BACKGROUND: The benefits of the Mediterranean-style dietary pattern in mitigating cancer risk among Americans is unclear and its role in obesity-related cancer risk has not been evaluated. OBJECTIVES: This study examines the prospective association between adherence to a Mediterranean style dietary pattern and cancer risk (including total, obesity related, breast and colorectal cancers) among men and women in the Framingham Offspring (FOS) cohort. In secondary analyses for breast cancer, we explore stratifying by hormone receptor status and menopausal status. METHODS: The Mediterranean style dietary pattern (MSDP) score was derived from a semi-quantitative food frequency questionnaire taken at examination visit 5 in the prospective FOS cohort. Subjects included 3199 participants (1703 women and 1496 men), aged 30 years old and older, who were free of prevalent cancer. The MSDP score was classified into tertiles and also dichotomized (MSDP score <19 vs. ≥19) to evaluate the association between the MSDP and cancer risk through the ninth examination cycle (2014). Cox proportional-hazards regression models were used to estimate hazard ratios (HRs) and 95% confidence intervals (CIs) for all subjects and for men and women separately, adjusting for confounding by age, physical activity, body mass Index, pack-years of cigarette smoking, supplement use, diabetes status and sex (for all subjects models). In the breast cancer model, we adjusted for age, waist-to-height ratio height ratio, pack-years, physical activity, diabetes, supplement use age at menopause. Factors found not to confound the effects of the MSDP on cancer risk were excluded from final models. RESULTS: During a median follow-up of approximately 18 years, 377 and 273 cancer cases were documented among men and women, respectively. Women in the upper two tertiles of the MSDP score had approximately 30% lower lower total cancer risks than women in the lowest tertile (tertile 2: HR, 0.69, 95% CI: 0.50-0.94; tertile 3: HR, 0.73; 95% CI: 0.54-0.99). Effects in men were weaker. Higher adherence to a MSDP was somewhat more strongly protective against total cancer risk among lower-risk individuals such as those who were leaner (BMI <25), drank less alcohol (<14 g/d), and did not currently smoke cigarettes. The association between MSDP adherence and total cancer risk was also modified by waist circumference and WHtR. We also observed a non-statistically significant protective effect of higher MSDP conformity and obesity-related cancer risk (tertile 3: HR, 0.80, 95% CI: 0.60-1.07). The association was present especially among women (tertile 2: HR, 0.76, 95% CI 0.53-1.09; tertile 3: HR, 0.73, 95% CI: 0.51-1.05). In analyses of effect modification by anthropometric measures of body fat, the combined effect estimates for higher adherence to the MSDP in women and body fat were more than additive for BMI and WHtR. The MSDP adherence was also inversely associated with BrCa risk (tertile 3 vs tertile 1: HR: 0.58, 95% CI: 0.34-0.98) especially in post-menopausal women (HR: 0.51, 95% CI: 0.29-0.91) and among those with any positive Estrogen Receptor/Progesterone Receptor BrCa (HR: 0.58, 95% CI: 0.31-1.06). We found no association between MSDP and colorectal cancer in these analyses. CONCLUSIONS: In this large cohort study, higher adherence to MSDP was associated with lower cancer risk (including total, obesity-related and breast cancers), among women aged 30 years old or older in the FOS study.

No description available.

This paper investigates the impact the LASSO, mRMR, SHAP, and Reinforcement Feature Selection techniques on random forest models for the breast cancer subtypes markers ER, HER2, PR, and TN as well as identifying a small subset of biomarkers that could potentially cause the disease and explain them using explainable AI techniques. This is important because in areas such as healthcare understanding why the model makes a specific decision is important it is a diagnostic of an individual which requires reliable AI. Another contribution is using feature selection methods to identify a small subset of biomarkers capable of predicting if a specific RNA sequence will have one of the cancer labels positive. The study begins by obtaining baseline accuracy metric using a random forest model on The Cancer Genome Atlas's breast cancer database to then explore the effects of feature selection, selecting different numbers of features, significantly influencing model accuracy, and selecting a small number of potential biomarkers that may produce a specific type of breast cancer. Once the biomarkers were selected, the explainable AI techniques SHAP and LIME were applied to the models and provided insight into influential biomarkers and their impact on predictions. The main results are that there are some shared biomarkers between some of the subsets that had high influence over the model prediction, LASSO and Reinforcement Feature selection sets scoring the highest accuracy of all sets and obtaining some insight into how the models used the features by using existing explainable AI methods SHAP and LIME to understand how these selected features are affecting the model's prediction.