Sex Differences in Cancer Incidence and Survival: A Pan-Cancer Analysis

Sex plays an important role in the incidence, prognosis, and mortality of numerous cancers (1), but is often not taken into consideration in the treatment of the disease. While the reasons behind sex differences in cancer incidence and survival are not well understood, recent studies have suggested that genetic and environmental factors, and their complex interactions, may contribute to sex differences, specifically to the worse overall prognosis for males compared with females (1–3). Prior research suggests that males have a higher incidence of cancer (1, 4), and recent studies in Europe (5), Korea (6), Canada (7), and Australia (8) have consistently reported that males are less likely to survive a diagnosis of cancer. Identifying disadvantaged groups is critical as it can provide useful information to improve cancer survival, as well as to better understand the etiology and pathogenesis of specific cancers.

Few large-scale studies have systematically examined sex differences in incidence and survival across all cancers. Many of these studies are older and use smaller, less comprehensive datasets (4, 9). A recently published study used the Centers for Disease Control and Prevention's (CDC) National Program of Cancer Registries (NPCR) and the National Cancer Institute's (NCI) Surveillance, Epidemiology, and End Results (SEER) Program datasets to examine incidence rates by sex of cancers diagnosed from 2011 to 2015 and death rates by sex of cancers diagnosed from 2012 to 2016 (10). However, this analysis was limited to patients between 20 and 49 years of age. Age is the most significant factor reported in cancer incidence and survival (2), and the risk of being diagnosed with cancer significantly increases in adults older than 50 years.

The purpose of this study was to use current population-based data to quantify differences in cancer incidence and survival by sex that was comprehensive for both age and race. In this study, we investigated sex differences in cancer incidence among those ages 0–85+ years from 2001 to 2016 using data from the United States Cancer Statistics (USCS) public use database, and for sex differences in cancer survival using the SEER dataset (2001–2016). Datasets were age-adjusted and only included diagnostically confirmed primary, malignant tumors from 2001 to 2016, resulting in a sample population of about 11.4 million for incidence and 2.9 million for survival. In addition, given the interests of our group, we examined sex differences in incidence and survival for specific brain tumor histologies (3, 11–13). Identifying the role of sex in cancer incidence and survival may improve outcomes by providing healthcare professionals additional insight regarding sex-specific approaches to treatment.

Population-based cancer incidence data by primary site, sex, race, and age groups were obtained from the USCS public use database, which combines data from the CDC's NPCR dataset and the NCI's SEER Program dataset. The USCS dataset includes de-identified cancer incidence data from central cancer registries (state) reported to NPCR from 46 states and the District of Columbia, and from SEER for 4 states, and represents 99.9% of the U.S. population (cases diagnosed in Mississippi from 2001–2002 were not available).

Survival data by primary site and sex were obtained from the NCI SEER 18 registries which represents approximately 28% of the U.S. population.

Cases were classified according to the International Classification of Diseases for Oncology, Third Edition (14). Primary site and histology groupings were classified by the site recode ICD-O-3/WHO 2008 definition provided by SEER. Only those with a malignant behavior code (/3) were kept for all analyses. For brain-specific analyses, histology groupings were defined by the Central Brain Tumor Registry of the United States (CBTRUS) Brain and Other Central Nervous System Tumor Histology Groupings, and primary sites C71.0–C72.5, C72.8–C72.9, C70.0–C70.9, C75.1–C75.3, and C30.0 (15) Data were restricted to patients with primary, first sequence tumors diagnosed between 2001 and 2016. Only diagnostically confirmed cases (microscopically or radiographically) were included. Sex-specific cancers (e.g., cervical, ovarian, and prostate) were excluded.

Self-reported race categories included in this study were White, Black, American Indian/Alaska Native (AIAN), and Asian/Pacific Islander (API). Other race, unspecified, and unknown race were included in analyses that were not race-specific. Age at diagnosis groups used in this analysis were 0–14, 15–39, 40–64, and 65 years and older. For brain-specific analyses, histology groupings were defined by the Central Brain Tumor Registry of the United States (CBTRUS) Brain and Other Central Nervous System Tumor Histology Groupings, and primary sites C71.0–C72.5, C72.8–C72.9, C70.0–C70.9, C75.1–C75.3, and C30.0 (15). Counts are not presented when fewer than 16 cases were reported for any specific site.

Data were collected from January 1, 2001 to December 31, 2016. Data analysis took place from June 18 to July 1, 2019.

Incidence rate ratios (IRR) with 95% confidence intervals (95% CI) were generated to compare male-to-female incidence across all cancers using the SEER*Stat software, version 8.3.5. IRRs were calculated by sex, sex and race, and sex and age at diagnosis. All rates were age-adjusted and standardized to the 2000 U.S. population (19 age groups–Census P25-1130) to adjust for differences in age distribution (16). 95% CIs were calculated using the method described in Tiwari and colleagues (17) Cox proportional hazard models were used to calculate male-to-female HRs adjusted for age in R 3.6.0. Statistical significance was set at P < 0.05.

Incidence was calculated for 14,281,801 patients [6,423,073 male (45.0%) and 7,858,728 female (55.0%)] diagnosed with cancer from 2001 to 2016 (Supplementary Table S1). IRRs were calculated by sex (Supplementary Table S1), sex and race (Supplementary Table S2), and sex and age (Supplementary Table S3). Survival analyses were conducted on 3,705,584 patients [1,654,454 male (44.6%) and 2,051,130 female (55.4%)].

Overall, males had lower cancer incidence (IRR: 0.958; 95% CI, 0.957–0.959; P < 0.001), but worse survival (HR, 1.568; 95% CI, 1.564–1.573; P < 0.001) compared with females (Supplementary Table S1 and S4).

Figure 1 and Table 1 depict age-standardized male-to-female incidence rate ratios from 2001 to 2016 for all major cancer sites and groupings. Highest male incidence rate ratios were exhibited in Kaposi sarcoma (IRR: 9.751; 95% CI, 9.287–10.242; P < 0.001), larynx (IRR: 4.567; 95% CI, 4.511–4.624; P < 0.001), mesothelioma (IRR: 4.112; 95% CI, 4.011–4.216; P < 0.001), and liver (IRR: 3.381; 95% CI, 3.349–3.412; P < 0.001).

Females exhibited higher incidence of cancer in breast (IRR: 0.010; 95% CI, 0.093–0.103; P < 0.001), peritoneum, omentum, and mesentery (IRR: 0.098; 95% CI, 0.093–0.103; P < 0.001), thyroid (IRR: 0.313; 95% CI, 0.311–0.315; P < 0.001), gallbladder (IRR: 0.543; 95% CI, 0.532–0.554; P < 0.001), anus, anal canal, and anorectum (IRR: 0.683; 95% CI, 0.672–0.693; P < 0 0.001), and appendix (IRR: 0.900; 95% CI, 0.883–0.919; P < 0.001).

Figure 2 shows age-adjusted male-to-female IRRs from 2001 to 2016 for all major cancer site groupings stratified by race. Across all sites, the male-to-female IRR was highest in whites (IRR: 0.955; 95% CI, 0.954–0.956; P < 0.001) and lowest in American Indians or Alaska Natives (IRR: 0.924; 95% CI, 0.919–0.930; P < 0.001; Supplementary Table S2). For cancers of the anus, anal canal, and anorectum, females had higher incidence compared with males for whites (IRR: 0.628; 95% CI, 0.617–0.638; P < 0.001), American Indians or Alaska Natives (IRR: 0.721; 95% CI, 0.575–0.901, P = 0.004), and Asians or Pacific Islanders (IRR: 0.774; 95% CI, 0.671–0.891, P < 0.001), while males had higher incidence among blacks (IRR: 1.107; 95% CI, 1.058–1.158, P < 0.001). Asian or Pacific Islander males exhibited particularly high incidence in larynx cancers (IRR: 8.199; 95% CI, 7.203–9.363; P < 0.001) compared to whites (IRR: 4.443; 95% CI, 4.383–4.503, P < 0.001), blacks (IRR: 5.123; 95% CI, 4.957–5.296, P < 0.001), and American Indians or Alaska Natives (IRR: 4.530; 95% CI, 3.809–5.414, P < 0.001).

Across site, cancer incidence was higher for females ages 15–39 and 65 and older, and higher for males ages 0–14 and 40–64. (Supplementary Table S3). Male-to-female IRRs were highest in children aged 0–14 years (IRR: 1.125; 95% CI, 1.113–1.136; P < 0.001). The IRR for appendix cancers was higher in female age groups 0-14, 15-39, and 40–64 years but higher in males 65 years or older (IRR: 1.087; 95% CI, 1.047–1.129; P < 0.001). Higher incidence of pancreas cancer was exhibited in females ages 0–14 years (IRR: 0.574; 95% CI, 0.417–0.786; P < 0.001) and 15–39 years (IRR: 0.944; 95% CI, 0.897–0.993; P = 0.03), whereas males expressed higher incidence in the 40 to 64 years (IRR: 1.476; 95% CI, 1.460–1.492; P < 0.001) and 65 years or older (IRR: 1.17; 95% CI, 1.16–1.18, P < 0.001) age groups. Similarly, in melanomas of the skin, incidence was higher in females ages 0–14years (IRR: 0.912; 95% CI, 0.832–0.999; P = 0.05) and 15–39 years (IRR: 0.579; 95% CI, 0.572–0.586; P < 0.001), and higher in males ages 40–64 years (IRR: 1.248; 95% CI, 1.24–1.256, P < 0.001) and 65 years or older (IRR: 2.319; 95% CI, 2.302–2.336; P < 0.001). Male-to-female incidence of trachea, mediastinum, and other respiratory organs cancers was highest in the 15–39 age group (IRR: 7.051; 95% CI, 6.171–8.085; P < 0.001). Kidney and renal pelvis cancers were characterized with higher incidence in males in all age groups except 0–14 years (IRR: 0.864; 95% CI, 0.828–0.902; P < 0.001). Incidence of mesothelioma in males was lowest in the 15–39 age group (IRR: 0.801; 95% CI, 0.685–0.935; P = 0.005) and highest in adults 65 years or older (IRR: 5.160; 95% CI, 5.005–5.32; P < 0.001). Incidence of Kaposi sarcoma in males was lowest in the 65 years or older age group (IRR: 3.048; 95% CI, 2.846–3.265; P < 0.001), compared with IRRs of 22.935 and 19.924 in ages 15–39 and 40–50, respectively.

Figure 3 shows age-adjusted male-to-female HRs from 2001 to 2016 for all major cancer site groupings. Males experienced better overall survival in Kaposi sarcoma (HR, 9.751; 95% CI, 9.287–10.242; P < 0.001), other lymphocytic leukemia (HR, 0.812; 95% CI, 0.721–0.914; P < 0.001), cancers of the larynx (HR, 0.926; 95% CI, 0.894–0.960; P < 0.001), and oropharynx (HR, 0.878; 95% CI, 0.800–0.964; P = 0.01), despite having a higher incidence compared with females in those specific cancers (Supplementary Table S4). Females were observed to have better overall survival in most other cancers, particularly cancers of the thyroid (HR, 1.773; 95% CI, 1.707–1.845; P < 0.001), salivary gland (HR, 1.610; 95% CI, 1.515–1.712; P < 0.001), breast (HR, 1.540; 95% CI, 1.474–1.610; P < 0.001), anus, anal canal, and anorectum (HR, 1.488; 95% CI, 1.421–1.558; P < 0.001), and melanoma of the skin (HR, 1.410; 95% CI, 1.383–1.437; P < 0.001).

Males had significantly lower incidence and worse survival compared to females in cancers of the thyroid (IRR: 0.313; HR, 1.773), breast (IRR: 0.010; HR, 1.540), gallbladder (IRR: 0.543; HR, 1.066), anus, anus canal, and anorectum (IRR: 0.683; HR, 1.488), and cranial nerves and other nervous system (IRR: 0.956; HR, 1.434; Supplementary Tables S1 and S4). There are no sites in which males had lower incidence and better survival. Incidence was higher for males in all four sites that males experienced better overall survival (Kaposi sarcoma, other lymphocytic leukemia, larynx, and oropharynx cancers).

Given our group's research focus in sex differences in brain tumors, we performed additional analyses on specific malignant brain tumor histologies (Figs. 4 and 5; Supplementary Table S4 and S5; refs. 11–13, 15). Among these, males had the highest incidence rate ratio of germ cell tumors (IRR: 3.03; 95% CI, 2.798–3.284; P < 0.001), hematopoietic neoplasms (IRR: 1.842; 95% CI, 1.540–2.206; P < 0.001), and glioblastoma (IRR: 1.554; 95% CI, 1.537–1.570; P < 0.001). Females had higher incidence than males among other neuroepithelial tumors (IRR: 0.425; 95% CI, 0.305–0.587; P < 0.001), and malignant meningioma (IRR: 0.731; 95% CI, 0.687–0.777; P < 0.001). Males had shorter survival among nerve sheath tumors (HR, 2.739; 95% CI, 1.589–4.720; P < 0.001), neoplasms related to the meninges (HR, 1.859; 95% CI, 1.072–3.223; P = 0.03), and malignant meningioma (HR, 1.488; 95% CI, 1.260–1.757; P < 0.001). Females only exhibited shorter survival among germ cell tumors (HR, 0.720; 95% CI, 0.526–0.985; P = 0.04).

Using data representative of 99.9% of the U.S. population for incidence and data from approximately 28% of the population for survival, this analysis of cancer incidence and survival across a 15-year period identified significant differences by sex. Males had significantly higher incidence and worse survival outcomes in the majority of cancer sites, consistent with what has been previously reported (1, 10). There was further variation of incidence rates by age groups, with females' ages exhibiting higher cancer incidence than males in the 15–39 years age group. The male-to-female IRR was similar among all examined race groups.

Sex disparities in cancer incidence and survival can be attributed to behavioral and environmental factors as well as biological differences. Higher incidence and worse survival in males can be partially attributed to increased tobacco use (18) compared with females and increased exposure to oncogenic agents such as oral HPV (19). Males have also been found to be more likely to engage in high-risk behaviors and less likely to utilize health care services than females. Occupational exposures to carcinogens such as asbestos may influence sex differences in incidence and survival. Sex differences in immune system functions (20), along with genetic and hormonal differences (21), are also likely to play a role in the observed sex disparities.

Previous analyses have found that, in contrast to other cancers, females are more likely to be diagnosed with melanoma than males until age 40 years (22). This pattern was present in the incidence analyses of this study, as the male-to-female incidence rate ratio was less than 1 for the age groups 0–14 and 15–39 years, and more than 1 for age 40–59 and 60 years or older. Numerous studies suggest that the increased incidence of melanoma in adolescent and young adult women is influenced by the increased usage of indoor tanning facilities (23). Male hormones and physiologic sex differences are possible factors that may contribute to the higher overall male incidence of skin and other cancers (24).

On the basis of our findings, men have significantly lower incidence, but worse survival, for breast cancer. These results are supported by previous studies that have found similar trends (25, 26). These studies noted that the majority of male breast cancer cases are associated with BRCA2 mutations, as well as Klinefelter syndrome. Studies have also noted that a lack of screening and early detection, and the impact this has on stage migration, contributes to these lower survival outcomes (27).

In this study, thyroid cancers were observed to have lower incidence and worse survival in males compared with females. There have been few molecular explanations for sex differences in thyroid cancers, though estrogen receptors may have an important role (21, 28). Other studies have suggested that a substantial percentage of thyroid cancer diagnoses could be due to factors such as obesity and smoking. Higher female incidence may be influenced by recent advances in thyroid cancer detection, possibly contributing to overdiagnosis and overtreatment of female patients relative to male patients (24, 28, 29).

Kaposi sarcoma had the highest male-to-female IRR and the lowest female survival in this analysis. A recent study found that despite advances in treatment, HIV-infected persons have an 800-fold elevated risk of Kaposi sarcoma relative to the general population (29). However, male incidence of Kaposi sarcoma has declined considerably since the beginning of the HIV/AIDS epidemic, as evidenced by a similar analysis of SEER data from 1975 to 2004 that exhibited an incidence rate ratio approximately three times larger than the ratio calculated in this study. The survival disadvantage for females has not been well-studied in the United States, but an analysis of the incidence and clinical outcomes of HIV-positive patients with Kaposi sarcoma in Uganda suggested that females were more likely to experience more advanced and severe case of Kaposi sarcoma. Delayed diagnosis of Kaposi sarcoma may account for sex differences in survival, though more investigation is necessary as other studies showing similar patterns are over a decade old (30, 31).

This study identified sex differences in multiple brain tumor histologies, such as germ cell tumors, malignant meningioma, and glioblastoma. Glioblastoma is the most common malignant brain tumor and is also the most fatal. We found males with this tumor had both increased incidence and risk of death. There are not many proven risk factors for this tumor, but previous studies have identified potential sex-related genetic factors related to incidence and survival (3, 11–13, 32). This study demonstrates the necessity for detailed investigations into these potential differences allowing us to uncover the biological underpinnings of these sex differences thereby helping us to advance treatments that could be tailored for males and females separately.

There are several limitations to this analysis that are worth noting. Survival data were unavailable in the dataset used for incidence analyses, and therefore a subset was used. In addition, the survival analyses were conducted only by sex differences with adjustment for age, and did not assess potential variations in sex difference by race. Furthermore, there may be other variables not considered in this analysis or not available in the datasets used for analysis that could influence sex disparities in cancer incidence and survival, such as screenings, treatments, molecular testing, comorbidities, and risk factors (smoking, obesity, etc.).

To our knowledge, this study represents the most recent comprehensive reporting of cancer incidence and survival by sex in the United States. Cancer registry data for incidence included approximately 99.9% of the U.S. population while SEER survival data represented approximately 28% of the population (1, 2). Among patients diagnosed with non-sex-specific cancers from 2001 to 2016, males had higher incidence and worse overall survival compared with females in the vast majority of cancer sites. In addition, male-to-female incidence varied by race and age groups. Further examination is necessary to obtain a better understanding of sex disparities in cancer etiology and prognosis, as well as the clinical implications of observed sex differences.

No potential conflicts of interest were disclosed.

Conception and design: G. Cioffi, J.S. Barnholtz-Sloan

Development of methodology: G. Cioffi, J.S. Barnholtz-Sloan

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): M. Dong, K.A. Waite, J.S. Barnholtz-Sloan

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): M. Dong, G. Cioffi, J. Wang, J. Connor, J.S. Barnholtz-Sloan

Writing, review, and/or revision of the manuscript: M. Dong, G. Cioffi, J. Wang, K.A. Waite, Q.T. Ostrom, C. Kruchko, J.D. Lathia, J.B. Rubin, M.E. Berens, J.S. Barnholtz-Sloan

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): K.A. Waite, J.S. Barnholtz-Sloan

Study supervision: J.S. Barnholtz-Sloan

This work was funded by departmental funds. The authors acknowledge Crystal Miller, the director of the Science Research & Engineering Program (SREP), for her mentorship. They also acknowledge Nirav Patil and Sindoosha Malay for their guidance and collaborative discussion.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.