Hada et al. Page 189

Previous studies suggested inflammation increases risk for ovarian cancer while aspirin use decreases risk, though the mechanisms are unknown. Hada and colleagues evaluated prediagnostic circulating levels of 31 lipid molecules, including arachidonic, linoleic, and alpha-linoleic acid metabolites. Metabolites in the linoleic acid pathway, precursors of inflammatory markers, were associated with increased risk of ovarian cancer, supporting a pro-inflammatory mechanism. Further, prostaglandins/thromboxane derived from arachidonic acid via COX pathway were not associated with risk. The identification of novel, risk-related fatty acid metabolites provides mechanistic insights into the etiology of ovarian cancer and supports future research of these pathways.

Bond-Smith and Stone Page 22

Since the influential contribution by McCormack and dos Santos Silva (CEBP, 2006), no comprehensive meta-analysis of the association between mammographic density and breast cancer has been published. At the same time, the literature has expanded substantially. This article by Bond-Smith and Stone fills this gap and provides guidance on the impact of using different mammographic density measurement and analysis strategies as a risk marker for breast cancer. The authors specifically discuss methodological challenges for meta-analyses with small and noisy data where appropriate estimation of heterogeneity is critical (and frequently neglected) for obtaining reliable and robust clinical findings.

Warnecke et al. Page 59

Conceptualization, data collection, and analysis at multiple levels can help identify determinants of health disparities that are often systemic. In a study of non-Hispanic white, non-Hispanic black, and Hispanic newly diagnosed female breast cancer patients (N=989), Warnecke and colleagues examined patient resources, neighborhood primary care access, and screening and diagnostic facility characteristics as potential determinants of disparities in stage at diagnosis. The authors found that racial/ethnic differences in mode of breast cancer detection (through screening or symptoms), and differences in the characteristics of facilities where patients are diagnosed, accounted for much of the disparity in later stage at diagnosis.

Usher-Smith et al. Page 67

Up to 40% of cancer cases are thought to be attributable to lifestyle factors. Usher-Smith and colleagues developed lifestyle-based models for the five cancers for which the most cases are potentially preventable through lifestyle change in the UK using estimates of relative risk from meta-analyses of observational studies. They assessed the performance of the models in a UK-based cohort and found that the discrimination and calibration of the models are both reasonable. These models could be used to demonstrate the potential impact of lifestyle change on risk of cancer to promote behaviour change.