Current Cancer Drug Targets · Oct 29, 2026
Alcohol Consumption Patterns as a Predictor of Pan-Cancer Risk and Clinical Prognosis Across Multi-Modal Evidence
By Wang Y, Cai S, Sun Y et al.
Background: Alcohol is a modifiable carcinogen with heterogeneous risk profiles across consumption patterns, yet robust causal evidence and translational strategies to mitigate alcohol-related cancer burdens remain limited. Methods: We performed bidirectional MR and meta-analyses to assess causal links between drinking status/beverage types and 24 cancers. For esophageal carcinoma (ESCA), two-step MR evaluated potential mediators. Alcohol history-informed prognostic models were developed by machine learning and internally validated in ESCA and head and neck squamous cell carcinoma (HNSC). Results: Per-cancer FDR-significant associations (q<0.05) included previous drinking with ESCA (OR=5.54), cholangiocarcinoma (OR=2.10), and lung squamous cell carcinoma (OR=1.49), and alcohol dependence with ESCA (OR=1.49); current drinking showed an inverse association with ovarian cancer (OR=0.75). Red wine (OR=9.13) and beer/cider (OR=7.97) were associated with higher ESCA risk, and white wine with lung squamous cell carcinoma risk (OR=1.69); these beverage-specific estimates are not comparable across scales. H. pylori was not established as a mediator, and none of the 731 immune-cell traits remained significant after FDR correction. A five-gene, alcohol history-derived signature was internally validated for ESCA survival (AUCs 0.62-0.74), and screening of >100 algorithms in HNSC supported framework feasibility. Discussion: Our multimodal analysis reveals causal heterogeneity across cancers with beverage-specific risk profiles. The apparent inverse ovarian-cancer association may reflect survival bias rather than true protection. The alcohol history-derived signature provides exploratory molecular targets; independent external validation is required before clinical use. conclusion: This study establishes alcohol consumption patterns as a pan-cancer predictor through multi-modal evidence, bridging causal etiology to clinical translation for prevention and prognostic stratification. Conclusions: These findings support alcohol consumption patterns as pan-cancer risk factors and propose an exploratory framework for repurposing drinking history toward prognostic risk stratification, subject to independent external validation.