Multidimensional Evidence Chain of Nutritional Impact on Digestive System Tumors: Integration of Mendelian Randomization Causal Inference, NHANES Epidemiological Data, and Transcriptomic Analysis.

Background: The global burden of digestive system tumors (e.g., hepatocellular carcinoma, gastric cancer, colorectal cancer) continues to rise, with low survival rates in advanced stages, necessitating etiology-based intervention strategies. Nutritional metabolic disorders are closely linked to tumor progression, but existing studies predominantly focus on correlational analyses, leaving causal relationships and molecular mechanisms poorly understood.

Objective: Integrating Mendelian randomization (MR), National Health and Nutrition Examination Survey (NHANES) epidemiological data, and transcriptomic data analysis, this study aims to elucidate the impact of nutritional status on digestive system tumors, identify key genes, develop prognostic models for precise nutritional interventions, and propose potential therapeutic directions.

Methods: Causal Inference: MR analysis using OpenGWAS data to assess causal associations between nutritional symptoms and digestive system tumor. Epidemiological Validation: Restricted cubic spline models analyzed nonlinear associations between nutritional status (Geriatric Nutritional Risk Index [GNRI]) and telomere length (DNAmTL) in NHANES 1999-2002 data (n = 2,532). Gender stratification and mediation effects tested nutritional pathways. Molecular Mechanism: Transcriptomic data from TCGA (CHOL, LIHC, COAD, etc.) identified malnutrition-related genes. Survival analysis, immune subtype classification (C1-C6), and tumor microenvironment scoring (ESTIMATE/RNAss) were integrated to build a Cox prognostic model. Machine learning (ML; Random Forest/Support Vector Machine) screened key genes. CellMiner database linked gene expression to drug sensitivity.

Results: Causal Association: MR confirmed significant causal effects of nutrition on digestive system tumor risk (P_IVW < 0.05). Gender Heterogeneity: Males exhibited accelerated telomere shortening at GNRI > 115 (β = -0.0123/unit, P = 0.035), while females showed no significant association. Dietary quality (Healthy Eating Index) directly protected telomeres (β = 0.069, P = 0.003), linked to seafood/plant protein (β = 0.066) and whole fruit intake (β = 0.067). Key Genes: ML identified ACTG2, MSX1, and COL7A1 as core drivers. Prognostic Model: A risk score model (ATP6V0A1*0.70 + TP63*0.37 + SLC7A7*0.49 + ARHGAP29*0.33 + CDH1*(-0.50)) distinguished high/low-risk groups (AUC = 0.977). Potentially Available Drugs: Zoledronate, LY-294002, and Everolimus may be potential choices for malnutrition digestive tract tumors.

Conclusion: This study systematically reveals multidimensional mechanisms linking malnutrition to digestive system tumor progression via genetic, molecular, and immune pathways. Key genes (ACTG2, MSX1, COL7A1) and gender-specific interventions offer novel strategies for precision oncology. Future validation in cross-ethnic cohorts and clinical trials is warranted.