Epidemiology of Primary CNS Tumors in Iran : A Systematic Review

BACKGROUND
Although primary malignant CNS tumors are registered in the national cancer registry (NCR) of Iran, there are no available data on the incidence of the primary malignant or benign CNS tumors and their common histopathologies in the country. This study analyzed the 10-year data of the Iranian NCR from March 21, 2000 to March 20, 2010, including a systematic review.


MATERIALS AND METHODS
The international and national scientific databases were searched using the search keywords CNS, tumor, malignancy, brain, spine, neoplasm and Iran.


RESULTS
Of the 1,086 primary results, 9 papers were selected and reviewed, along with analysis of 10-year NCR data. The results showed that primary malignant brain tumors have an overall incidence of 2.74 per 100,000 person-years. The analysis of the papers revealed a benign to malignant ratio of 1.07. The most common histopathologies are meningioma, astrocytoma, glioblastoma and ependymoma. These tumors are more common in men (M/F=1.48). Primary malignant spinal cord tumors constitute 7.1% of the primary malignant CNS tumors with incidence of 0.21/100,000.


CONCLUSIONS
This study shows that CNS tumors in Iran are in compliance with the pattern of CNS tumors in developing countries. The NCR must include benign lesions to understand the definitive epidemiology of primary CNS tumors in Iran.


Introduction
The incidence of primary CNS Tumors has been estimated as 3.9 and 3.2 per 100,000 person-year worldwide in males and females, respectively (Ferlay et al., 2010).However, the reported incidence of primary CNS tumors is higher in developed countries (Bondy et al., 2008).Nevertheless, the incidence of primary malignant CNS tumors ranges from 2.1 to 5.8 per 100,000 population in the world (Bondy et al., 2008).
The reported incidence of primary CNS tumors has increased from 11.5 in 1994 to 20.1in 2008 per 100,000 persons in the US (Surawicz et al., 1999;Bondy et al., 2008).The incidences of brain and spinal cord tumors in USA are 19.5 and 0.6 per 100,000 individuals, respectively.The importance of descriptive data on the full spectrum of primary brain and spinal cord tumors is mainly for its role of assessing risk factors, finding at-risk populations and as a rich database for studying relatively rare brain tumors (Kurland et al., 1982;Deorah et al., 2006).For the first time, a CNS tumor registry was begun in 1902 by Harvey Cushing (Wahl et al., 2009) and subsequently the registry expanded to other countries.The epidemiologic data from
A local cancer registry in Iran began in 1955 with the establishment of Cancer Institute at the University of Tehran (Etemadi et al., 2008).The first formal cancer related data from Iran were published by Habibi in 1962(Habibi, 1962;Listed, 1977).The late professor Ameli (Ameli et al., 1979;Alimi and Rahimi-Movaghar, 2012) and Ardehali (1990) published the first reports on the brain and spinal cord tumors in Iran in 1979and 1990, respectively. Since 1984, all , all pathology centers in Iran are required to (by law from parliament) report all tissues "diagnosed or suspected of cancer/malignant" to the Ministry of Health.Malignant cases are reported according to International Classification of Diseases-Oncology (ICD-O).After the foundation of the National Cancer Registry (NCR) module, several reports for incidence and prevalence of malignancies were published from single centers and/or different provinces/cities of the country (Larijani et al., 2004;Babaei et al., 2005;Mohagheghi et al., 2009;Mousavi et al., 2009).There are no accurate published nationwide CNS tumor epidemiology data for Iran, and this makes policy making for prevention, tumor screening and treatment modalities difficult.
This study attempts to report a systematic review of the epidemiology of CNS tumors in Iran, in published and recently unpublished reports of NCR data, and in all available published data from authors in English and Farsi journals, and estimates the incidence of primary benign and malignant brain and spinal cord tumors and their common pathologies.

Materials and Methods
This systematic review was designed in accordance with the PRISMA group guideline of systematic review requirements (Moher et al., 2010).Keywords were selected under the supervision of a neurosurgeon and a medical librarian who designed the appropriate search strategy for each database (Table 1).On April 17 th 2012, Medline, Embase, ISI web of knowledge and Google scholar for international papers and three Iranian databases of Scientific Information Database (SID) (http://www.sid.ir),MagIran (www.magiran.com), and IranMedex (www.iranmedex.com) for Farsi articles were searched.Keywords included: central nervous system, brain, intracranial, cerebroventricular, intraventricular, cerebrum, choroid plexus, infratentorial, posterior fossa, pontine, mesencephalic, midbrain, medullary, cerebellum, supratentorial, hypothalamus, pituitary, neoplasm, cancer, malignancy and Iran.Results were integrated into an EndNote X5 library and duplicates were then automatically removed.Two independent reviewers (VRM and SBJ) reviewed the remaining papers.We used ten-year (March 21, 2000-March 20, 2010) raw data of NCR (2000NCR ( -2009) ) and percent of coverage of registered primary malignant tumors in 10 years to estimate the incidence of primary malignant brain and spinal cord tumors and their different pathologies.
During the ten years of the registry program, the coverage of 20% in 2000 has progressed to 93% in 2009.
To estimate the incidence of common primary malignant brain and spinal cord tumors, we used the accurate data from the last verified year (from March 21, 2009to March 20, 2010).In this study, metastatic tumors were excluded.Since papers have reported the total count of astrocytoma and glioblastoma together as gliomas, data of glioma were not included in the analysis.NCR data did not include the primary benign tumors of the brain and spinal cord.Therefore, we used data from all available comprehensive single center published papers for primary brain and spinal cord tumors.The ratio of benign to malignant tumors were calculated based on these reports and was used to estimate the incidence of benign tumors in the country.The same process was used for sex and age distributions of the tumors.

Results
Primary searches found 1086 entities in addition to the database of NCR data.Of the primary results, 562 duplicates were removed, and this limited resources to 524 papers.One hundred thirty four papers were then selected based on their relevance to the subject after reading titles and abstracts.Of the initial 134 selected articles, 36 papers were removed because the report was limited to specific tumor types or specific treatment methods, 17 papers were deleted because the reports were limited to data of a specific age groups, 22 papers were reporting the same data of the NCR along with 6 case reports, 17 papers which had no epidemiologic data and 27 papers were rejected following quality control.Therefore, the 10-years data of NCR (2000NCR ( -2009) ) and nine papers (Ameli et al., 1979;Ketabchi and Ghodsi, 1989;Ardehali, 1990;Rezaee and Hadadian, 1997;Meshkini et al., 2000;Mahzoni and Mohammadizadeh, 2003;Reyhani-Kermani, 2003;Mehrazin et al., 2006;Seddighi et al., 2010) related to the CNS tumors were selected (Figure 1).These nine CNS papers consisted of seven papers for brain (Ameli et al., 1979;Ketabchi and Ghodsi, 1989;Meshkini et al., 2000;Mahzoni and Mohammadizadeh, 2003;Reyhani-Kermani, 2003;Mehrazin et al., 2006;Seddighi et al., 2010) and two papers for spinal cord tumors (Ardehali, 1990;Rezaee and Hadadian, 1997).
The NCR reports have been published regularly in      ) of all primary malignant tumors (Table 2).A total of 10,868 cases of primary malignant brain tumors were registered between 2000 and 2009.In the rechecking the details of the entered data, 66 cases were found to have incompatible site and histologic coding values in the raw data; these cases were removed from the study, hence, 10,802 cases are shown (Table 3).Histologic confirmation was achieved in 94.7% (10,227/10,802) of the cases.Astrocytoma and glioblastoma together form 60.4% of the primary malignant registered brain tumors in Iran (Table 4).Astrocytoma is the most common pathology of the registered tumors, except in 2000 where glioblastoma is registered as the most common type.Astrocytoma and glioblastoma are both more common in males (M/F=1.5 and 1.8, respectively).The incidence of primary malignant CNS tumors was 2.73 per 100,000 for primary malignant   (1978-2003) (1977-1986) (1949-1978) (1997-2006) (1996-2000) (1997-2001)   brain tumors (Table 5), which comprises incidence rates of 3.10 and 2.37 for males and females, respectively.Benign brain tumors are not registered in NCR data.A summary of published data is shown in Table 6.Benign CNS tumors range from 30.3% (Seddighi et al., 2010) to 58.2% of all CNS tumors indifferent studies (Meshkini et al., 2000).In the analyzes of the seven available studies, benign tumors composed 51.9% of all primary brain tumors.Meningioma (27.8%), pituitary adenoma (11.3%) and schwannoma (5.9%) are the most common pathologies in benign brain tumors (Table 6).The gender distribution of benign brain tumors was reported in three papers, (Meshkini et al., 2000;Reyhani-Kermani, 2003;Mehrazin et al., 2006) which were analyzed to determine the sex distribution of benign tumors in Iran.Although brain tumors are generally more common in males, meningioma was more common in females (M/F: 0.69).
The mean age and incidence of the most common pathologies of primary brain tumors are shown in Table 7. Analysis of NCR data and published evidence revealed incidence of 2.95 per 100,000 persons for primary benign brain tumors.Three papers have reported the total count of astrocytoma and glioblastoma together as gliomas: these data were not included in the analysis (Ameli et al., 1979;Mahzoni and Mohammadizadeh, 2003;Reyhani-Kermani, 2003).Meningioma as the most common pathology of benign brain tumors has incidence of 1.58 per 100,000 person-years.
Primary malignant spinal cord tumors registered in the NCR are shown in Table 7.In the rechecking process for spinal cord registered tumors, 21 cases were excluded because of coding miss-match.Primary malignant spinal cord tumors composed 7.1% of primary malignant brain tumors which has incidence of 0.21 per 100,000.Malignant neoplasm composes 34.4% of all malignant spinal cord tumors in the NCR data.Primary malignant spinal cord tumors are more common in males, with a male to female ratio of 1.26.A total of 767 cases of primary malignant spinal cord tumors were registered between 2000 and 2009 (Table 8).Published papers of spinal cord tumors in Iran are limited to two studies, (Ardehali, 1990;Rezaee and Hadadian, 1997) which reported the most common pathologies as nerve sheath tumors, meningioma and neuroepithelial tumors (Table 9).Nerve sheath tumors and meningioma together formed 56.4% of primary spinal cord tumors.Therefore, the incidence of these two benign tumors is 0.27 per 100,000.The total primary spinal cord tumors are estimated to be 0.48 per 100,000.Three common pathologies of spinal cord tumors consist of neurofibroma (30.7%), meningioma (25.7%), and malignant neoplasm (15.0%).
In a study between 1986-1993 in Spain the incidence of primary brain tumors was 8.34 and 5.40 per 100,000 population of males and females, respectively (Lopez-Abente et al., 2003).The incidence of primary brain tumors in France (Bauchet et al., 2007).was 15.8 per 100,000.The French study shows 39.6% benign, 56.3% malignant and 4.1% unknown cases (Bauchet et al., 2007).In the report of the cancer registry of Korea in 2005, the incidence of primary brain tumors was 11.69 per 100,000 population (Lee et al., 2010).Manoharan et al. (2012) reported primary brain tumors in Delhi showing an incidence rate of 3 and 2 for males and females per100,000, respectively (Manoharan et al., 2012).
The worldwide incidence of primary malignant brain tumors is reported as 3.7 for males and 2.6 for females per 100,000.These rates are higher in developed countries: 5.8 and 4.1 in 100,000 males and females, respectively.Underdeveloped countries report a lower incidence as 3 and 2.1 per 100,000 cases in males and females (Bondy et al., 2008).
It seems that the reported incidence of brain tumors is related in part to the economy of various countries.The highest rates of incidence are seen in North America, Australia (Dobes et al., 2011), and Western Europe, (Deltour et al., 2009) and the lowest incidences are reported in Asia, Central and South American regions (Ferlay et al., 2010).The GLOBOCAN 2008 study published in 2010, estimated the highest incidence of primary brain tumors in North Europe, and the lowest incidence in Eastern Africa (Ferlay et al., 2010).The incidence of brain tumors in Iran is lower than in developed countries.Primary malignant brain tumors in Iran are more common in males with a male to female ratio of 1.48; the higher proportion of tumors in males is in accordance with global distribution of tumors.In

Discussion
Based on systematic reviews, the incidence of primary brain tumors in Iran is estimated to be 5.69 per 100,000; our study all tumors except meningioma were diagnosed more commonly in males, which is in compliance with the worldwide distribution of primary brain tumors.However, the incidence of primary brain tumors in the US (Dolecek et al., 2012), Korea (Lee et al., 2010) and France (Baldi et al., 2011) is more frequent in females.This might be due to the higher rate of meningiomas (30.6% and 35%) and lower rates for gliomas (17% and 30%) in Korea and US, respectively.It might be noted that the broad category of benign tumors is registered in the series from the US and Korea, while benign tumors in Iran are estimated based on available evidence.
Spinal cord tumors in Iran composed 7.1% of all CNS tumors, while in the US spinal cord tumors form only 3% of all CNS tumors (Dolecek et al., 2012).The most common pathologies for spinal cord tumors of US are meningeal tumors, nerve sheath tumors and ependymoma.(Dolecek et al., 2012).Su et al. (1997) have reported the results of surgically treated spinal cord tumors between 1988 and 1995 in Taiwan.In their report of 117 patients, 69 were males and 48 were females.The most common pathologies were metastasis (45.3%), nerve sheath tumors (28.2%), meningioma (12%), and neuroepithetial (6%) tumors.Studies in Iran (Ardehali, 1990;Rezaee and Hadadian, 1997) have also reported the same common pathologies.However, it should be noted that one third of primary malignant spinal cord tumors are in malignant neoplasm group of unknown pathology, some of which may be due to radiologic diagnosis, non-operated or inoperable cases.Nevertheless, in the US only 2.6% of spinal cord tumors are in the unclassified neoplasm category (Dolecek et al., 2012).
In the registry process many cases had failed to be diagnosed pathologically due to early death, refusing surgery, inappropriate tumor location for biopsy or surgery, etc.These tumors are collectively accounted for as malignant neoplasm in the registry system.
The main limitation of the NCR system is that benign lesions are not registered.Although every effort was made to select the best available data in analyzing benign tumors, the single center or single author nature of these studies may affect the overall analysis.
Meanwhile, metastasis is another interesting subject, which we did not include in this study.The age distribution of each tumor, especially in the pediatric age group will be discussed in a separate paper.
To our this report contains the most recent incidence data for primary malignant CNS tumors in Iran.Iranian NCR data have an average lag time of 4 years from data gathering to the final report, similar to the CBTRUS program of the US (Dolecek et al., 2012).Contrary to the CBTRUS program in the US, the Iranian NCR does not register benign CNS tumors.Our study emphasizes the need for separating within the CNS tumor registry program both benign and malignant tumors.The incidence of CNS tumors in Iran is in accordance with the incidence of CNS tumors in developing countries.These data estimate incidence of 2.74 malignant and 2.95 benign lesions per 100,000 person-years, and a total of 5.69 per 100,000 person-years for primary brain tumors in Iran.

Figure
Figure 1.The Algorithm of Paper and Resource Selection

Table 3 . Summary of Registered Brain Tumors between 2000 and 2009 by Sex and Histology in Iran
Farsi since 2000-2009 and while in its first year, only 20% of the estimated cancer cases were reported; this rate has increased to 93% in 2009.Primary malignant CNS tumors in Iran encompass 2.3% (95%CI: 2.3-2.4

Table 7 . Incidence Rate and the Mean Age of Patients of Common Pathologies of Malignant Brain Tumors in Iran Based on NCR Data
Seyed Behzad Jazayeri et al Asian Pacific Journal of Cancer Prevention, Vol 14, 2013 3982