Sex differences in cancer risk and survival

Abstract

Aim: The objective of study I was to delineate and quantify sex differences in cancer risk and survival together with assessing the potential gain achieved by eliminating the excess cancer risk in men. Study II and III aimed to in detail characterize the superior non-small cell lung cancer survival and the inferior urinary bladder cancer survival, in women, with the underlying objective to identify underlying drivers to these two phenomena. In study IV we wanted to explore to what extent taller body stature can explain the excess cancer risk in men.

Methods: All of the studies are Swedish population-based cohort studies. Study I included all incident cancer cases (n=872,397) recorded in the Swedish Cancer Register in 1970-2014 at age 15-84. The association between sex and cancer risk and sex and cancer survival was assessed by estimating male-to-female incidence rate ratios (IRRs) and excess mortality ratios (EMRs), respectively, using Poisson regression models adjusted for age and calendar year. All incident lung squamous cell carcinoma (n=10,325) and adenocarcinoma (n=23,465) cases recorded in the Swedish Lung Cancer Register in 2002-2016 formed the basis in Study II. Flexible parametric models were applied to compute adjusted female-to-male hazard ratios (HRs) and standardized survival proportions over follow-up, including; age, year, education, marital status, birth country, health care region, ECOG performance status, smoking history, comorbidity, TNM stage, and tumor location, in the final model. A subgroup analysis of lung adenocarcinoma, additionally adjusting for EGFR mutational status, was additionally performed. In study III we included all records of urothelial bladder cancer diagnosed in 1997-2014 at age 18-89 in the Swedish Urinary Bladder Cancer Register (n=36,344). We estimated empirical survival proportions and mortality rates in men and women as well as female-to-male adjusted HRs and standardized survival proportions, using flexible parametric models including; age, year, WHO grade, TNM stage, marital status, education, health care region, birth country, and comorbidity, in the fully-adjusted models. In study IV individual-level information on height from the Swedish Passport Register, the Conscription Register, and the Medical Birth Register (n=6,156,659) was linked to the Swedish Cancer Register where 285,778 cancer cases were identified. Contemporary mediation analysis was applied to assess the effect of male sex, explained by height, on cancer risk.

Results: In study I we found that men are at a higher risk of 34 of 39 malignancies, and have a poorer survival in 27 of 39. Except for smoking-associated malignancies, the excess risk in men is stable over calendar time. In male predominant sites, IRRs range from 1.05; 95% CI, 1.02-1.1 (lung adenocarcinoma) to 8.0; 95% CI, 7.5-85 (laryngeal cancer). Women with non-small cell lung cancer (study II) are younger, smoke less, and present better performance status, compared to men. Women with lung adenocarcinoma additionally present lower comorbidity burden, less advanced stage, and more often harbor activating EGFR mutations. Women with non-small cell lung cancer have a superior survival that is most consistent in lung adenocarcinoma where female-to-male HRs ranged from 0.69; 95% CI 0.63-0.76 (stage IA-IIB) to 0.94; 95% CI 0.88-0.99 (stage IIIB-IV). HR estimates remain largely unchanged after meticulous adjustments. Except for an unfavorable stage distribution in women, we found sparse evidence of sex differences in clinical management or tumor aggressiveness, in urothelial bladder cancer (study III). Women, overall, have a poorer bladder cancer survival (adjusted HR 1.15; 95% CI 1.08-1.23) which is driven by muscle invasive tumors (adjusted HR 1.24; 95% CI 1.14-1.34) and restricted to the first two years from diagnosis. Study IV confirmed that a majority of investigated cancers are associated with male sex (here, 33 of 39) and body height (27 of 39). A fair proportion of the excess male cancer risk is explained by taller body stature, and ranges from 0.5% (laryngeal) to 100% (salivary, colon, melanoma, and AML). The effect of body height and the mediated.

Conclusion: In Study I we found that male sex is a consistent risk as well as a negative prognostic factor for a majority of cancers. Identifying and eliminating underlying factors to the excess cancer risk in men could substantially reduce the global cancer burden. Men with lung adenocarcinoma have a consistently poorer survival that remained largely unchanged after adjustments for a range of prognostic factors, indicating sex differences in tumor biology (study II). The excess bladder cancer mortality in women is limited to muscle-invasive tumors, only noticeable within the first two years from diagnosis, and cannot be explained by the examined clinicopathological factors (study III). This warrants further investigation of sex differences in outcomes and complications to radical cystectomy. A large proportion of the excess cancer risk in men is explainable by height (study IV). This finding corroborate that a considerable proportion of cancer cases are a result of random processes during DNA replication (i.e., bad luck) rather than underlying hereditary and/or environmental factors.