zegeye Abitew

MSc

Flinders University, Flinders Health and Medical Research Institute, College of Medicine and Public Health, Adelaide, South Australia

PhD Student

PLCO (Learn more about this study)

PLCO-1166

Feb 27, 2023

Reduced rank regression derived dietary patterns and risk of colorectal cancer and mortalit

Colorectal cancer (CRC) is the fourth most common cancer worldwide and the third most common in the United States (1). According to the global burden of disease report, the incidence of CRC is steadily increasing, with the number of cases and deaths related to it reaching 2,17 million by 2022 (2). It is estimated that 3.2 million new cases of CRC will be diagnosed in 2040, and 1.6 million people will die from CRC-related causes (3).
CRC development from precancerous lesions (conventional adenoma or sessile serrated lesions) is significantly influenced by the combination of the host genetics and environmental risk factors (physical, biological, chemical, or behavioural patterns) (4). Diet is one modifiable behavioural risk factor that has been linked with CRC through inflammation, immune function, oxidative stress, insulin resistance, and estrogen metabolism (5).
The relationship between diet and CRC risk and mortality has been studied in several studies. The majority of research evaluated the impact of diet on CRC risk using a priori methods such as dietary scores or indices and limited studies used posterior methods such as principal component analysis (PCA). Dietary scores/indices, however, cannot account for the interaction between various dietary components, indicate the overall dietary pattern, or provide particular information on multiple foods (6). PCA considers the interaction of different dietary components but does not consider the previous nutritional knowledge of how dietary intake links to CRC risk (7, 8). Therefore, improving the prediction of diet-related risk factors and CRC risks, requires further studies, including biomarkers to identify objective dietary patterns and novel statistical analysis methods which account for the interaction of nutrients and integrate previous nutritional knowledge with empirical data to provide further insights into the link between diet and CRC risk (5).
Currently, hybrid dietary pattern analysis methods which use existing subject matter knowledge to select intermediate response variables and dietary data to drive the dietary patterns have been recommended (9). Reduced rank regression is a commonly used method among hybrid methods of dietary pattern analysis. It is used to identify a linear combination of food groups that explain as much variation as possible in a set of intermediate response variables (10). Very limited studies have dietary patterns using nutrients or biomarkers as a response variable and examine these patterns with CRC risk and mortality. Therefore, we will use reduced rank regression analysis to identify and examine dietary patterns predictive of CRC risk in this analysis.
As the prostate, lung, CRC, and ovarian (PLCO) study has a larger sample size and a longer follow-up period, it is particularly important for this study to determine diet related CRC risk and mortality. Further, the dietary data is comprehensively assessed so we can compare the association between different dietary patterns and CRC risks.