AI/ML-Integrated Nutritional and Metabolic Multi-Omics in Cancer: A Systematic Review, Meta-Analysis, and Proposed Translational Framework

Cancer is increasingly recognized as a metabolic disease influenced by nutritional factors, with multi-omics technologies and artificial intelligence (AI), particularly machine learning (ML), enabling integrative analyses of diet, metabolism, and tumor biology interactions. This study aimed to synthesize evidence on these approaches for understanding the nutrition–metabolism–cancer axis and assess their translational potential in oncology, especially in low-resource settings. A PRISMA-compliant systematic review and meta-analysis searched PubMed, EMBASE, and Cochrane databases from 2018 to 2025, including studies on human cancers using ≥2 omics layers integrated via AI/ML and addressing nutritional/metabolic exposures. Random-effects pooling evaluated area under the curve (AUC), odds ratios (OR), and clinical endpoints, with subgroup analyses and quality assessments via QUADAS-2, ROBINS-I, TRIPOD, and PRISMA-AI. From 4812 records, 42 studies were included, yielding a pooled AUC of 0.88 (95% CI: 0.86–0.91) and OR of 2.4 (95% CI: 1.2–3.5), demonstrating encouraging but early-stage exploratory evidence of predictive performance. Cancer-specific signatures emerged in colorectal, breast, pancreatic, liver, and hematologic malignancies. A conceptual translational framework was proposed, integrating nutrition, omics, AI/ML, and oncology to illustrate a potential implementation pathway for developing countries like Saudi Arabia. These findings represent preliminary, hypothesis-generating evidence; the proposed framework requires prospective validation before clinical deployment, particularly in resource-limited settings.