Abstract: Objective To analyze the epidemiological characteristics of upper gastrointestinal cancer in the population under 60 years old in Bozhou area, and construct an early-onset risk prediction model. Methods Data from 13 288 high-risk individuals for upper gastrointestinal cancer, aged less than 60 years, at Mengcheng County Second People's Hospital between 2022 to 2024 were collected. A case-control design was employed, matching by age, sex, and residence at a 1: 2 ratio, which resulted in a case group (n=598) and a control group (n=1196). Data on demographic, clinical, and lifestyle factors were collected. Conditional logistic regression analysis was utilized to identify risk factors, which subsequently informed the construction of a risk prediction model. Results Significant differences were observed between the overall patient population with upper gastrointestinal cancer and the healthy control group with respect to history of digestive system diseases, alcohol consumption, frequency of anger, consumption of spicy foods, preference for scalding-hot food, intake of pickled products, consumption of leftover food, and garlic intake (all χ²<0.01). Within the patient population, the early-onset group (<50 years) showed significant differences compared to the late-onset group in terms of history of digestive system diseases, alcohol consumption, frequency of anger, preference for scalding-hot food, intake of pickled products, and garlic intake (all P<0.01). Multivariate conditional logistic regression analysis revealed that a history of digestive system diseases (OR=3.043), long-term alcohol consumption (OR=3.701), frequent anger (OR=3.533), consumption of scalding-hot food (OR=3.385), frequent intake of pickled products (OR=3.485), and frequent garlic consumption (OR=0.255) were significant factors associated with early-onset upper gastrointestinal cancer in the 40 to <50 age group in the Bozhou region (all P<0.05). A nomogram-based risk prediction model constructed from the logistic regression demonstrated good discrimination, with a Concordance Index (C-index) of 0.835. Conclusions In Bozhou area, individuals under 60 with upper gastrointestinal cancer exhibit significant differences from the healthy population regarding history of digestive system diseases, alcohol consumption, frequent anger, preference for spicy diets, consumption of scalding-hot food, intake of pickled products, consumption of leftover food, and insufficient garlic intake. A significantly higher proportion of early-onset patients, compared to late-onset patients, reported a history of digestive system diseases, long-term alcohol consumption, frequent anger, consumption of scalding-hot food, frequent intake of pickled products, and insufficient garlic intake. These factors were identified as determinants for the early onset of upper gastrointestinal cancer in the high-risk population under 60 in this region. The nomogram risk prediction model, developed based on these factors, demonstrates excellent discrimination and accuracy.

Key words: Middle-aged, Upper gastrointestinal cancer, Epidemiology, Risk prediction model, Nomogram