Increasing Incidence of Testicular Germ Cell Tumors among Racial/Ethnic Minorities in the United States

Testicular germ cell tumors (TGCT) are rare in the general population, but are the most commonly occurring malignancy among men ages 15 to 44 years in the United States (1). TGCTs are classified by histologic subtype as seminomas, nonseminomas, and spermatocytic tumors. Seminomas and nonseminomas comprise the majority of TGCTs (98%–99%) and have a peak incidence at 35 and 25 years of age, respectively. Spermatocytic tumors account for only 1% to 2% of TGCTs, and have a peak incidence at 55 years of age (2).

Incidence rates of TGCT have been rising in the United States since the mid-20th century (1, 3). Although rates are notably higher among non-Hispanic white (NHW) men than men of other ancestries, recent studies have reported that incidence rates have been rising in other racial/ethnic groups, especially Hispanics (1, 4, 5). A prior study by our group found that between 1998 and 2011, the largest increase in TGCT incidence was experienced by Hispanics, followed by only a slight increase in rates among NHWs (4). Incidence rates also increased, but not significantly, among Asian/Pacific Islander (A/PI) men. Reasons for the reported increases are unclear as there are few well-identified risk factors for TGCT.

The availability of more recent data on incidence in the United States prompted further examination of TGCT incidence trends by race/ethnicity. As such, the purpose of this study was to examine whether the patterns noted in our prior publication remained evident in more recent years and to determine whether any new patterns have been emerging.

Data for this study were drawn from the U.S. Cancer Statistics (USCS) public use databases maintained by the Centers for Disease Control and Prevention (CDC; ref. 6). High-quality, population-based cancer incidence data reported to the CDC's National Program of Cancer Registries (NPCR) and the National Cancer Institute's Surveillance, Epidemiology, and End Results (SEER) programs are combined by the USCS. Registries included in the current analysis met USCS's data quality criteria for the November 2018 submission cycle. Data from 51 registries, including the 50 states and the District of Columbia, were included for the years 2001 to 2016, covering approximately 99% of the U.S. population.

TGCT was defined using the International Classification of Diseases for Oncology (3rd ed.) topography (C62) and morphology codes (seminoma: 9060–9062, 9064; nonseminoma: 9065–9102; spermatocytic tumors: 9063; ref. 7). Data on race, Hispanic ethnicity, histology, and year of diagnosis were available for each case. Identification of Hispanic ethnicity was based on the North American Association of Central Cancer Registries (NAACCR) Hispanic/Latino Identification Algorithm (NHIA; ref. 8). Incidence rates per 100,000 man-years, age-adjusted to the U.S. 2000 standard population, and their 95% confidence intervals were calculated. Estimates of annual percent change (APC) were calculated for the 2001 to 2016 time period using the annual rates and weighted least squares regression (9). For temporal analysis, years of diagnosis were grouped into four periods: 2001 to 2004, 2005 to 2008, 2009 to 2012, and 2013 to 2016. For geographical analyses, states were grouped into the four census regions (Northeast, Midwest, South, and West) based on the U.S. Census Bureau's system of State groupings (10). Statistics were suppressed where there were fewer than 16 cases, for purposes of confidentiality. The Wald test was used to examine differences in rates by geographic region. Statistical significance was based on a two-sided P value <0.05. The Wald test was calculated using SAS (version 9.4). APCs were estimated using the Joinpoint Regression Program (version 4.7.0.0). All other statistical analyses were performed using the SEER*Stat statistical package (version 8.3.6).

During the years 2001 to 2016, 126,575 TGCTs (72,171 seminomas; 53,476 nonseminomas; 928 spermatocytic tumors) were recorded by the USCS registries (Table 1). TGCT incidence rates were highest among NHW men (6.63 per 100,000 man-years), followed by Hispanic (4.20), AI/AN (3.27), A/PI (1.72), and NHB men (1.27; Table 1). Rates of both seminomas and nonseminomas followed the same ranking. Among all men, temporal analysis showed that TGCT incidence modestly increased (APC: 0.41, P < 0.01), although the increase was due to a significant increase only in nonseminoma (APC: 0.97, P < 0.01) as rates of seminoma showed little change (APC: 0.05, P = 0.62). Spermatocytic tumor rates decreased significantly (APC: −1.57, P = 0.02). The majority of tumors were diagnosed at localized stages in all men (68.8%; Table 2); however, non-Hispanic white men had the highest percentage of localized disease (70.1%) whereas non-Hispanic black men had the lowest percentage (61.9%).

Trends in TGCT incidence by race/ethnicity are shown in Table 3 and Fig. 1. While rates increased only slightly, but significantly, among NHW men (APC: 0.41, P < 0.01), larger increases were seen in all other groups. The largest increase was experienced by A/PI men (APC: 2.47, P < 0.01), then Hispanics (APC: 2.10, P < 0.01), AI/ANs (APC: 1.71, P = 0.03), and NHBs (APC: 1.28, P < 0.01). Incidence trends by histologic subtype are also presented in Table 3 and Fig. 1. Seminoma rates increased significantly among AI/AN (APC: 2.25, P = 0.02), A/PI (APC: 1.81, P = 0.03), Hispanic (APC: 1.40, P < 0.01), and NHB (APC: 1.37, P = 0.01) men but not among NHW men (APC: 0.14, P = 0.22). Nonseminoma rates increased significantly among almost all racial/ethnic groups: A/PI (APC: 3.34, P < 0.01), Hispanic (APC: 2.98, P < 0.01), NHB (APC: 1.28, P = 0.03), and NHW (APC: 0.82, P < 0.01) men. Nonseminoma rates also increased among AI/AN men (APC: 1.04, P = 0.29), but not significantly so. Joinpoint analyses revealed no significant joinpoints for any of the histologic subtype and racial/ethnic group combinations.

TGCT rates for the most recent time period (2013–2016) by census region are presented in Table 4. TGCT rates among Hispanic (5.38), AI/AN (4.47), and A/PI (2.37) men were highest in the West whereas rates among NHW (7.60) and NHB (1.51) men were highest in the Northeast. There was a statistically significant difference in rates by region among NHW (P < 0.01), Hispanic (P < 0.01), A/PI (P < 0.01), and AI/AN (P < 0.01) men, but not among NHB men (P = 0.28). Similar findings were observed for seminomas and nonseminomas, where statistically significant regional rates were observed for all groups except NHBs. Seminoma and nonseminoma rates among AI/AN men were not calculated for all census regions due to small numbers.

Trends in TGCT incidence by census region among NHW, Hispanic, NHB, and A/PI men are shown in Fig. 2A. Incidence among NHW men was stable in the West (APC: 0.10, P = 0.51), and the South (APC: 0.33, P = 0.05), but modestly increased in the Northeast (APC: 0.71, P < 0.01) and the Midwest (APC: 0.62, P < 0.01). Among Hispanic men, incidence increased significantly in all four regions: Midwest (APC: 3.89, P < 0.01), West (APC: 2.79, P < 0.01), Northeast (APC: 1.84, P < 0.01), and South (APC: 0.97, P = 0.02). Among NHB men, incidence increased significantly in the South (APC: 1.74, P < 0.01), but remained stable in the Northeast (APC: 0.34, P = 0.73), and the Midwest (APC: 0.69, P = 0.43). Although rates increased in the West (APC: 1.70, P = 0.16), the increase was not statistically significant. Among A/PI men, rates increased significantly in the West (APC 3.03, P < 0.01).

Trends in the incidence of seminoma by census region are shown in Fig. 2B. Among NHWs, the only significant increase was seen in the Northeast (APC: 0.47, P = 0.01). Among Hispanics, there were significant increases in both the Midwest (APC: 3.16, P < 0.01) and the West (APC: 2.18, P < 0.01), whereas among NHBs, the only significant increase was in the South (APC: 1.99, P = 0.02). Among A/PIs, rates increased significantly in the West (APC 2.30, P = 0.04). Trends in the incidence of nonseminoma by census region are shown in Fig. 2C. Among NHWs, there were significant increases in the Northeast (APC: 1.03, P < 0.01), Midwest (APC: 1.00, P < 0.01), and South (APC: 0.80, P < 0.01). Among Hispanic men, there were significant increases in all four regions: Midwest (APC: 4.76, P < 0.01), West (APC: 3.51, P < 0.01), Northeast (APC: 2.46, P < 0.01), South (APC: 2.11, P < 0.01). Among NHB men, rates increased significantly in the South (APC: 1.50, P = 0.04) whereas among A/PI men, rates increased significantly in the West (APC 4.14, P < 0.01).

Hispanic men had the greatest APC for TGCT and both histologic subtypes in the Northeast and Midwest and the same was true for A/PI men in the West (Fig. 2A–C). Hispanic men also had the greatest APC in nonseminoma rates in the South. NHB men had the greatest APC for TGCT and seminoma in the South. Joinpoint analyses revealed two significant findings; a joinpoint was observed among NHW men in the Northeast for TGCT (APC 2001–2012: 1.24, P < 0.01; APC 2012–2016: −1.57, P = 0.11) and seminoma (APC 2001–2011: 1.05, P < 0.01; APC 2011–2016: −1.15, P = 0.11). APCs among A/PI men could not be calculated in any census region, except the West, due to small numbers. Similarly, APCs could not be calculated for AI/AN men by census region.

State-level data on TGCT incidence are presented in Fig. 3 and by histologic subtype in Supplementary Table S1. The states with the highest rates of TGCT were New Hampshire (7.07 per 100,000 man-years) and Utah (6.92). The states that experienced the greatest increase in rates were Hawaii (APC: 2.07, P = 0.03) and Kansas (APC: 1.94, P = 0.01). TGCT rates were lowest in Mississippi (3.59 per 100,000 man-years) and Georgia (4.00).

This study found that although NHW men had the highest incidence of TGCT between 2001 and 2016, the greatest increase in incidence occurred among A/PI men. By histologic type, A/PI men experienced the greatest increase in nonseminoma rates, whereas AI/AN men experienced the greatest increase in seminoma rates. Geographic analyses revealed that Hispanic men were the only racial/ethnic group to experience significant increases in TGCT incidence in all four U.S. census regions.

As in our prior report (4), the current analysis found a significant increase in TGCT incidence among Hispanic men and a modest, yet significant, increase among NHW men. In contrast, however, this study found significant increases in TGCT incidence among men of all other racial/ethnic ancestries, including NHB, A/PI, and AI/AN men. Furthermore, our previous study reported that Hispanic men were the only racial/ethnic group to experience increases in rates for both seminomas and nonseminomas. This study, however, found significant increases in seminoma and nonseminoma rates among not only Hispanic men, but to a greater degree among A/PI men, and among NHB men.

Asians are the fastest growing population group in the United States, with a growth rate between 2000 and 2010 more than four times that of the total U.S. population (11). The main driving force for this rapid growth has been attributed to international migration (12). A comparison of rates among A/PI men in the United States (1.7 per 100,000) to rates among men in Asia finds that rates are somewhat higher in Japan (2.1), but lower in China (1.5), the Philippines (0.8), Thailand (0.5), and India (0.6). It is possible that rates among United States A/PIs have risen with migration. Prior studies of migrants from lower to higher rate countries have reported that changes in TGCT incidence do not occur among the first generation of migrants, but in subsequent generations (13, 14). Thus, it is possible that the increase in TGCT rates among A/PIs could be a result of exposures that are more prevalent in the United States. In this study, this could not be examined as migration status was not available. A previous study examining TGCT rates in the SEER-13 registries (1992–2004) found that TGCTs were 1.4-fold more likely to develop in A/PIs than in NHWs (15). This study, however, only examined rates among boys of ages 0 to 14 years. The authors hypothesized that the rate difference might reflect a difference in tumorigenesis related to early-life endocrine disrupting chemical exposure.

The significant increase in TGCT rates among NHB and AI/AN men is a noteworthy finding. A previous study examining trends in TGCT incidence among black men found that rates increased between 1988 and 1992 (16). In studying the subsequent time period, 1992 to 2011, our group found that incidence rates among black men continued to increase significantly (17). These studies, however, did not distinguish NHB men from Hispanic black men. Two other studies conducted by our group examined TGCT rates among NHB men and found no significant increase in TGCT incidence rates (1, 4). This study, however, found that rates among NHB men increased for both histologic subtypes and for overall TGCT, mainly in the South. Although the rates of TGCT among NHB men remain notably lower than the rates of other racial/ethnic groups, this study's significant finding suggests that TGCT among NHB men should not be disregarded. Similarly, this is the first study to demonstrate a statistically significant increase in TGCT rates among AI/AN men. Although previous studies have noted an increase in rates among AI/AN men, those increases have never reached a level of statistical significance, possibly due to the misidentification of AI/AN status.

It remains unclear why there are differences in the risk of TGCT among racial/ethnic groups in the United States, or in other countries. The only established risk factors for TGCT include family history of TGCT (18, 19), cryptorchidism (20), hypospadias (21), and impaired spermatogenesis (22). The collection of these male reproductive disorders, collectively known as the testicular dysgenesis syndrome (TDS), has been suggested to have an in utero etiology (23). Whether the prevalence of each of the TDS conditions varies by racial/ethnic group, however, remains unclear.

Genetic susceptibility to TGCT has been evidenced by the increased risk among first degree relatives of men with TGCT (19, 24–26) and by the results of genome-wide association studies (GWAS). Approximately 40 to 50 loci, thus far, have been found by GWAS to be associated with TGCT susceptibility (27–34). Among these loci, genetic variation in the KIT ligand (KITLG) gene has been found to have the strongest association (33) with some SNPs varying by ancestry. For example, the KITLG rs3782181 SNP has an allele frequency that varies greatly between men of African ancestry (A: 70%, G: 30%) and men of other ancestries, such as European (A: 20%, G: 80%), Hispanic/Latino (A: 17%, G: 83%), and east Asian (A: 23%, G: 77%). Thus, it is possible that variation in genes which are associated with TGCT risk, combined with the presence of an environmental risk factor, can lead to an additive or multiplicative increase in TGCT risk as a gene–environment interaction (35).

Although genetic susceptibility to TGCT may be related to some of the difference in incidence by race/ethnicity, it is unlikely to explain the decades long continuing increase in rates. Such increases are consistent with environmental risk factors playing an important role in the TGCT etiology (36). Prenatal exposures such as maternal nutrition and exposure to endocrine disrupting chemicals (EDC), as well as other factors, have been suggested to be related to risk (37). An assessment of prenatal exposures, however, is difficult as the exposure and the outcome are among different persons (mothers and sons). Meta-analyses, however, have found that TGCT is significantly associated with cryptorchidism, inguinal hernia, twinning, low birth weight, short gestational age, maternal bleeding, birth order, sibship size, and possibly caesarean section (20, 38). Postnatal exposures that have been associated with TGCT include occupational exposures such as firefighting (39–44) and aircraft maintenance (45–49), age at puberty (50), adult stature (51), cannabis use (52–56), and exposure to particular EDCs (57, 58). The prevalence of cannabis use in the United States has recently increased in the general population, possibly due to the changing cannabis laws (59). Although it is possible that cannabis use could explain some of the increase in TGCT rates, most published studies on cannabis and TGCT are limited by being based on self-reported data from retrospective studies.

In this study, the increase in TGCT incidence was driven by an increase in nonseminoma rates. The peak age at diagnosis of nonseminoma is 25 years, whereas the peak age at diagnosis of seminoma is 10 years older, at age 35 years. The younger age at presentation of nonseminoma may explain why it is the most common TGCT among Hispanic men, as the Hispanic population is younger than other populations in the United States. Why nonseminoma rate increases would be statistically significant, whereas seminoma increases were not, is unclear as risk factors that differ between the two types, other than marijuana smoking, have yet to be identified. It is important to note, however, that although nonseminoma rates are propelling the increase in TGCT rates, nonseminomas are less common than seminomas among all groups, with nonseminoma to seminoma ratios of NHWs = 0.71, Hispanics = 0.91, NHBs = 0.54, A/PIs = 0.79, and AI/ANs = 0.78.

Although based on small numbers, a notable finding of this study is the significant decrease in the rates of spermatocytic tumor. Spermatocytic tumors are not clinically aggressive tumors and rarely warrant treatment other than orchiectomy (60). Even though the median age at diagnosis of spermatocytic tumor is age 50 years, the incidence is always lower than that of seminoma. Although previous studies have found modest, nonsignificant, increases (61) and decreases (4) in the rates of spermatocytic tumor, the finding a statistically significant decrease in incidence is novel.

In the most recent time period, A/PI men had the highest incidence of TGCT in the West (2.37 per 100,000 man-years), followed in order by the Midwest (1.90), the Northeast (1.84), and the South (1.32). The highest APC was also observed in the West. It is conceivable that the ancestral make-up of the A/PI population in various parts of the country contributed to the regional difference in rates, as cancer statistics are heterogenous among the various A/PI populations (62–64). The data used in the current analysis, however, did not permit the disambiguation of the A/PI population into its constituent components. Thus, further study of may provide insight into whether particular A/PI groups have greater susceptibility to TGCT.

A strength of this study was the use of population-based cancer registry data from 51 registries, which included virtually the entire U.S. population (99%). In addition, this study was able to use state-level TGCT data. Limitations included the inability to restrict AI/AN populations to PRCSDA, which better captures AI/AN status. In addition, it was not possible to disaggregate the A/PI population, thus the study could have failed to detect any differences that might exist within the component A/PI groups. Another limitation was the use of the NAACCR NHIA algorithm to categorize Hispanic ethnicity. This study was restricted to men, however, and as such, there was less chance for misclassification. The algorithm is based partially on surname and, as women are more likely to change their surnames, there is greater chance of misclassifying women than men. Finally, this study lacked information on birthplace and country-specific ancestry, as well as information on A/PI and Hispanic subpopulations, which could provide further insight into whether subpopulations may have greater genetic susceptibility to TGCT.

In summary, this study suggests that TGCT incidence is increasing among men of all racial/ethnic backgrounds in the United States, with the most notable increases in A/PI and Hispanic men in the West. Rising rates of TGCT among men of non-European ancestry suggests that both etiologic research and public health efforts among these populations are warranted. Reasons for the increase in rates could be related to as yet unidentified environmental exposures and/or genetic susceptibility to TGCT.

No potential conflicts of interest were disclosed.

Conception and design: A.A. Ghazarian, K.A. McGlynn

Development of methodology: A.A. Ghazarian, K.A. McGlynn

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): A.A. Ghazarian

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): A.A. Ghazarian, K.A. McGlynn

Writing, review, and/or revision of the manuscript: A.A. Ghazarian, K.A. McGlynn

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): A.A. Ghazarian

Study supervision: K.A. McGlynn

This work was supported by the Intramural Research Program of the NCI.

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.