LAUSR

Disparities undeniably exist in cancer, understandably arouse great concerns, and need to be addressed. Cancer disparities may arise from the interplay of risk factor differences, treatment adherence, and sexbiased response to treatment, in addition to underlying biologic mechanisms. Pinpointing the disadvantaged group, understanding the contributions of risk factors, and identifying actionable areas for improvement across the cancer continuum will abrogate cancer disparities. To date, disparity research has largely focused on racial and geographic disparities. In the past decade, sex disparities have been increasingly recognized in cancer (and other diseases) and have been actively studied. Leveraging a large, prospective cohort, the National Institutes of Health (NIH)AARP Diet and Health Study (https:// dieta ndhea lth.cancer.gov/, Accessed May 20, 2022), with 171,274 male and 122,826 female participants, Jackson et al. calculated ageadjusted cancer incidence and quantified contributions from many risk factors on patient survival in 21 nonreproductive cancers. The key observations on male predominance in cancer incidence and inferior male survival in the majority of the 21 cancer types are not new but were previously reported based on the Surveillance, Epidemiology, and End Results (SEER) program database, a cancer registry representative of the US population. However, the SEER database does not collect individuallevel risk factors, and thus previous reports have not evaluated the contributions of risk factors to the observed sex disparities. The authors comprehensively examined the contributions of a wide spectrum of nongenetic risk factors collected in the NIHAARP study to quantify sex disparities in cancer, including behavior anthropometrics (e.g., body mass index and height), lifestyle factors (physical activity, diet, smoking, alcohol drinking), medication use, and family cancer history. The authors should be applauded for the thorough and comprehensive analyses performed in the study. They quantified and tested whether and to what extent the distributions of the risk factors explained the observed sex disparities using the Peters– Belson method for timetoevent survival outcomes. The interactions of sex with the other risk factors under analyses were taken into account. The authors also examined the ageing effect by accommodating the interaction of sex with age (in both dichotomized and continuous scale). In consideration of the confounding impact of unmeasured infectious agents (e.g., human papillomavirus, hepatitis C virus) in infectionrelated cancers, the effect of sex on cancer risk was adjusted by a bias factor calculated according to the Steenland & Greenland method in sensitivity analyses. Subset analyses were conducted within strata of key risk factors for effect modification. Nonlinear effects of age and height were included into the models while model assumptions were examined. Sex disparity is cancer typespecific. For instance, immunotherapy tends to favor females for nonsmall cell lung cancer but favors males for colorectal cancer. Therefore, cancer typespecific analysis, as done in this study, is preferred rather than the naive pancancer analyses in which data from all of the cancers are simply lumped together to assume that the sex effect is unidirectional across all cancers. In summary, this study has furthered our understanding on sex disparities in cancer, particularly in terms of the contributions of risk factors. In this study, Jackson and colleagues report that all the nongenetic risk factors under examination only explained less than one half of the sitespecific male predominance. Contribution to the male predominance is multifaceted. Going beyond sex chromosomes and sex hormones, many essential human biologic mechanisms and epigenetics are sexdistinct, constituting the biologic fundamentals for sex disparities in cancer. In addition to the genetic and epigenetic factors, the lifestyle (examined in part in this study) and environmental factors also play a role in sex disparities, although to a various extent by cancer site. Yet these risk factors do not stand alone but, rather, interplay intricately while exerting their effects on sex disparities. Examining the contributions of risk factors to patient survival, smoking and alcohol drinking were retained after variable selection for nearly all cancer types in this study, underscoring the importance of these two (and potentially