In this population based study, we developed a volumetric brain tumor atlas of all intracranial meningiomas from a total of 602 patients referred from a well-defined geographical catchment region. Meningiomas were often located in the frontal convexity region and parasagittally, along the falx anteriorly and on the skull base region of the frontal and middle cranial fossa, with gradual decline in frequencies posteriorly towards the posterior fossa and foramen magnum region. Thus, although clinically useful, the common neuroanatomical classifications of meningiomas into specific anatomical locations may be seen as arbitrary and does not directly reflect predilection sites.
The meninx, which is the primordium for the meninges, the skull and the scalp, are formed by differentiation of a mesoderm- and neural crest-derived layer that stem from mesenchymal cells completely encasing the brain and spinal cord during embryological development . Meningiomas are believed to arise from the arachnoid cap cells of the arachnoid granulations, which protrude into the venous sinuses and form ‘portals’ essential for absorption of cerebrospinal fluid (CSF). Although arachnoid granulations are more numerous along the intracranial venous sinuses and in particular along the parasagittal plane adjacent to the superior sagittal sinus , our atlas of meningiomas also demonstrate that the distribution of meningiomas does not follow the distribution of major veins or venous sinuses. In a study by Hirayama et al. of patients with meningiomas, they speculated that anatomical locations with high proportions of arachnoid cells are more frequently prone to harbor meningeal neoplasms . However, their study concentrated on lesions on one side of the brain only in selected patients, whereas our population-based study included all patients with meningiomas who underwent surgery and/or were treated conservatively. Imaging studies of arachnoid granulations located along the main intracranial venous sinuses have been reported [19,10], but distribution maps of arachnoid granulations within the meningeal layers covering the entire brain are scarce. Nevertheless, there is no known anterior- to posterior gradient of arachnoid granulations in the brain , suggesting that there is not a perfect association between the density of arachnoid granulations and meningiomas. Also, CSF efflux along the parasagittal dura is greater in the middle to posterior segments of the superior sagittal sinus , which could indicate more numerous arachnoid granulations in these regions, rather than anteriorly. However, more meningiomas are located anteriorly in the brain according to our study. Interestingly, even though the falx cerebri is shorter/narrower in its anterior portion while the posterior portion is broader and attaches to the tentorium cerebelli , meningiomas were most abundant along the convexity and anterior cranial vault compared with the posterior region. Molecular stem-cell marker positive cells associated with meningiomas have been reported in which some histological tumor types are more frequently located in non-skull base regions [18,1] suggesting that potential stem cells may be widely spread among different regions of the brain. In the early embryonic and early postnatal development, a prostaglandin D2 synthase (PGDS) positive meningeal precursor has been indicated to play a role in meningioma formation. This also accounts for the different meningioma subtypes when the biallelic NF2 gene is inactivated . Also, PDGS positive meningeal cells has been identified as a common precursor to both the dural border cells and arachnoid border cells . Vascular supply might also play a role in predilection of tumor and tumor growth as meningiomas. The vascular construction forms a more complex network in the cranial base than over the convexity. The dural territories often have overlapping vascularization from several sources such as in the parasellar dura, tentorium and falx . Furthermore, dural anatomy such as the single-layered dura of the medial wall of the cavernous sinus  may also play a role in growth of meningiomas in different intracranial regions.
The locations of meningiomas in the present study were comparatively similar overall to published series on both conservatively managed and/or surgically treated meningiomas [31,25,33,15,28], depending on how tumor locations are categorized. Also, meningiomas were more often seen in women (male to female ratio 1:2.5), a finding consistent with a population based cancer registry report , and most meningiomas were convexity tumors representing approximately one – quarter of all tumors, in line with previous reports [15,33,16,28,31]. We did not find any significant associations between age and tumor locations, in contrast to a study by Sun et al. where frontal and occipital structures were more frequently associated with older patients, males and high grade meningiomas . However, their study was restricted to surgical cases only. Some reports have dichotomized meningioma locations to skull base and non-skull base regions where the ratio of females were significantly higher in the skull base region [25,26,40,41]. In comparison, we found that men are more likely to have meningiomas located supratentorially compared to females. Thus, hormonal factors may play a role in tumor distribution. Females younger than 50 years old were also more likely to have multiple meningiomas than men in our study, but not compared to females ≥ 50 years old. Meningiomas can be associated with progesterone and estrogen receptor activity , and hormone replacement therapy (HRT) has been associated with higher risk of meningiomas in females 26 – 55 years of age compared to controls . Still, no specific predilection sites were detected for meningiomas in females less than 50 years old.
In contrast to some studies where high grade meningiomas more frequently were found over the convexity [11,23,33,42] and non-skull base regions [18,39], we found no such association between WHO grade and tumor locations. In our population based selection the overall median tumor volume was 6.1 cm3 (mean 16.8 cm3), which is somewhat larger than in reports of incidentally discovered and surgically treated meningiomas [12,27,8,23]. Meningiomas were significantly larger in the supratentorial compartment and in males in our study, but we did not find any significant association between tumor volume and tumor locations.
To the best of our knowledge, our study is the first to report an intracranial atlas of meningioma distribution from a well-defined geographical region, thus reducing risk of referral bias. Whereas some similar reports have studied meningioma distributions in selected patients and/or tumor locations, our study is population-based including both patients who did and did not undergo surgical intervention for meningioma. All tumor segmentations was performed manually and assessed in a 3D map based approach, thereby minimizing the risk of classification bias.
Possible limitations of the study include a potential discrepancy between histologically and radiologically diagnosis of meningiomas where the former is established as the gold standard. Even though it is commonplace and routine practice within the Norwegian Health Care system to refer most patients with meningiomas to the neurosurgical treatment center for treatment or recommendations concerning follow-up, there is still a possibility of under-representation of patients with tumors of small sizes and/or the oldest age group harboring meningiomas who might have not have been referred, but rather managed locally (depending on their co-morbidities/circumstances).