Trends in intracranial meningioma incidence in the United States, 2004‐2015

Abstract Background Meningioma incidence was reported to have risen substantially in the United States during the first decade of the 21st century. There are few reports about subsequent incidence trends. This study provides updated data to investigate trends in meningioma incidence by demographic and tumor characteristics at diagnosis in the United states from 2004 to 2015. Methods Trends in meningioma incidence were analyzed using data from the Surveillance, Epidemiology, and End Results‐18 (SEER‐18) registry database of the National Cancer Institute. The joinpoint program was used to calculate annual percent change (APC) in incidence rates. Results The overall incidence of meningioma increased by 4.6% (95% CI, 3.4‐5.9) annually in 2004‐2009, but remained stable from 2009 to 2015 (APC, 0; 95% CI, −0.8 to 0.8). Females (10.66 per 100 000 person‐years) and blacks (9.52 per 100 000 person‐years) had significant predominance in meningioma incidence. Incidence in many subgroups increased significantly up to 2009 and then remained stable until 2015. However, meningioma incidence in young and middle‐aged people increased significantly throughout the entire time period from 2004 to 2015 (APC: 3.6% for <20‐year‐olds; 2.5% for 20‐39‐year‐olds; 1.8% for 40‐59‐year‐olds). The incidence of WHO II meningioma increased during 2011‐2015 (APC = 5.4%), while the incidence of WHO III meningioma decreased during 2004‐2015 (APC = −5.6%). Conclusion In this study, the incidence of meningioma was found to be stable in recent years. Possible reasons for this finding include changes in population characteristics, the widespread use of diagnostic techniques, and changes in tumor classification and risk factors in the US population.


| INTRODUCTION
Meningiomas are generally benign tumors that originate in cerebral dura mater and can grow at any site, especially at the skull vault and the skull base, with 10% of meningioma located in the spinal cord. 1 Although most meningiomas are asymptomatic 2 and patients often only experience mild headaches in the early stages, there are a large number of complications and poor functional outcomes when tumors progress. According to the latest Central Brain Tumor Registry of the United States (CBTRUS) statistical report, 3 meningioma was the most frequently reported histologic type, accounting for 37.1% of all CNS tumors and 53.1% of nonmalignant CNS tumors from 2011 to 2015. Meningioma also has the highest incidence rate, that is, 8.33 per 100 000 person-years in among CBTRUS histology groupings and an estimated 31 990 new cases in 2019.
Few studies have focused on the trends in the incidence of meningioma. A study of patients aged over 65 years old with meningioma found nonmalignant meningioma incidence increased significantly for both females (APC = 4.69%) and males (APC = 4.79%) from 2005 to 2009, whereas the incidence of malignant meningioma decreased significantly for females (APC = −5.45%) and males (APC = −2.88%) during 2005-2015. 4 Furthermore, Kshettry et al revealed that WHO II meningioma incidence increased by 3.6% per year from 2004 to 2010 and WHO III meningioma incidence decreased by 5.4% per year during 2000-2010. 5 However, these studies did not cover all populations or all types of meningioma. A report from the UK demonstrated that the meningioma incidence remained stable for the local population over the study period. 6 However, comprehensive investigations on the trends in meningioma incidence in the United States are lacking.
In our study, we used data from 2004 to 2015 of the SEER registry database to provide an updated report focusing on meningioma incidence trends according to demographic and tumor characteristics at diagnosis. Data

| Patients' characteristics
We investigated trends in meningioma incidence according to demographic and tumor characteristics obtained from medical records of various registries. Demographic characteristics mainly included sex, race, and age at diagnosis (divided into the following subgroups <20, 20-39, 40-59, 60-79, and ≥80 years old).
Tumor characteristics included histologic types, WHO grade, and tumor size. Histologic types were classified into meningioma NOS, meningothelial, fibrous, psammomatous, angiomatous, hemangioblastic, transitional, clear cell, atypical meningioma, and meningeal sarcomatosis by ICD-O-3 codes (Table 1). Only the first matching record and tumors located in cerebral meninges (ICD-O-3 topography code C70.0), meninges NOS (ICD-O-3 topography code C70.9), or the brain (ICD-O-3 topography code C71) were included. Meningiomas were divided into grade I, grade II, and grade III according to National Comprehensive Cancer Network guidelines and ICD-O-3 behavior code. Collaborative Staging codes (CS) were used to determine tumor size during 2004-2015. Diagnostic confirmation was provided by the special code of the SEER database.

| Statistical analysis
Data regarding all cases and incidence rates were obtained from SEER*Stat version 8.3.5 (https ://seer.cancer.gov/seers tat/). All incidence rates were age-adjusted to the 2000 US standard population and expressed per 100 000 person-years. The Joinpoint Regression Program, version 4.6.0.0 (https :// surve illan ce.cancer.gov/joinp oint/), was used to evaluate the trends in incidence rates during the time accessed, and the simplest joinpoint model allowed by the data was applied. The Monte Carlo permutation method was used to evaluate an apparent change in trend. 8 A two-sided P value of 0.05 was considered statistically significant.

| RESULTS
Of the 83 030 patients ( Age-adjusted incidence rates according to demographic and tumor characteristics are represented in Table 2. The overall age-adjusted incidence rate of meningioma was 7.92 (95% CI, 7.86-7.97) per 100 000 person-years during 2004-2015. The incidence rate of women was about two times higher than that of men (10.66 vs 4.75 per 100 000 person-years). In terms of race, black people (9.52 per 100 000 person-years) had the highest incidence, followed by white people (7.82 per 100 000 person-years), Asian/Pacific Islander (APIs) (6.56 per 100 000 person-years), and American Indian/Alaskan Native (AIANs) (4.76 per 100 000 person-years). The incidence rate increased sharply with age, from 0.13 per 100 000 person-years for patients <20 years old to 46.83 per 100 000 person-years for those ≥80 years old. When stratifying by histologic type, the incidence rate of meningioma NOS was highest at 6.68 per 100 000 person-years, while the incidence rates of other histological types were extremely low. In terms of WHO grade, grade I meningioma had the highest incidence at 7.48 per 100 000 person-years and grade III meningioma had the lowest incidence at just 0.10 per 100 000 person-years. According to tumor size, the age-adjusted incidence of meningioma was 4.23 per 100 000 person-years for tumors ≤3 cm, 1.44 per 100 000 person-years for tumors >3 to ≤5 cm, and 0.74 per 100 000 person-years for tumors >5 cm.
Our research also explored the diagnostic confirmation method (Figure 4). The interesting finding was that there were more cases with microscopic diagnosis than those with radiographic diagnosis in the earliest period of our research. However, the proportion of radiographic diagnosis has increased in recent year, even exceeding 60% of the total in 2012, 2014, and 2015.
The annual number of cases and age-adjusted incidence rates from 2004 to 2015 are shown in Table S1-S6 in the supplementary data.

| DISCUSSION
Many studies have reported an increase in the incidence of meningioma in recent decades, not only in the United States, 3 but also in many other countries. 9-13 However, not all cancer registries in the United States were required to collect data on nonmalignant tumors until 2004, which restricted us to conduct a longitudinal analysis. Thus, we aimed to investigate temporal trends in the incidence of meningioma by demographic and tumor characteristics using updated 12-year data from SEER registries. Our main finding was that overall meningioma incidence in- The SEER-18 registry database  .402 Age at diagnosis <20 a 3.6 (0.5 to 6.9) . .289 Histologic type Meningioma,

Note:
Each time period is determined by Joinpoint Program when a statistically significant change in the APC occurred.
a Denotes only an overall trend for this subgroup after joinpoint analysis.
T A B L E 3 (Continued) to overall incidence, we also found that many APCs increased up to 2009 and then remained stable until 2015, for example, both sexes, white people, age ≥60 years old, meningioma NOS, WHO grade I, and medium-sized (＞3 to ≤5 cm) meningioma. This finding may be related to the changes in population characteristics, the widespread use of diagnostic techniques, changes in tumor classification and risk factors. There was also a period of significant increase in the incidence rates in our study, consistent with previous studies. To our knowledge, the incidence of meningioma increases sharply with age and peaks at 46.83 per 100 000 person-years in patients above 80 years old in our study, which was dozens of times higher than young patients. Interestingly, the incidence rate of meningioma in the elderly patients (>60 years old) did not increase between 2009 and 2015, consistent with the overall trend. However, the incidence of meningioma in young people showed a significantly increasing trend during 2004-2015. We used age-adjusted incidence, thus eliminating the difference in age distribution among the population. However, an aging population may contribute to an increase in crude incidence rates. A wider availability of CT and MRI examinations was responsible for the increased in incidence of meningioma. The effects of this aspect may continue into this century, and may lead to an increase in the possibility to detect incidental meningioma. A related study reported asymptomatic incidental meningioma was the most frequent brain tumor in the general population, with a prevalence between 0.29% and 0.90%. 14,15 After the significant increase in detecting meningioma linked to more powerful imaging equipment (such as CT and MRI), the increase in meningioma incidence will reach a plateau level. This may be a F I G U R E 1 Trends in annual meningioma incidence rates by demographic characteristics (2004-2015). A, shows overall meningioma incidence rates, and incidence by sex. B, shows meningioma incidence rates by race. C, shows meningioma incidence rates by age groups. All rates are age-adjusted to the 2000 US standard population. Abbreviations: AIAN, American Indian/Alaskan Native; API, Asian/Pacific Islander F I G U R E 2 Trends in annual meningioma incidence rates by tumor characteristics (2004)(2005)(2006)(2007)(2008)(2009)(2010)(2011)(2012)(2013)(2014)(2015). A, shows meningioma incidence rates by histologic type. B, shows meningioma incidence rates by WHO grade. C, shows meningioma incidence rates by tumor size. All rates are ageadjusted to the 2000 US standard population. Hemangioblastic meningioma and Meningeal sarcomatosis were not shown due to <16 cases in the time interval plausible explanation for the stable overall incidence between 2009 and 2015.
In addition, SEER registries collected meningioma data starting from 2004 and there may be bias at an early stage. The observed increase in incidence may be partly attributed to improved accuracy and documentation of meningioma in SEER registries partly. Another important change was a minor adjustment regarding meningioma in the WHO central nervous system classification, which upgraded WHO I meningioma with microscopic brain invasion to WHO II in 2007, and downgraded WHO III meningioma with brain invasion but without anaplasia to WHO I or II in 2000. 5 Although this change would not affect the trend in overall incidence, our results also reflected the fact that the incidence  (2000), as well as improvement in diagnostic accuracy of meningioma, resulting in some malignant meningiomas diagnosed as nonmalignant.
We also found significant predominance in the incidence of meningiomas in female and black patients, consistent with previous reports. Incidence of meningioma in females was more than twice that in males. A number of studies [16][17][18][19] have suggested sex hormones and genetic differences between males and females were responsible for the differences in meningioma incidence. Another study also reported that meningiomas are related to breast cancer, uterine fibroids, and endometrioses, diseases that are associated with female hormones. 20 Interestingly, the incidence trends and the magnitude of APCs of males and females were very similar during 2004-2015 according to our findings. We also identified trends in meningioma incidence by race. Black patients had the highest incidence, and the AIAN population had the lowest incidence, consistent with the report by Achey et al 4 The reasons for these differences were not only genetic or environmental factors, but also inequalities in health-care delivery. 21 Perhaps the concern for the black population has increased in recent years, which may account for the significant decrease in the incidence of meningioma in blacks.
Moreover, some extrinsic risk factors play an important role in the occurrence of meningioma. The only identified risk factor linking to an increase in meningioma is ionizing radiation. Research on ionizing radiation was mainly focused on the tinea capitis radiotherapy studies, [22][23][24] atomic bomb survivor studies, 25,26 and medical or occupational exposure. [27][28][29] Both high doses and low doses of ionizing radiation increase the incidence of meningioma and the latency periods shorten with increasing doses. A large amount of research has been devoted to studying the relationship between cell phones and brain tumor risk. 30,31 No studies have found that the use of cell phone is associated with a higher risk of meningioma. Other risk factors including the environment, genes and lifestyle have been researched with inconclusive results. High-quality research is thus essential in the future to integrate these risk factors in order to obtain conclusions.
We must acknowledge several limitations in our study. As this is a retrospective analysis, we can only speculate on the trends in meningioma incidence and the underlying factors. All our data comes from the medical records of the SEER registries, which have documented meningioma data since 2004. Continuous improvement in case records and reports may increase the observed incidence and bias in registries records, which could affect our results. The SEER database only covers 27.8% of the US population, thus not representing the whole US population. Furthermore, only patients diagnosed with meningioma for the first time were enrolled in this study, ignoring recurrent meningioma which could be more malignant. This may lead to an increase in the incidence of WHO I meningioma and a decrease in the incidence of WHO II and III meningioma. In addition, our analysis was limited by the available variables for each patient. Many factors not documented by SEER registries may substantially contribute to the incidence of meningiomas, such as environmental exposures, lifestyle, and meningioma detection methods. Finally, a large proportion of meningiomas is of unknown size, which may lead to an overestimation of the incidence of other size classifications. It is therefore necessary to continue to track trends in meningioma incidence to confirm whether the trends that we observed are sustainable.

| CONCLUSION
In this study, we provide an updated analysis of incidence rates and temporal trends for meningioma by demographic and tumor characteristics in the US population during 2004-2015. We found the overall incidence increased by 4.6% annually between 2004 and 2009, and remained stable during 2009-2015. Incidence and trends of meningioma varied significantly by demographic and tumor characteristics, and many subgroups had similar trends to overall incidence. These findings could be related to changes in population characteristics, the widespread use of diagnostic techniques, changes in tumor classification and risk factors, etc, in the US population. Further studies are needed to monitor the incidence of meningioma to determine if the observed incidence trends are persistent.