Enrichment analysis linked these epigenetic alterations to three major pathway groups: 1) tumor metabolism and microenvironment, 2) anti-EGFR resistance, and 3) longevity/mTOR signaling pathways. Notably, hypermethylation of the mTOR inhibitor DEPTOR (Δβ = 0.35) and hypomethylation of SOX2-OT (Δβ = -0.33) suggest a role in enhancing tumor plasticity and stemness.

Comprehensive ctDNA profiling reveals tumor plasticity as a key mechanism of resistance to therapy in mCRC, driven by both genomic instability and epigenomic dysregulation. Expanding the METACC cohort will help explore actionable biomarkers like ARID1A loss, which may inform immunotherapy strategies.