Abstract
Colorectal cancer is a global health challenge with high morbidity and mortality, but its causative factors remain unclear. In recent years, associations between various dietary patterns and colorectal cancer have been identified, but no studies have examined the association between macro- and micronutrient intake and colorectal cancer. This study analyzed the association between colorectal cancer and dietary intake using the logistic least absolute shrinkage and selection operator (LASSO). The data were derived from national data from the 1999–2010 National Health and Nutrition Examination Survey (NHANES) cycle. These data were further filtered to select those aged 50 years or older who self-reported having colorectal cancer (n = 168) and those who did not self-report having colorectal cancer (n = 649). LASSO regression is a new statistical shrinkage technique based on the R statistical software. In this study, LASSO was used to analyze the association between colorectal cancer and the variables from which the most relevant variables were selected. These variables included currently recognized risk factors for colorectal cancer and nutrients related to dietary intake. Age, sex, and race, which are recognised risk factors, still showed a significant association with colorectal cancer after LASSO regression shrinkage. For dietary intake of macro- and micronutrients, only thiamine (beta = 0.003) and zinc (beta = 0.0007) were positively associated with colorectal cancer. The results suggest that thiamine and zinc may be strongly associated with colorectal cancer. However, the results of the LASSO regression are based on statistically derived propensities and have not been validated by ex vivo experiments.