Background and AimsThe contribution of suboptimal diets to gastrointestinal (GI) cancer incidence globally remains unquantified, and we aimed to evaluate it. MethodsComprehensive meta-analyses and rigorous evidence grading assessment identified the associations between suboptimal diets and six GI cancers and their subtypes. A comparative risk assessment model was employed to estimate the proportional attributable burden and attributable rate of GI cancers to suboptimal diets by using the corroborative association estimates. Additionally, correlation assessments with the Sociodemographic Index (SDI) were carried out. ResultsIn 2018, 21.5% (95% Uncertainty Interval (UI): 19.1%, 24.5%) of incident GI cancer cases globally were attributable to suboptimal diets, maintaining a relatively stable proportion since 1990 (22.4% (19.7%, 25.6%)), while the absolute diet-attributable cases doubled from 581 thousands (511 thousands, 664 thousands) in 1990 to 1,040 thousands (923 thousands, 1,187 thousands) in 2018. Excessive processed meat consumption (5.9% (4.2%, 7.9%)), insufficient fruit intake (4.8% (3.8%, 5.9%)), and insufficient whole grain intake (3.6% (2.8%, 5.1%)) were the most significant dietary risk factors in 2018, a shift from 1990 when the third major concern was insufficient non-starchy vegetable intake. Additionally, Central and Eastern Europe, and Central Asia experienced the highest attributable burden across regions in both 1990 (31.6% (27.0%, 37.4%)) and 2018 (31.6% (27.3%, 36.5%)), and a positive correlation (P < .01) between the SDI and the attributable GI cancer incidence was observed. ConclusionsAlthough the proportional attributable GI incidence remains relatively stable, the doubling of absolute cases from 1990 to 2018, along with the discrepancies among urbanicity and countries/regions, informs dietary priorities and more targeted preventive measures.