KMT2 Mutations Predict Immunotherapy Benefit in Esophageal Adenocarcinoma, Not Squamous Cell Carcinoma
核心洞察
A genomic analysis of 787 esophageal cancer patients found KMT2 (搜索) mutations were markedly more frequent in esophageal squamous cell carcinoma (搜索) (20.8%) than in esophageal adenocarcinoma (搜索) (4.3%).
KMT2 (搜索)-mutant esophageal adenocarcinoma (搜索) patients had significantly longer immune-related overall survival on immunotherapy (21.42 vs 13.42 months; HR 0.66, P=0.004), but no such benefit emerged in ESCC.
KMT2 (搜索) mutations were strongly associated with microsatellite instability-high status and higher tumor mutational burden in both histologies, supporting their role as an immunotherapy biomarker.
A comprehensive genomic analysis of esophageal carcinoma has identified recurrent mutations in the KMT2 (搜索) family of histone methyltransferases as a potential biomarker for immunotherapy response—but only in esophageal adenocarcinoma (搜索) (EA), not in esophageal squamous cell carcinoma (搜索) (ESCC). The study, based on a database of 787 patients established by the CLIA-certified laboratory Caris Life Sciences, found that KMT2-mutant EA patients experienced significantly longer immune-related overall survival (irOS) on immune checkpoint inhibitor (ICI) therapy compared with KMT2-wild-type patients.
The work addresses a pressing clinical need. Although anti-PD-1 inhibitors have reshaped treatment of esophageal cancer—supported by results from ATTRACTION-3, KEYNOTE-181, KEYNOTE-590, and CheckMate 648—some patients still do not benefit from ICI therapy. The use of biomarkers to guide ICI treatment in esophageal cancer is clinically required, a need recently endorsed by the Food and Drug Administration's Oncologic Drug Advisory Committee (ODAC).
KMT2 (搜索) Mutations Are More Common in Squamous Cell Carcinoma
The study examined 604 EA and 183 ESCC patients, analyzing the two histologies separately due to their distinct biological features. KMT2 (搜索) mutations were significantly more frequent in ESCC than in EA (20.8% versus 4.3%). In ESCC, 5 patients (2.7%) carried KMT2A mutations, 7 (3.8%) carried KMT2C mutations, and 26 (14.2%) carried KMT2D mutations. In EA, the corresponding figures were 7 (1.2%), 7 (1.2%), and 12 (2.0%). In both cohorts, KMT2D was the most commonly mutated member of the KMT2 family.
Survival Benefit Restricted to Esophageal Adenocarcinoma (搜索)
The key clinical finding diverged sharply between the two histologies. EA patients with KMT2 (搜索)-mutant tumors had significantly longer irOS than those with KMT2-wild-type tumors (21.42 versus 13.42 months; HR = 0.66, 95% CI: 0.50–0.88, P = 0.004). This association persisted in multivariable Cox analysis (HR 0.74, 95% CI 0.54–1.01, P = 0.06).
By contrast, no significant difference in irOS was observed between ESCC patients with KMT2 (搜索)-mutant and KMT2-wild-type tumors, in either univariate or multivariable analyses (median OS: 11.94 months versus 10.20 months; univariate HR = 0.96, 95% CI 0.67–1.38, P = 0.825; multivariable HR = 1.0, 95% CI 0.69–1.45, P = 0.99).
Association with MSI-H and Tumor Mutational Burden
KMT2 (搜索) mutations were strongly associated with established biomarkers of immunotherapy response. KMT2 mutations correlated with microsatellite instability-high (MSI-H) status in both EA (19.2% vs 1.0%) and ESCC (5.4% vs 0%), and with higher tumor mutational burden (TMB; median 9 vs 8 mutations per megabase) in both cohorts (EA P < 0.0001; ESCC P = 0.042). Even after excluding MSI-H samples, the association of KMT2 mutations with higher TMB remained significant in both the EA (P = 0.008) and ESCC (P = 0.015) cohorts. There was no significant association between KMT2 mutations and PD-L1 expression in either histology.
Distinct Co-Mutation and Transcriptomic Profiles
KMT2 (搜索)-mutant EA was associated with higher rates of co-occurring mutations in genes including CIC (13.0% vs 1.1%), NF1 (12.5% vs 2.2%), ATM (11.5% vs 2.6%), BRCA1 (11.5% vs 1.0%), FLCN (11.5% vs 0%), RNF43 (11.5% vs 1.4%), ATRX (10.0% vs 0.5%), and others (all q < 0.05). In a microsatellite-stable subgroup analysis, higher mutation rates of NF1 (10% vs 2%), PIK3R1 (5% vs 0.5%), and BMPR1A (4.8% vs 0.4%) remained significant in KMT2-mutant EA.
Transcriptomic analysis using the TCGA database identified 464 differentially expressed genes between KMT2 (搜索)-mutant and wild-type EA patients, enriched in fat digestion and absorption and cholesterol metabolism pathways. Activated B-cell infiltration was significantly higher in KMT2-mutant EA. In contrast, only 50 differentially expressed genes were found in ESCC, with no enriched signaling pathways and no significant difference in immune-cell infiltration.
Mechanistic Rationale and Clinical Implications
Members of the KMT2 (搜索) family methylate histone H3 on lysine 4 (H3K4), promoting genome accessibility and transcription, which may increase TMB and neoantigen production. The authors note that many genes co-occurring with KMT2 mutations in EA—including ATM, BRCA1, SMARCA4, ATRX, BRCA2, RAD50, and WRN—relate to DNA damage repair pathways. This suggests KMT2C/D mutations may sensitize cancer cells to Poly(ADP-ribose) polymerase inhibitors (PARPi).
A transcriptome-wide association study (TWAS) found the KMT2 (搜索) gene family was not associated with esophageal cancer risk in either UTMOST or FUSION analyses, suggesting KMT2 mutations are likely concomitant rather than driver events in esophageal cancer pathogenesis.
Study Limitations
The authors acknowledge several limitations. The retrospective design may introduce selection bias, and prospective randomized studies with larger sample sizes integrating genome-wide genotype, transcriptomic, and immunotherapy outcome data are warranted to confirm the predictive value of KMT2 (搜索) mutations. Clinical information regarding line of therapy and combination regimens was not available. Finally, the functional roles of KMT2 mutations in esophageal cancer require further exploration in vitro and in vivo to clarify their impact on metabolic processes and the tumor immune environment.
