The relation between surgical cleavage and preoperative neuroradiological findings in intracranial meningiomas

European Journal of Radiology 80 (2011) e109–e115

Contents lists available at ScienceDirect

European Journal of Radiology journal homepage: www.elsevier.com/locate/ejrad

The relation between surgical cleavage and preoperative neuroradiological findings in intracranial meningiomas Erhan Celikoglu, Hikmet Turan Suslu ∗ , Julide Hazneci, Mustafa Bozbuga Neurosurgery Clinic, Dr. Lutfi Kirdar Kartal Research and Training Hospital, Petrol Is Mahallesi, Raman Sokak, Kartal, Istanbul, Turkey

a r t i c l e

i n f o

Article history: Received 5 April 2010 Accepted 9 June 2010 Keywords: Meningioma Surgical resection Cleavage Neuroradiology Recurrence

a b s t r a c t Background: Meningiomas are generally benign masses, and in many cases they do not invade the brain. Therefore their potential to provide cures is high. The most important cause of the development of recurrence in the post-operative period is subtotal resection. Any information that will allow us to perform total mass resection will be beneficial in terms of long-term good clinical procedure. Our aim in this study is to obtain the radiological data from which we can obtain accurate information in terms of the surgical cleavage between the tumor and parenchyma during the surgical planning of the meningiomas. Methods: We evaluated 85 cases with intracranial meningioma that were treated by the same microsurgical technique. All posterior fossa and skull base meningiomas were not included in the study. Results: Tumor size was smaller than 3 cm in 19 cases, between 3 and 6 cm in 46 cases, and bigger than 6 cm in 20 cases. The cleavage line between the tumor capsule and the cortex underneath was extrapial in 32 cases, subpial in 29 cases, and mixed in 24 cases. Dominant arterial supply was dural in 46 cases. Thirty-three cases were predominantly mixed and 6 cases were predominantly corticopial. At magnetic resonance images, 16 of 28 cases which showed clear tumor-cortex interface, had an extrapial cleavage line. Conclusions: When surgical treatment of intracranial meningiomas are considered, it is necessary to examine if there is a surgically safe border between the cortex underneath in the preoperative images. It can be concluded that it is appropriate to operate small meningiomas which are on the sensitive regions of the brain when they are in their earlier stages and still have an extrapial cleavage. © 2010 Elsevier Ireland Ltd. All rights reserved.

1. Introduction Meningiomas are benign encapsulated intracranial tumors. However, the involvement of critical vascular and neural structures linked to the meningioma, the invasive surgical cleavage may restrict the total microscopic inference targeted. The subtotal resection caused due to this may lead to an increased risk of morbidity in lesions in the sensitive regions of the brain, recurrence in long-term follow-up. The amount of surgical resection based on its system is directly linked with the risk of recurrence. Recurrence is 9% in patients with Grade I excision, 19% in patients with Grade II excision, 29% in patients with Grade III excision, and 44% in patients with Grade IV excision [1]. Adegbite et al. have stated that the rate of recurrence increases more in tumors that are harder to reach [2]. The lifespan in total resection without recurrence is found to be approximately 80% in all degree of resection [3]. The lifespan

Abbreviations: AgNOR, argyrophilic nucleolar organizer regions; CT, computed tomography; KPS, Karnofsky Performance Scale; MR, magnetic resonance. ∗ Corresponding author. Fax: +90 0216 3068059. E-mail address: [email protected] (H.T. Suslu). 0720-048X/$ – see front matter © 2010 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.ejrad.2010.06.016

in partial resection without recurrence is found to be 63%, 45%, and 9%, in the same order. In the study composed of 101 cases, Nakasu et al. have stated that after Simpson’s Grade I excision the recurrence was most frequent at the dural resection limit, after Simpson’s Grade II or III the recurrence happened locally, and that recurrence was seen more in mushroom shaped lobule meningiomas compared to the round ones [4]. Jaaskelainen et al. have reported the total recurrence rate for 20 years as 19%, and the coagulation of dural joints, bone invasion, and the soft structure of the tumor tissue have been found to be high risk factors for recurrence [5]. The 20-year recurrence rate for patients having none of these risk factors is reported to be 11%. In patients having one or two risk factors, the rate increases up to anything between 15–24% and 34–56% [6]. In the light of all this information, it can be said that the most important factor in the recurrence of cranial meningioma is post-operative tumor residual. There is a clear demarcation line on the outside of some extrinsic tumors that allow them to be removed easily. This may be seen frequently during preoperative scanning. Despite not having this appearance, some can still be easily dissected from the surrounding tissues. However, in some tumors showing distinct tumor-cortex interface during the preoperative scanning, resec-

e110

E. Celikoglu et al. / European Journal of Radiology 80 (2011) e109–e115

Table 1 The 85 cases making up to study group were separated into three groups according to the size of the tumor. In a majority of the cases (54%) the tumor is of medium size. Tumor size

Case

%

Small (<3 cm) Medium (3–6 cm) Large (>6 cm) Total

19 46 20 85

22 54 24 100

tion may be extremely difficult due to tighter cohesion. Despite the most careful dissection efforts, the resection of these tumors may result with the resection or damage of the normal parenchyma [7,8]. Whether or not the separate plans between the pia arachnoid and dura disappear with adhesive reactions cannot revealed even with the most developed screening methods. Therefore, computed tomography (CT) and magnetic resonance (MR) imaging inability to provide the true information in terms of the nature of the tumor-cortex interfaces which will be helpful to the neurosurgeon in surgical terms. When the surgical treatment of intracranial meningiomas is in question, the careful inspection of the presence of a surgically safe limit between the cortex and tumor during preoperative screening may be significant in surgical planning. Our familiarity with the screening characteristics of an adhesive tumor from our past experiences does not increase our assumption power enough in this regard. These screening characteristics must be measurable and classifiable. The purpose of this study is to show the connection between the intraoperative observed surgical cleavage type and the findings reached after making the pretumor changes in the preoperative MR screenings of some cases with meningioma and the angiographic findings measurable and classifiable. 2. Methods and materials 2.1. Characteristics of the cases This study was carried out in accordance with the files of 85 intracranial meningioma cases treated at our clinic with the same microsurgical technique and intraoperative observation records. The average post-operative follow-up period of the cases was 22 months (maximum 48 months, minimum 6 months). At the end of the 6th month following surgery, all patients were assessed in terms of “general neurological functional state”, and the Karnofsky Performance Scale (KPS) values corresponding to these were recorded. All posterior fossas and some skull base meningioma of the brain-tumor interface that were not wide enough for assessment were excluded from the study. 36 of the patients were male, 49 were female. The average age of patients is 52.4 ± 13.7 years (varies between 23 and 79 years). From their files, operation notes, and video recordings, the preoperative neurological examinations, neuroradiological examinations (CT, MR, and cerebral angiography), neurologic functional state assessments (KPS), surgical cleavage plan, and degree of resection (Simpson’s Score) of all patients were assessed. The average diameter of the tumors in the entire series was 47.1 ± 15.2 (minimum 15 mm, maximum 90 mm). The widest diameter that can be measured in the T1 predominant MR screenings was taken for the length of the tumor. Accordingly, tumors are separated into 3 groups (Table 1): those which the diameter is smaller than 3 cm (small), between 3 and 6 cm (medium), and larger than 6 cm (large). Accordingly, 19 cases were detected to be smaller than 3 cm, 46 to be between 3 and 6 cm, and 20 larger than 6 cm. After the histopathological examination, cases were separated into

subtypes in accordance with the classification of the World Health Organization (WHO). Accordingly, in the entire series there were 38 transitional, 15 meningotheliomatosis, 13 fibroblastic, 7 anaplastic, 5 atypical, 3 psamomatous, 2 angiomatosis, and 2 microcystic meningiomas. Some characteristics were studied in the preoperative screening examinations to understand whether there is an extrapial surgical cleavage plan. These radiologically examined characteristics are as follows: (1) size of the tumor, (2) width of the peritumoral edema in the CT and MR screenings, (3) the distance between the tumor capsule and the adjacent cortex surface in the MR screening, (4) cause of the supply of the tumor in the angiographies (from which of the dural arteries on the side of the tumor adjacent to the tumor or the corticopial arteries on the brain cortex surface the supply was caused). The peritumoral edema was assigned in the T2 predominant sequences of the MR screenings. The distance between the farthest point of the edema and the tumor capsule was measured on the axial section which the cross-section thickness of the hyperintense area surrounding the tumor. Accordingly, the cases were classified as follows: (1) no edema (no peritumoral hyperintensity in the T2 predominant MR screenings); (2) focal edema (3 cm or less peritumoral hyperintensity in the T2 predominant MR screenings); (3) lobar or hemispheric edema (peritumoral hyperintensity exceeding 3 cm in the T2 predominant MR screenings). The tumor-cortex interface was examined in the T2 predominant MR screenings and classified as follows: (1) clear distance (there is a clearly distinct gap (>1 mm) in more than 50% of the surface between the tumor and the cortex surrounding it); (2) well demarcated (there is no distinct gap between the tumor and cortex, however there is a limit that can be defined as good in more than 50% of the total surface); (3) poorly demarcated (there is no limit that can be defined as good in more than 50% of the tumor and cortex surface). Arterial supply was studied in the angiographies composed of the external carotid artery (ECA), internal carotid artery (ICA), and vertebral artery (VA) injections performed selectively by means of the transfemoral method. The contribution of the pial-cortical and dual-meningial arteries in the arterial supply of the tumor was examined here. Accordingly, they are classified as (1) pial supply (75% or more of the supply in the tumor is from the pial-cortical); (2) mixed type (the pial-cortical arteries contribute to the total arterial supply of the tumor at rates varying between 25% and 75%); (3) dural supply (more than 75% is from the dural–meningial arteries). All cases were operated on using standard microsurgical dissection methods. A surgical microscope (Zeiss, OPMI® Neuro/NC 4), bipolar coagulation, and other microsurgical instruments were used for this. In general, during the surgery, the tumor was first released from its dural connections via bipolar coagulation, and reduced “in pieces” from inside with bipolar coagulation, microscissors, tumor forceps, and microaspirators. Finally, the tumor was dissected from the cortex below it. After the tumor was removed, the appearance of the cortex below was especially noted. When defining the dissection plan, cases in which at least two thirds of the surface between the tumor capsule and brain cortex is extrapial are classified as extrapial (Type I); cases with brain-tumor interface which the extrapial cleavage is more than one third, less than two thirds are classified as mixed (Type II); and cases which the tumor capsule is observed to exceed the pia mater underneath in more than two thirds of the total tumor-cortex interface are classified as subpial (Type III). During the statistical calculations, the “VassarStats: Website for Statistical Computation” website was used for the “multibox chi-square relationship test”, and the “SPPSS 10.0 for Windows” software was used for the t-tests.

E. Celikoglu et al. / European Journal of Radiology 80 (2011) e109–e115

e111

Table 2 Preoperative and post-operative KPS values were compared. In the post-operative period, an average 4.97-point improvement was detected in the KPS value and this is statistically significant. Paried differences

Preoperative KPS − Post-operative KPS

t

Mean

Std. deviation

Std. error of mean

−4.9412

9.7130

1.0535

3. Results The rate of the cases in the entire series on which gross total tumor resection (Simpson Grade I–II) was applied is 96%. Subtotal (Simpson Grade III) was applied on two cases, and partial resection (Simpson Grade IV) was applied on 1 case. While the preoperative average KPS value in the entire series is 88.2 ± 10.8, the post-operative average increased up to 93.17 ± 9.78 in the 6th month. The significance of the 4.97-point improvement of patients amongst the pre- and post-operation KPS values was studied with the “t-test on matched samples”, and the change was considered statistically significant (p < 0.05) (Table 2). It was detected that the cleavage plan between the tumor capsule and the cortex under it is extrapial in 32 cases, subpial in 29 cases, and mixed in 24 cases. When the relationship of the surgical cleavage with the post-operative neurological functional improvement is studied, the post-operative KPS average of 32 cases with extrapial cleavage was 95.9. On the other hand, the post-operative KPS average of cases with non-extrapial cleavage (mixed and subpial) was 91.5 (Table 3). The significance between the averages was studied with the “t-test in independent groups”. In the test that was carried out, the difference between the post-operative functional improvements of the group with extrapial cleavage and the group with non-extrapial cleavage is considered statistically significant (Table 4). The tumor size and surgical cleavage relationship was assessed in all cases. 84% of the tumors that are smaller than 3 cm had extrapial cleavage, and 16% had mixed type cleavage. On the other hand, there were no small tumors exceeding the pial limit. Cases with medium sized (3–6 cm) tumors showed a more homogeneous distribution: 16 cases (35%) were extrapial, 19 cases (41%) were mixed, and 15 (34%) were subpial. While the cleavage plan was subpial in 18 (90%) of the 20 cases with large (>6 cm.) tumors, and mixed type in 2 (10%) cases, there were no large tumors with extrapial cleavage in the entire series. The relationship between the size of the tumor and the surgical cleavage was also statistically significant (p < 0.0001) (Table 5). The peritumoral edema and surgical cleavage relationship was assessed. There was no peritumoral edema in 30 cases (35%). Focal edema was detected in 37 cases (44%), lobar-hemispheric edema was detected in 18 cases (21%). While there was no peritumoral edema in a majority (22 cases, 73%) of tumors with extrapial cleavage, almost all of the cases with lobar-hemispheric edema (14 cases, 78%) had subpial cleavage (Table 6). While there was a statisti-

df

Sig. (2-tailed)

% 95 Confidence interval of the difference Lower bound

Upper bound

−7.0362

−2.8461

−4.690 84

0.000

cally significant relationship between the width of the peritumoral edema and the surgical cleavage plan of the tumor (p < 0.0001), there was no significant relationship between the edema and the size of the tumor (p: 0.0055). The predominant arterial supply detected in the cerebral angiography was predominantly dural in 46 cases (54%), mixed in 33 cases (39%), corticopial in 6 cases (7%) (Table 7). When the relationship of the surgical cleavage type and arterial supply type is reviewed, the cleavage plan was extrapial in 27 of the 46 cases (59%) with dural type arterial supply. Subpial cleavage was observed in only 3 (7%) of the cases in which arterial supply is of this type. There were only 6 cases in which corticopial arterial supply is predominant, and subpial cleavage was detected in 5 of these 6 cases. While there was dural predominant arterial supply present in 84% of the tumors with extrapial cleavage, in cases which the arterial supply is not dural, the probability of coming across a fully extrapial surgical cleavage was quite low (3%). There was a statistically significant relationship between the arterial supply type of the tumor group and the surgical cleavage plan (p < 0.0001) (Table 8). When the relationship between the size of the tumor and the arterial supply is studied, it was observed that all of the small tumors (all 19 cases) had dural arterial supply, and with the increase of the tumor size, the corticopial vascularisation gained significance. A statistically significant relationship between the tumor vascularisation type seen at angiography and the extent of the edema was found. 16 of the 28 cases (57%) showing evident tumor-cortex interface in the T2 predominant MR screenings were observed to have an extrapial cleavage plan. Only 2 of the cases showing evident tumor-cortex interface in the preoperative MR screenings in this group were observed to have preoperative subpial cleavage plan. While the group showing benign limit in terms of the cleavage plan showed partial homogenous distribution, 18 of the 34 cases (53%) showing malign limit in the MR examinations had subpial cleavage, 6 (18%) had mixed cleavage, and 10 (29%) had extrapial cleavage. No statistically significant relationship was detected between the radiological tumor-cortex interface and the surgical cleavage (p > 0.1). Due to almost all small sized tumors being extrapial and large sized ones being subpial, in 46 cases forming the “medium size tumor” group, the relationship between the radiological tumorcortex interface and the surgical cleavage plan was studied to deem the relationship between the radiological findings and the tumor cleavage independent from the length of the tumor. No statisti-

Table 3 The distribution of cases into surgical cleavage groups in the neurological functional state assessment (KPS) in the 6th month after surgery. There is evident improvement in the post-operative period in the KPS values in cases with extrapial cleavage. Surgical cleavage

Extrapial Non-extrapial (Mixt + Subpial)

Case

32 53

Post-operative KPS Mean

Std. deviation

Std. error of mean

95.9375 91.5094

7.5602 10.6331

1.4606 1.3365

−0.1530 −0.4887

Table 5 The distribution of cases into contingency groups based on size and surgical cleavage. A homogeneous distribution was observed in the cleavage plan types in medium sized tumors. The extrapial cleavage plan was dominant in small tumors, while the subpial cleavage plan was dominant in large tumors. No subpial cleavage was detected in small tumors, and no extrapial cleavage was detected in large tumors. In the light of these findings, a direct relationship was found between the tumor size and the surgical cleavage type (chi-square: 50.95 df: 4 and p < 0.0001).

−8.7032 −8.3674

Surgical cleavage

2.1494 1.9798

Difference of standard error

% 95 Confidence interval for mean Lower bound

Upper bound

E. Celikoglu et al. / European Journal of Radiology 80 (2011) e109–e115

−4.4281 −4.4281 0.043 0.028

Mean difference Sig. (2-tailed)

Tumor size

Extrapial Mixt Subpial Total

83 80.669 −2.060 −2.237 0.202 Post-operative KPS

Variances assumed to be equal Variances assumed to be unequal

1.654

df t Sig. F

Small

Medium

Large

16 3 0 19

16 19 11 46

0 2 18 20

Large Medium Small Total

Edema

32 24 28 85

Total

Absent

Focal

Lobar-hemispheric

3 18 9 30

7 21 9 37

10 7 1 18

20 46 19 85

Table 7 Contingency table showing the relationship between the arterial supply type and the surgical cleavage plan (chi-square: 37.36 df: 4 and p < 0.0001). While dural supply is dominant in tumors with extrapial cleavage, mixed type supply is predominantly seen in tumors with subpial cleavage. Surgical cleavage

t-test for equality of means

Total

Table 6 There was no edema in medium sized tumors or while there focal edema was present, lobar-hemispheric edema was dominant in the large tumors. There was no statistically significant relationship between the size of the edema and tumor (chi-square: 14.65 df: 4 and p: 0.0055). Tumor size

Levene test for equality of variances

Table 4 Results of the “t-test in independent groups” amongst the post-operative averages. The variances are considered equal because the P-value 0.202 > ␣ from the Levene test is the level of significance (0.1 or 0.5 or 0.01). In the statistical study, a significant difference was found in the KPS values in the post-operative period between cases with extrapial and non-extrapial cleavage.

e112

Arterial supply

Extrapial Mixt Subpial Total

Total

Dural

Mixt

Corticopial

27 16 3 46

4 8 21 33

1 0 5 6

32 24 29 85

cally significant connection could be established here for the group (chi-square test p > 0.1). 4. Discussion Yas¸argil separated the meningiomas into three groups according to the difficulty of dissection [8]. These are: (1) non-adherent (5%) (tumors that can be easily separated from surrounding tissues), (2) adherent (85%) (dissection of tumors are hard but possible), and (3) adhesive (10%) (tumors that are impossible to dissect from neural structures without any harm). Yas¸argil has emphasized that good preoperative definition and differentiation of these three groups is important in terms of surgical resection [8]. Alvernia and Sindou Table 8 When the relationship between arterial supply and edema is studied, it was observed that peritumoral edema in meningiomas with supply from the dural arterias was more limited. On the other hand, the surrounding edemas of cases showing pial and mixed type arterial coloration in the angiography are wider (chi-square: 20.93 df: 2 and p < 0.0001). Arterial supply

Dural Non-dural* Total *

Mixed + corticopial.

Edema

Total

Absent

Focal

Lobar-hemispheric

25 5 30

18 19 37

3 15 18

46 39 85

E. Celikoglu et al. / European Journal of Radiology 80 (2011) e109–e115

have studied the relationship between surgical cleavage and neurological functional improvement in meningioma cases on sensitive cortex regions [9]. In this study, at the end of the 1-year postoperative a significant amount of difference was found between the cases with subpial cleavage and extrapial cleavage in functional neurological performance. In the same study, a statistically significant relationship was detected between the tumor length and the surgical cleavage plan. It is emphasized that the presence of a strong arachnoid barrier between the meningioma and the neural tissue makes the dissection of the tumor easy, and prevents recurrence and post-operative neurological functional performance connected to this [8,10,11]. Sekhar et al. who studied 75 cases with cleaval meningioma have stated that in the onset stage the tumor is restricted in the subdural compartment, and that it separates from the pia mater with the two layers of the arachnoid membrane [10]. At this stage there is a subarachnoidal plan that allows the tumor to be dissected easily from the brain stem. In stage two, this subarachnoidal plan disappears depending on the pressure or invasion of the tumor on the arachnoid membrane. In stage three, the brain’s pia mater is invaded. This causes edema, demyelinization, or gliosis. Similar changes based on tumor invasion are also defined in the supratentorial meningiomas [4], and all of these parenchymal changes are considered edema [12]. Yas¸argil has stated that there are intrasubdural lesions in the first stages of the meningiomas, and that during this stage they are not stuck to the arachnoid, and that as the tumor grows the arachnoid, pia, and other restricting layers are invaded by the tumor [8]. The probability of coming across an extrapial surgical cleavage decreases as the length of the meningioma grows. This is emphasized in other similar studies as well [9,10,13,14]. It is observed from the cerebral angiography examinations that the supply in meningiomas mostly originate from the meningial arteries. Large tumors generally have double arterial supply; while the supply in the core of the tumor is from the arteries of the meninx, the blood build up in the surrounding is from the pial stems of cerebral arteries. It is reported that as the meningioma grows, the weight of pial contribution in the supply increases [9,10,14]. In this case, as the tumor grows, the probability of the subpial coming across a surgical cleavage plan decreases depending on both pressure and the increase of pial vascularisation. The significance of the corticopial arteries on the blood build up of intra-axial brain tumors is clear. On the other hand, its contribution on the supply of meningiomas is less. Sekhar et al. have stated that when the size of the tumor is ignored, the supply of the tumor from the basilar artery is the most important preoperative determinative factor in the post-operative permanent neurological performance [10]. They have shown that such cases having corticopial supply have 4.4 times less risk in terms of permanent neurological dysfunction. On the other hand, in another study it was asserted that as a result of the selective angiographies of 62 cases with medium sized meningioma, whether dominant or mixed type, cases with corticopial supply cannot have an extrapial surgical cleavage [9]. Whereas in our study there was no statistically significant relationship between the arterialisation of the tumor and the surgical cleavage (chi-square 37.36 and p < 0.0001); 4 of the 32 cases having extrapial cleavage and 1 case having corticopial type supply is partially contradictory to this study in the literature. Meningiomas are unique lesions in terms of peritumoral edema. Despite their benign structures, slow growth rate, extracerebral settlement, a leptomeningial barrier that is impermeable and separates them from the brain parenchyma, all meningiomas form peritumoral edema in at least half of them [15,16]. It is still unknown how this occurred. The tumor-brain interface has been shown to be the most important region in the formation of edema in meningiomas. Yas¸argil stated that there is mode edema in menin-

e113

gioma showing adhesion to the arachnoid and pia, but on the other hand that the dissection of lesions having less peritumoral reaction from surrounding tissues is easier [8]. The peritumoral low-density in the CT of cases with meningioma was studied by many researchers to figure out ‘what’ it is. Does the edema fluid leak from the tumor itself or from the brain tissue which the blood–brain barrier around it is deteriorated? In their series consisting of 38 meningioma cases with wide peritumoral edema, Go et al. have looked at the contrast involvement on the brain tissue around the tumor and have studied whether the blood–brain barrier is intact [17]. The authors have been able to show peritumoral contrast involvement in 1 case only, and have asserted that the remaining majority have an intact blood–brain barrier. Accordingly, they support the opinion that the blood–brain barrier surrounding the meningioma is generally intact, and that this cannot be the cause of the edema of the peritumoral brain tissue. It is also believed that the origin of the edema fluid is the tumor itself [18]. This time it needs to be explained if the formation of the edema fluid depends on “the diffusion of water, or the active secretion from the tumor to the brain”. Various hypotheses have been suggested. A possibility is that the meningiomas cause edema in the form similar to those in intrinsic brain tumors. The mechanism here can be summarised as increased capillary permeability, a pressure gradient from the artery to the extracellular compartment, and the retention of fluid in the extracellular region. Despite the meningioma causing changes in the capillary structures similar to those in the glial tumors, this will not cause the formation of edema on its own unless the tumor is physically penetrated in the brain substance [19]. This is not common, however in its existence, there is a linear relationship between the amount of the peritumoral edema. Cerebral edema settles mostly in white matter. What separates the white matter from the meningioma is the arachnoid membrane, subarachnoidal gap, pia mater, and the cerebral cortex. The arachnoid membrane is fluid impermeable. Pia mater easily permeates water and electrolytes, however it is not equally permeable against macromolecular substances which the proteins of vasogenic edema fluid is included. Similarly, the cerebral cortex is also quite resistant to the spread of the vasogenic edema fluid due to its tight cell tissue structure [1]. Even if the edema fluid is actively secreted by the tumor, the barrier that develops from the abovementioned anatomic structures is believed to prevent the edema fluid from spreading into the white matter [12]. While the fluid is extravased from the permeable capillaries, the leptomeninxes are impermeable. In this case, this is not how it will happen while the edema fluid is expected to accumulate extracerebrally. If the “permeating” tumor arteries are the true source of the edema accompanying the meningioma, it should be accepted that there is a functional defect in the leptomeningial barrier as well. The deficiency in accepting this as it is will not only corrupt this theory, it will also weaken the other various theories that will be discussed below. The peritumoral capillaries of the normal brain tissue surrounding a meningioma show symptoms of advanced level pinocytic vesicular and dedifferentiation [20]. Therefore, the decomposition of the cortical and leptomeningial layers in between with the mass effect of the tumor facilitates fluid permeability towards the brain tissue in accordance with the existing pressure gradient [1,19]. As the tumor size increases in the meningiomas, the performance of these barriers really do contribute to the formation of edema even if partially. In the second hypothesis it is claimed that the edema formation occurs as a response to the mass effect of the tumor. Some studies have shown that as the size and surface area of the tumor increase the edema increases [15]. The common opinion in clinical practice is that there is a relationship between the tumor length and edema that cannot be predicted in advance. While sometimes

e114

E. Celikoglu et al. / European Journal of Radiology 80 (2011) e109–e115

the edema does not accompany large meningiomas, sometimes small meningiomas may be accompanied by edema holding the entire hemisphere. In some of the studies carried out, the relationship between the size of the tumor and amount of edema was studied. Abe et al. have reached the conclusion that 40% of the 68 cases with meningioma having evident edema have peritumoral low absorption area [21]. These researchers have mentioned two separate types of edema in these cases. The first of these is a diffused white matter process that reflects the active transudation of water in white matter, and it is seen in 43% of the tumors in the series with edema. The edema seen in the remaining 57% is the localized type. According to the authors who emphasize the clinical and histopathological significance of the differentiation between these two peritumoral processes, the diffused pattern does not reflect the size of the tumor [21]. Whereas in other studies, a positive connection was found between the degree of the edema and the size of the tumor, and it was stated that they may cause edema by deluting the cortical and leptomeningial layers of wide tumors [1]. A significant connection was reported between the length of the tumor and the peritumoral edema, however it was detected that as the length of the tumor increases the edema index (the ratio of the edema volume to the tumor) decreases [12]. This may be the result of an anatomical restriction in the large lesions. The length of the tumor cannot be the only factor in the formation of edema. Many possible mechanisms have been suggested. These are the age of the patient, localization of the tumor, its vascularisation, histological subtype, cellularity, mitotic efficiency, the existence of steroid receptors, pressure of the drainer vein, and the secretory efficiency of tumor cells [6,15,18,21–28]. Whereas in our study no statistically significant relationship could be found between the size of the tumor and the formation of edema (chisquare: 14.65 df:4 and p = 0.0055). Actually, it is generally considered that more than the size of the tumor its rate of growth may be a more reliable messenger of the size of the edema [19]. The same argument is also used to explain the edema accompanying metastatic tumors and edemas that are typically seen disproportionately wide compared to the tumor. Mitotic activity, increased cellularity and necrosis are the traditional measurements of proliferative capacity. Thus, the rate of growth of the tumor can only be calculated to a certain point. Peritumoral edema is seen in more than 90% of the meningioma having these histological characteristics, and almost 70% of these are evident edemas. Only half of all meningiomas having this morphological aggressiveness have evident edema [29]. In the “flow cytometric” analyses of the meningiomas, a positive relationship is shown between the proliferative capacity of the tumor and peritumoral edema by determining the ones in the G2 and S phases of the cells in the cycle [30]. Another mechanism that is commonly referred to explain the edema surrounding the meningioma is venous occlusion. Yet as the length of the tumor increases, venous congestion will also increase, and as a result this will reflect on the size of the edema. Theoretically, the bass or invasion of the venous sinuses and cortical veins may be the cause of peritumoral edema. This is because increased cerebral venous pressure results in fluid transudation. However, this cannot be the rich exuded type vasogenic edema fluid from the common protein [19]. Despite this opinion being a mechanically appealing one, while the progressive venous occlusion settles, since we know that the collateral circulation is developing, it would be a little contradictory. Furthermore, in some clinical studies researching this matter, in cases which the venous sinuses and drainage veins are angiographically open, it has been shown that there may be evident peritumoral edema, and on the other hand there may be no edema in the existence of a total sinus occlusion [31]. The recent studies claim that the peritumoral edema in most of the cases is not linked to venous occlusion [32].

Even though venous occlusion has no role on the formation of edema in most meningiomas, the arterial supply of meningioma seems extremely important in the formation of peritumoral edema. While no peritumoral edema could be seen in meningiomas with no pial supply, it was shown that there are more widespread peritumoral edemas in tumors with pial coloration [33]. The surrounding edema of meningiomas which the supply is largely from dural arteries was mostly less compared to those with supply from pial arteries (p > 0.0001). The peritumoral edema surrounding the meningiomas may also be the result of some chemical and biological mediators. Their significances in the formation of cerebral edema are previously discussed, known mediators. A positive relationship was found between the prostaglandin level and the peritumoral edema [23]. In another study, higher VEGF levels were found in meningiomas containing a lot of peritumoral edema [18,35]. Bitzer et al. reported, all meningiomas showing high VEGF expression have pial supply [18]. On the other hand, only 50% of those not showing the VEGF expression have been able to show pial supply. This may even explain the relationship between the peritumoral edema and tumor vascularisation. This is because we now know the relationship between the vascularity of the tumor and the peritumoral edema. In some studies, histologically and angiographically determined vascularity is found to have a connection with the prevalence of the peritumoral edema [16,36]. However, there are also studies with findings that are not compatible with these [15,37]. When we examined the results of our study, a significant improvement was detected in tumors with extrapial cleavage between the pre- and post-operative KPS scores. The best improvement in the KPS score was in tumors with extrapial cleavage. There is a direct relationship between the extrapial cleavage and the size of the tumor and the peritumoral edema. As the size of the tumor decreases, the possibility of coming across an extrapial cleavage plan increases. For tumors larger than 6 cm there is no possibility of coming across an extrapial cleavage plan. As the peritumoral edema increases, the possibility of coming across an extrapial cleavage increases. However, we did not detect any significant relationship between the size of the tumor and the peritumoral edema. This is most likely because the formation mechanism of the peritumoral edema could be caused by the difference in the causes of the peritumoral change. There is no significant relationship between the tumor-cortex interface seen in the MR and the surgical cleavage. However, the size of the tumor detected with the MR is helpful in terms of determining the surgical cleavage type. The tumor supply type that is detected with the cerebral angiography gives accurate data in terms of the separation of the cleavage plan. In our studies, dural arterial supply was detected in small tumors and tumors with extrapial cleavage.

5. Conclusion When the surgical treatment of intracranial meningiomas is in question, the presence of a surgically safe limit between it and the cortex under it during preoperative screening must be studied carefully. The tumor invasion of the pia may be masked by peritumoral hyperintensity in the T2 predominant MR screening, and by peritumoral hypodensity in the CT. The data obtained from the CT and MR concerning the length of the tumor and the width of the edema are able to provide adequate power to make assumption regarding the invasion of the pial limit, and the selective cerebral angiography provide important information concerning the estimation of the state of the tumor cleavage and the state of the post-operative neurological function. While the dissection plan in most small tumors is extrapial, with large tumors it has been seen that dissection is possible at a high rate but only in the subpial plan. With medium scale tumors, the width of the peritumoral edema in the CT and

E. Celikoglu et al. / European Journal of Radiology 80 (2011) e109–e115

MR is a significant indication to estimate the dissection plan. From these findings we can conclude that it would be better to operate on the small meningiomas particularly on the sensitive brain regions at the early stages which they have an extrapial cleavage. References [1] DeMonte F, Marmor E, Al-Mefty O. Meningiomas. In: Kaye AH, Laws RE, editors. Brain Tumors. London: Harcout Publishers Limited; 2001. p. 719–50. [2] Adegbite AB, Khan MI, Paine KW, Tan LK. The recurrence of intracranial meningiomas after surgical treatment. J Neurosurg 1983;58:51–6. [3] Mirimanoff RO, Dosoretz DE, Linggood RM, Ojemann RG, Martuza RL. Meningioma: analysis of recurrence and progression following neurosurgical resection. J Neurosurg 1985;62:18–24. [4] Nakasu S, Hirano A, Llena JF, Shimura T, Handa J. Interface between the meningioma and the brain. Surg Neurol 1989;32:206–12. [5] Jaaskelainen J, Haltia M, Servo A. Atypical and anaplastic meningiomas: radiology, surgery, radiotherapy and outcome. Surg Neurol 1986;25:233–42. [6] Kawase T, Ohira T, Murakami H. Influence of peritumoral edema on rCBF and on cerebral function: analysis by xenon-enhanced CT and EEG topography. In: Inaba Y, Klatzo I, Spatz M, editors. Brain Edema. Berlin: Springer Verlag; 1985. p. 484–9. [7] Yasargil MG. Meningiomas. Microneurosurgery IVB. New York: Thieme; 1996. pp. 134–165. [8] Yasargil MG. Neuropathology. Microneurosurgery IVA. New York: Thieme; 1996. pp. 115–191. [9] Alveria JE, Sindou MP. Preoperative neuroimaging findings as a predictor of the surgical plane of cleavage: prospective study of 100 consecutive cases of intracranial meningioma. J Neurosurg 2004;100:422–30. [10] Sekhar LN, Swamy NK, Jaiswal V, Rubinstein E, Hirsch Jr WE, Wright DC. Surgical excision of meningiomas involving the clivus: preoperative and intraoperative features as predictors of postoperative functional deterioration. J Neurosurg 1994;81:860–8. [11] Toth S, Vajda J, Pasztor E, Toth Z. Separation of the tumor and brain surface by “water jet” in cases of meningiomas. J Neurooncol 1987;5:117–24. [12] Bradac GB, Ferszt R, Bender A, Schorner W. Peritumoral edema in meningiomas. A radiological and histological study. Neuroradiology 1986;28: 304–12. [13] Inamura T, Nishio S, Takeshita I, Fujiwara S, Fukui M. Peritumoral brain edema in meningiomas—influence of vascular supply on its development. Neurosurgery 1992;31:179–85. [14] Sindou MP, Alaywan M. Most intracranial meningiomas are not cleavable tumors: anatomic-surgical evidence and angiographic predictibility. Neurosurgery 1998;42:476–80. [15] Gilbert JJ, Paulseth JE, Coates RK, Malott D. Cerebral edema associated with meningiomas. Neurosurgery 1983;12:599–605. [16] Smith HP, Challa VR, Moody DM, Kelly Jr DL. Biological features of meningiomas that determine the production of cerebral edema. Neurosurgery 1981;8:428–33. [17] Go KG, Wilmink JT, Molenaar WM. Peritumoral brain edema associated with meningiomas. Neurosurgery 1988;23:175–9. [18] Bitzer M, Wockel L, Morgalla M, et al. Peritumoral brain edema in intracranial meningiomas: influence of tumor size, location and histology. Acta Neurochir (Wien) 1997;139:1136–42.

e115

[19] Thapar K, Taylor MD, Laws ER, Rutka JT. Brain edema, increased intracranial pressure, and vascular effects of human brain tumors. In: Kaye AH, Laws RE, editors. Brain Tumors. London: Harcourt Publishers Limited; 2001. p. 189–215. [20] Vaz R, Borges N, Cruz C, Azevedo I. Cerebral edema associated with meningiomas: the role of peritumoral brain tissue. J Neurooncol 1998;36:285–91. [21] Abe T, Black PM, Ojemann RG, Hedley-White ET. Cerebral edema in intracranial meningiomas: evidence for local and diffuse patterns and factors associated with its occurrence. Surg Neurol 1994;42:471–5. [22] Benzel EC, Gelder FB. Correlation between sex hormone binding and peritumoral edema in intracranial meningiomas. Neurosurgery 1988;23:169–74. [23] Constantini S, Tamir J, Gomori MJ, Shohami E. Tumor prostaglandin levels correlate with edema around supratentorial meningiomas. Neurosurgery 1993;33:204–10. [24] De Vries J, Wakhloo AK. Cerebral edema associated with WHO-I, WHO-II, and WHO-III-meningiomas: correlation of clinical, computed tomographic, operative and histological findings. Acta Neurochir (Wien) 1993;125:34–40. [25] Hino A, Imahori Y, Tenjin H, et al. Metabolic and hemodynamic aspects of peritumoral low-density areas in human brain tumor. Neurosurgery 1990;26:615–21. [26] Ildan F, Tuna M, Gocer AP, et al. Correlation of the relationships of brain-tumor interfaces, magnetic resonance imaging, and angiographic findings to predict cleavage of meningiomas. J Neurosurg 1999;91:384–90. [27] Lobato RD, Alday R, Gomez PA, et al. Brain edema in patients with intracranial meningioma. Correlation between clinical, radiological, and histological factors and the presence and intensity of edema. Acta Neurochir (Wien) 1996;138:485–93. [28] Ohno K, Matsushima Y, Aoyagi M, et al. Peritumoral cerebral edema in meningiomas: the role of the tumor-brain interface. Clin Neurol Neurosurg 1992;94:291–5. [29] Schrell UM, Fahlbusch R, Adams EF, Nomikos P, Reif M. Growth of cultured human cerebral meningiomas is inhibited by dopaminergic agents. Presence of high affinity dopamine-D1 receptors. J Clin Endocrinal Metab 1990;71:1669–71. [30] Crone KR, Challa VR, Kute TE, Moody DM, Kelly Jr DL. Relationship between flow cytometric features and clinical behavior of meningiomas. Neurosurgery 1988;23:720–4. [31] Lindley JG, Challa VR, Kelly DL. Meningiomas and brain edema. In: Al-Mefty O, editor. Meningiomas. New York: Raven Press; 1991. p. 59–73. [32] Bitzer M, Topka H, Morgalla M, Friese S, Wockel L, Voigt K. Tumor-related venous obstruction and development of peritumoral brain edema in meningiomas. Neurosurgery 1998;42:730–7. [33] Bitzer M, Wockel L, Luft AR, et al. The importance of pial blood supply to the development of peritumoral brain edema in meningiomas. J Neurosurg 1997;87:368–73. [35] Yoshioka H, Hama S, Taniguchi E, Sugiyama K, Arita K, Kurisu K. Peritumoral brain edema associated with meningioma: influence of vascular endothelial growth factor expression and vascular blood supply. Cancer 1999;85: 936–44. [36] Stevens JM, Ruiz JS, Kendall BE. Observations on peritumoral edema in meningioma. Part II. Mechanisms of oedema production. Neuroradiology 1983;25:125–31. [37] Maiuri F, Gangemi M, Cirillo S, et al. Cerebral edema associated with meningiomas. Surg Neurol 1987;27:64–8.