J Neurooncol (2015) 121:521–529 DOI 10.1007/s11060-014-1659-z
CLINICAL STUDY
A comparative study of intraventricular central neurocytomas and extraventricular neurocytomas Zhongwei Xiong • Jianjian Zhang • Zhengwei Li Jingjing Jiang • Qingdong Han • Shoujia Sun • Xiaolin Wu • Yu Wang • Ting Lei • Jincao Chen
•
Received: 22 February 2014 / Accepted: 6 November 2014 / Published online: 11 November 2014 Ó Springer Science+Business Media New York 2014
Abstract Similar histology and clinical behavior of both intraventricular central neurocytomas (CNs) and extraventricular neurocytomas (EVNs) may argue against the idea that EVNs were the distinct entity to distinguish from CNs in the 2007 World Health Organization classification. To explore respective characteristics and compare similarities and differences in CNs and EVNs, relevant clinical, radiological, operative and pathological data of 49 patients (35 CNs and 14 EVNs) in the Department of Neurosurgery at our hospital from 2005 to 2012 was reviewed and some comparisons between CNs and EVNs were conducted. The factors affecting posttreatment recurrence of CNs and EVNs were assessed by Cox regression analysis. In comparison, CNs showed a more typical clinical manifestation, and radiological and histopathological features, while EVNs demonstrated more malignant biological behavior, with higher MIB-1 index (p = 0.006), higher rate of atypia (p = 0.042), higher recurrence rate (p = 0.028), and shorter time to recurrence (p = 0.049). Subtotal resection
Z. Xiong J. Zhang Q. Han S. Sun X. Wu Y. Wang T. Lei J. Chen (&) Department of Neurosurgery, Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, 1095# Jiefang Avenue, Wuhan 430030, Hubei, People’s Republic of China e-mail: [email protected] Z. Li Department of Neurosurgery, Center Hospital of Wuhan City, Wuhan 430014, Hubei, People’s Republic of China J. Jiang Department of Radiology, Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, 1095# Jiefang Avenue, Wuhan 430030, Hubei, People’s Republic of China
was associated with higher rates of recurrence in both CNs (hazard ratio [HR] 6.16, p = 0.046) and EVNs (HR 5.26, p = 0.045), and atypia was also associated with a higher recurrence rate in CNs (HR 5.03, p = 0.042). CNs were thus easier to diagnose than EVNs, with typical clinical, radiological, and histopathological features, while the latter were more likely to show malignant biological behavior associated with atypia and recurrence. Total surgical resection is the optimal treatment choice for both CNs and EVNs, and patients with either CN or EVN with typical and/or totally resected lesions showed favorable clinical outcomes. Keywords Intraventricular central neurocytomas Extraventricular neurocytomas Central nervous system Tumor
Introduction Intraventricular central neurocytomas (CNs) are well-differentiated tumors of neuronal origin. They were first reported as a new entity with a typical immunohistochemical (IHC) profile and ultrastructural features of neuronal differentiation by Hassoun et al. in 1982 [1]. With increased recognition, CNs displayed some common features, including predominant occurrence in young adults, location chiefly at the septum pellucidum and fornix, presentation with increased intracranial hypertension due to obstructive hydrocephalus, and radiological features of calcification and cystic degeneration [2–8]. CNs may also occur as a periventricular parenchymal mass or even in locations remote from the ventricles, as so-called extraventricular neurocytomas (EVNs). Case reports and series have suggested that EVNs have a poorer prognosis than
123
522
CNs [9, 10], and EVNs were classified as a distinct entity in the 2007 World Health Organization classification of tumors of the CNS [11]. However, similar histology and clinical behavior of both CNs and EVNs in some studies [9, 12] may argue against the idea that EVNs are actually a distinct disease from CNs. Now a few case reports have suggested EVNs appear to exhibit a wider morphological spectrum compared with CNs [9, 13], though little has been reported regarding other clinical and clinic-related aspects of EVNs. The specific features of EVNs, including clinical presentations, radiological features, and clinical outcomes, especially in comparison with CNs have never been systematically explored. In the present study, we compared the clinical features, imaging findings, pathological characteristics, and management and clinical outcomes in patients with CN and EVN, and assessed the factors affecting recurrence in CNs and EVNs.
Materials and methods This retrospective analysis included 49 patients (35 CNs and 14 EVNs) with surgically and histologically proven CN or EVN who visited the Department of Neurosurgery of Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology between 2005 and 2012. Their medical records, radiological findings, surgical records, pathological results and follow-up data were analyzed. Medical records included variable demographic data and information on the main symptoms and the duration of the presentation. Radiological findings were obtained by preoperative computed tomography (CT) of the head, pre- and postoperative cerebral magnetic resonance imaging (MRI), and MRI during the follow-up period. Tumoral volumes were determined by measuring the three major axes (a, b, and c) and calculating abc/2. The degree of enhancement was graded as none, minimal-to-mild, or moderate-tomarked. Cystic degeneration components were defined as areas that appeared hypointense on T1-weighted images, hyperintense on T2-weighted images and hypointense on flair images. MR spectroscopy (MRS) was carried out in the solid tumor part and peritumoural white matter in three patients with CNs and two with EVNs. The extent of removal was classified on the basis of postoperative MRI. Gross total resection (GTR) was indicated by the absence of either contrast enhancement or heterogeneous signals on T1- and T2- weighted images, depending on how the tumor was documented preoperatively. Subtotal resection (STR) was indicated by the presence of any residual mass. The features of CNs and EVNs were studied by IHC staining for synaptophysin (Syn), neuron-specific enolase (NSE), neuron-specific nuclear protein (NeuN), glial fibrillary
123
J Neurooncol (2015) 121:521–529
acidic protein (GFAP), Ki-67 cell proliferation-associated nuclear antigen (with monoclonal antibody MIB-1), oligodendrocyte transcription factor 2 (Olig2) and mutated isocitrate dehydrogenase 1 protein (IDH1). Atypical CNs or EVNs, defined as atypical tumors with MIB-1 labeling index C3 % and atypical histological features, were compared with typical tumors. Atypical histological features included necrosis, microvascular proliferation (MPV), infiltration of surrounding parenchyma, and high mitotic count (C3 mitoses/10 high-power fields). A high MIB-1 labeling index may reflect an elevated proliferation potential, and the atypical histological features were the classic histopathological features of malignancy, and both of which are associated with poorer clinical outcome. Clinical outcome of recurrence was considered as the prognostic indicator reflecting the long-term malignant potential of the tumor. Recurrence in this study referred to both tumor progression after STR, defined as an increase in residual size between two follow-up imaging results, and relapse after GTR. The mean follow up period of CNs and EVNs was 55.2 ± 25.6 months (median, 50 months; range, 18–105 months). The median follow-up period was slightly higher in patients with atypical CNs compared with typical CNs (50 vs 48 months, p = 0.824). Continuous variables are presented with standard errors, and differences between variables were analyzed using independent-samples t test. Differences in categorical variables were analyzed using the v2 or Fisher’s exact tests, as appropriate. Cox regression analysis was carried out on the variables in CNs and EVNs that were found to be associated with recurrence. In all statistical tests, values of p \ 0.05 were considered significant. All descriptive and statistical analyses were performed using SPSS version 13.0 (IBM/SPSS).
Results Clinical and radiological findings The clinical features and neuroimaging characteristics of 35 patients with CN and 14 with EVN are summarized in Tables 1 and 2 respectively. CNs occurred in 19 male and 16 female patients, aged 14–47 years (median 28.2 ± 8.4 years), with a symptom duration of 1 month to 3 years (mean 9.03 ± 9.5 months). In contrast, the sex ratio tended to be more equally even in patients with EVNs, with seven male and seven female patients, and the age of onset tended to be older, ranging from 2 to 63 years (mean 33.14 ± 17.6 years), compared with CNs (p = 0.33). The symptom duration in patients with EVNs ranged from 1 day to 2 years (mean 4.14 ± 6.4 months), which was significantly shorter than in patients with CNs (p = 0.044). EVNs were mainly located in the cerebral hemisphere (71.5 %)
J Neurooncol (2015) 121:521–529
523
Increased ICP
29
82.9
1
7.1
features compared with 50 % with EVNs (p = 0.042). The features of atypia in CNs and EVNs are showed in Tables 5 and 6. Atypical CNs mainly demonstrated atypical histological features of MPV and high mitotic count, along with an MIB-1 index [3 %, while atypical EVNs showed various atypical histological features, together with a high MIB-1 index. Immunohistochemically, both CNs and EVNs were positive for the neuronal markers Syn and NeuN, and all EVNs were negative for Olig2 and IDH1 (Fig. 2). GFAP was positive in 34.3 % of CNs and 50 % of EVNs. The MIB-1 proliferation index in CNs ranged from 1 to 10 % (median 2.47 ± 2.2 %), compared with 1.5–30 % in EVNs (median 9.57 ± 8.1 %) (p = 0.006).
General seizure
1
2.9
5
35.8
Management and outcome
Oppression symptoms
2
8.6
7
50.0
Accidental or other
3
5.7
1
7.1
The extent of resection, adjuvant radiotherapy (RT) and clinical outcome in patients with CNs and EVNs are shown in Table 1. Postoperative chemotherapy was not included in this analysis because only a few patients received standard chemotherapy for a sufficient time. Two patients with CNs died within 1 month postoperatively because of an untreatable brain swelling and secondary bleeding after surgical resection, respectively, and one patient with an EVN died one month postoperatively because of central failure associated with tumor-cell dissemination along the cerebrospinal fluid pathways. One patient with CNs was lost to follow-up. These four patients were excluded from the analysis of factors affecting disease recurrence. The rate of recurrence was significantly higher in patients with EVNs (64.3 %) compared with CNs (29.4 %) (p = 0.028). The median time to recurrence after initial treatment was 37.6 ± 20.2 months (range 18–88 months) for CNs, and was 22.33 ± 8.5 months (range 10–42 months) for EVNs (p = 0.049). Multivariate COX analysis in patients with CNs (Table 4) identified atypia (HR 5.03, p = 0.042) and STR (HR 6.16, p = 0.046) as factors associated with higher risk of recurrence. Given the overall rarity of EVNs, only STR and atypia were analyzed as potential factors affecting recurrence, but STR also remained a predictor of recurrence in EVNs according to Cox regression analysis (HR 5.26, p = 0.045).
Table 1 Summary of the clinical and pathological findings and the outcomes of the CNs and EVNs in our series Variable
CNs(35) No.
Age, Y
EVNs(14) %
28.2 ± 8.4
No.
p value %
33.14 ± 17.6
0.33
Sex Female Male
16 19
Onset time, M
9.03 ± 9.5
45.7 54.3
7 7
50.0 50.0
4.14 ± 6.4
0.044
Main symptoms
Resection Total Radiotherapy IHC results
20/35
57.1
8/14
57.1
22/35
62.9
9/14
64.3
Syn
35/35
14/14
NeuN
34/35
12/14
NSE
5/5
GFAP
12/35
MIB 1 index, %
2.47 ± 2.2
Atypical lesions Outcome
2/2 34.3
7/14
50
9.57 ± 8.1
0.006
7
20
7
50
Recurrence
10/34
29.4
Recurrent time (m)
37.6 ± 20.2
9/14
64.3
Lost
1
2.9
0
0.0
Death
2
5.7
1
7.1
22.33 ± 8.5
0.042
0.028 0.049
0.654
Y = year, M month, ICP intracranial pressure, IHC immunohistochemical, Lost lost to follow-up
and showed variable enhancement (Fig. 1) with less cystic degeneration (p = 0.019) and hydrocephalus (p \ 0.001) than CNs. Particularly on the T2-weighted images, half the EVNs demonstrated peritumoral edema and 60 % of CNs showed flow-void signs (FVS). The MRS findings in three CNs and two EVNs are summarized in Tables 3 and 4. Pathological findings Microscopically, CNs and EVNs were composed mainly of uniform, small, round cells with neuronal differentiation in the neuropil that are strongly immunoreactive for Syn. Seven patients (20 %) with CNs demonstrated atypical
Discussion The main clinical characteristics of CNs and EVNs in this study matched those reported in previous clinical studies [2, 14, 15]. The main symptom of CNs was increased intracranial hypertension due to obstructive hydrocephalus in young adults, while patients with EVNs often displayed general seizures or oppression symptoms due to nerve stimulation or mass effect. Compared with CNs, EVNs
123
524 Table 2 CT and MRI findings in CNs and EVNs
J Neurooncol (2015) 121:521–529
Neuroimaging
CNs(35)
p value
No/total
%
No/total
%
CT
Calcification
17/31
54.8
6/13
46.2
MRI
T1
Hypointensity
5/35
14.3
12/14
85.7
Isointensity
30/35
85.7
2/14
14.3
Hypointensity
9/35
25.7
Isointensity
20/35
57.1
11/14
78.6
Hyperintensity
6/35
17.1
3/14
21.4
No enhancement
4/35
11.4
2/14
14.3
Minimal-mild
20/35
57.2
5/14
35.7
Moderatemarked
11/35
31.4
7/14
50.0
T2
Enhanced
0.442
Tumor volume, cm3
43.42 ± 24.9
Cystic degeneration
34/35
97.1
10/14
71.4
FVS in CNs/Peritumoral edema in EVNs
21/35
60.0
7/14
50.0
Hydrocephalus
29/35
82.9
3/14
21.4 \0.001
Location
9 (R. ventricle)
25.7
4 (R. frontal)
28.6
10 (L. ventricle)
28.6
1 (L. temporal)
FVS Flow-void sign, R right, L left, B bilateral, T the third, P pineal region, F the fourth
showed a slightly older age at onset with no sex predilection, a wider age range, and a significantly shorter duration. In our series, EVNs were usually located in the frontal lobe, followed by the temporal and parietal lobes, and were also found in unusual intracranial locations such as the thalamus and hypothalamus [15], pons [16] and spinal cord [17]. Temporal lobe or frontal lobe lesions caused seizures. Oppression symptoms of headaches and hemiparesis were associated with lesions in the frontal lobe and parietal lobes, and occurred when the lesions were large enough. In contrast with the findings of Patil et al. [18] reported, increased intracranial pressure was rare in EVNs in this study, except in patients with the lesions in the posterior fossa. Increased intracranial pressure is commonly associated with an encephalic volume effect caused by cerebral edema or hydrocephalus, and EVNs usually occurred without peritumoral edema and were located remotely from the ventricles. CNs generally presented with a characteristic neuroradiological appearance. Computed tomography scans typically demonstrated an iso-dense or slightly hyper-dense mass with intratumoral calcification and cystic areas within the lateral ventricle near the foramen of Monro, with
123
EVNs(14)
28.85 ± 34.9
0.107 0.019
7.1
8 (B ventricle)
22.8
2 (R. temporal)
14.3
7 (B, T, ventricle)
20.0
2 (R. parietal-occipital)
14.3
1 (B, P, ventricle)
2.9
2 (F ventricle) 1 (R. parietal)
14.3 7.1
1 (vermis)
7.1
1 (sellar region)
7.1
occasional extension into the third ventricle. MRI findings of typical CNs included hypo- to iso-intensity on T1weighted images and iso- to hyper-intensity on T2weighted images, with mild-moderate enhancement. CNs had a distinctive ‘‘bubbly’’ appearance on MR, and were moderate-enhancing intraventricular lesions with frequent calcification. In contrast, EVNs presented with variable manifestations of circumscribed, sometimes large, complex, and heterogeneous-enhancement masses. EVNs were often partly or mainly cystic, frequently calcified, and sporadically showed a flow-void sign, and half of the EVNs in the present study were associated with peritumoral edema. Yi et al. [19], observed cystic degeneration in 40 % of EVNs, calcification in 30 %, enhancement on contrastenhanced T1-weighted images in 80 % and perilesional edema in 10 %. Such subtle differences in peritumoral edema and cysts in our study could be explained by the histological characteristics of EVNs, which exhibited a wide cell morphological spectrum, low cellularity, and ganglion differentiation [9]. Similar to the MRS results of Ueda [20] in our series, the presence of the N-acetylaspartate (NAA) signal and high MI and/or Gly signals may be the characteristic features of CNs. A combination of
J Neurooncol (2015) 121:521–529
525
Fig. 1 a–d Preoperative MRI and MRS of a patient with CN, a T1weighted image showing isointensity in the cystic tumor in the region of the septum pellucidum. b Hypointensity on T2-weighted image. c Post-contrast T1-weighted image showing moderate-marked enhancement in the solid component after administration of gadolinium. d Automated MR spectra showing small lactate doublet, small N-acetylaspartate resonance, prominent resonance from choline,
detectable creatine resonance, and high MI and/or Gly signals. e– h A 63-year-old man with EVN in the left temporal lobe. e T1weighted image. f T2-weighted image. g Post-contrast T1-weighted image showing a cortically-based infiltrative lesion with cyst and heterogeneous enhancement. h Automated MR spectra showing definite lactate doublet, small N-acetylaspartate resonance, prominent Cho resonance and detectable Cr resonance
prominent Cho resonance and detectable Cr resonance is an ordinary feature of both CNs and EVNs. Consequently, EVNs show variable imaging features compared with CNs. When a patient presents with a cortical based hemispheric lesion with variable contrast enhancement, cystic component, calcifications, remarkable Cho resonance and detectable Cr resonance on MRS, but lacking peritumoral edema, EVNs should thus be included in the differential diagnosis. Histopathologically, typical CNs and EVNs were composed predominantly of uniform small round cells with neuronal differentiation and clear cytoplasms embedded in a neuropil that are strongly immunoreactive for Syn, and which are consistent with current reports [7, 9, 13, 21]. The results of the present study showed that light microscopy of CNs and EVNs commonly revealed fibrillary matrices alternating with anuclear areas and focal calcifications. Tumor cells in EVNs exhibited a wider morphological spectrum than those in CNs, and were arranged in sheets, clusters, ribbons, or rosettes, with neuropils dispersed either in broad zones (neuropil islands). Atypical pathological features of CNs and EVNs in this study usually included high mitotic count and/or MVP, and focal necrosis and infiltration, also in accord with Brat et al.’s
study of patients with 35 EVNs [9]. In the present study, most CNs and EVNs with atypical histological features associated with malignancy also showed high MIB-1 indexes. Sharma et al. [22] also found that the MIB-1 monoclonal antibody index of multiple proliferation markers tended to be higher in CNs with mitosis and necrosis. Soylemezoglu et al. [23] found a MIB-1 labeling index [2 % in 39 % of the specimens of CNs which was significantly correlated with microvascular proliferation (p = 0.0006). Overall, these results suggest an association between MIB-1 index and atypical histological features, and that patients with atypia defined as atypical histological features with MIB-1 index [3, are more likely to have poorer clinical outcomes. CNs can be diagnosed based on their typical clinical, neuroradiological, histomorphological and immunohistochemical presentations. In contrast, it is difficult to distinguish between EVNs and other brain neoplasms because of their heterogeneous expressions. Syn has been recognized as the most reliable diagnostic marker for identifying neuronal differentiation in brain tumors [24], and may thus play a vital role in the diagnosis of EVNs, based on a histopathological appearance of monotonous neoplastic cells with round, regular nuclei embedded within a fine
123
526
J Neurooncol (2015) 121:521–529
Table 3 Metabolites observed in the 1H MR spectra of the CNs and EVNs Patient CNs
EVNs
MI/ Gly
Cho
Cr
NAA
Lac
Lip
1
H
H
H
H
–
–
2
H
H
H
H
H
H
3 4
H H
H H
H H
H H
– H
– H
5
–
H
H
–
–
–
Cho Choline, Cr creatine, Gly glycine, Lac lactate, NAA N-acetylaspartate, lip lipid
neuropil background, strongly immunoreactive for Syn [9, 25]. The main differential diagnoses of EVNs include oligodendroglioma with neurocytic differentiation, oligoastrocytoma with neurocytic differentiation, ganglioglioma, pilocytic astrocytoma, ependymoma and dysembryoblastic neuroepithelial tumor (DNT). Radiologically, EVNs demonstrate radiological features that are likely to be contrast-enhancing and/or cystic, depict a circumscribed, rather than a diffusely-infiltrating mass, while oligodendrogliomas and other EVNs-mimicking gliomas are typically more infiltrative than EVNs. Morphologically, ependymomas show more typical rosettes [9] and DNTs show more mucoid areas and are less cellular than EVNs [26]. Immunohistochemically, oligodendrogliomas can show some Syn positivity, but to a lesser degree or extent
Table 4 Cox Regression
Dependent variable
than EVNs [27, 28]. Positivity for NeuN also supports a diagnosis of EVNs rather than oligodendrogliomas or oligoastrocytomas [25] and Olig2 is useful for differentiating between oligodendrogliomas and EVNs [29]. Further, an absence of IDH1 expression is a powerful indicator for distinguishing between EVNs and EVNs-mimicking gliomas [30]. However, ultrastructural studies are sometimes needed to demonstrate unequivocal neuronal features in EVNs [24]. Moreover, twelve cases of CNs showed the expression of GFAP accounting for a big proportion of CNs in our study, and 6 of these cases showed atypical histopathological features and 5 cases recurred during the follow up. Elek et al’ [31] also found a high incidence of GFAPpositivity, suggesting that GFAP positivity might be associated with a malignant course in CNs. Based on the above clinical data for CNs and EVNs, most patients have achieved total resection and had a favorable prognosis, though a few patients required more than one operation. The Mortality was around 6 % in our study, which was the same as that found by Rades et al. [32] in a larger series of neurocytomas patients. The recurrence rates in the present study was 29.4 % for CNs and 64.3 % for EVNs, compared with 21 % in 126 CN patients reported by Vasiljevic et al. [33]. Kane et al. [15] carried out a literature search including case reports and case series comprising 85 EVNs, and found a recurrence rate of 52 %, including 36 % in patients with typical lesions and 68 % in atypical lesions. These results were
CNs
EVNs
HR (95 % CI)
p value
HR (95 % CI)
p value
Male sex
0.75 (0.18, 3.14)
0.696
Age
1.01 (0.94, 1.10)
0.767
Histological atypia
5.03 (1.06, 23.96)
0.042
5.38 (0.99, 29.21)
0.051
Subtotal resection Tumor volume
6.16 (1.04, 36.68)
0.046
5.26 (1.04, 26.66)
0.045
1.01 (0.98, 1.04)
0.580
HR Hazard rate, CI confidence intervals
Radiological therapy
0.33 (0.07, 1.51)
0.154
Table 5 Atypical features of seven CNs cases
Atypical CNs
Case 1
Case 2 Case 3
Case 4
Case 5
Case 6 Case 7
MIB 1 labeling index (%)
5
4
8
7.5
4
10
5
–
3
2
–
3
–
3
Atypical histological features High mitotic count (C3 mitoses/ 10HPFs)
123
Microvascular proliferation (MVP)
?
–
?
–
–
?
?
Necrosis
–
–
–
–
–
–
–
Infiltration of parenchyma
–
–
–
–
–
–
–
J Neurooncol (2015) 121:521–529 Table 6 The location, pathological features and immunohistochemistry data of 14 EVNs No.
Age
Location
Main pathological features and immunohistochemistry data
1
40
R. temporal
Atypical, sheet-like and cell clusters growth with neuropil islands, 3 mitoses/ 10HPFs, MVP, no infiltration, Syn?, NeuN?, GFAP?, Olig2-, IDH1-, MIB1 7.5 %;
2
13
R. parietal occipita
Atypical, sheet-like and cell clusters growth with neuropil islands, 6 mitoses/ 10HPFs, focal infiltration, Syn?, NeuN?, GFAP?, Olig2-, IDH1-, MIB1 30 %;
3
56
Sellar region
Typical, sheet-like growth with neuropil islands, no mitoses/10HPFs, Syn?, NeuN-, NSE?, GFAP-, LH-, FSH-, GH-, ACTH-, PRL-, TSH-, Olig2, IDH1-, MIB1 1.5 %;
4
4
F. ventricle
Atypical, sheet-like and cell clusters growth with neuropil islands, focal ganglion cells, 3 mitoses/10HPFs, MVP, no infiltration, Syn?, NeuN?, GFAP-, Olig2-, IDH1-, MIB1 7.5 %;
5
32
R. parietal
Atypical, sheet-like and cell clusters growth with neuropil islands, 4 mitoses/ 10HPFs, focal ganglionic cells, no infiltration, Syn?, NeuN?, GFAP?, Olig2-, IDH1-, MIB1 7.5 %;
6
43
R. parietaloccipita
Typical, sheet-like and cell clusters growth with neuropil islands, 2 mitoses/ 10HPFs, Syn?, NeuN-, NSE?, GFAP?, Olig2-, IDH1-, MIB1 2.5 %;
7
29
R. frontal
Atypical, cluster-like growth with neuropil islands, 6 mitoses/10HPFs, MVP, necrosis, focal ganglionic cells, Syn?, NeuN?, GFAP-, Olig2-, IDH1-, MIB1 20 %;
8
37
R. temporal
Typical, sheet-like and cell clusters growth with neuropil islands, 2 mitoses/ 10HPFs, focal ganglionic cells, Syn?, NeuN?, GFAP?, Olig2-, IDH1-, MIB1 7.5 %;
9
38
R. frontal
Typical, sheet-like and cell clusters growth with neuropil islands, no mitoses/10HPFs; Syn?, NeuN?, GFAP-, Olig2-, IDH1-, MIB1 5 %;
10
2
R. frontal
Atypical, sheet-like growth with focal rosettes, 5 mitoses/10HPFs, no infiltration, Syn?, NeuN?, GFAP-, Olig2-, IDH1-, MIB1 10 %;
11
25
F. ventricle
Typical, sheet-like and cell clusters growth with neuropil islands, no mitoses/10HPFs, no infiltration, Syn?, NeuN?, GFAP?, Olig2-, IDH1-, MIB1 5 %;
12
41
Cerebellum vermis
Typical, sheet-like growth, 2 mitoses/ 10HPFs; no infiltration, Syn?, NeuN?, GFAP-, Olig2-, IDH1-, MIB1 5 %;
527 Table 6 continued No.
Age
Location
Main pathological features and immunohistochemistry data
13
63
L. temporal
Atypical, sheet-like and cell clusters growth with neuropil islands, 7 mitoses/ 10HPFs, focal ganglionic cells, MVP, necrosis, no infiltration, Syn?, NeuN?, GFAP?, Olig2-, IDH1-, MIB1 20 %;
14
42
R. frontal
Typical, sheet-like growth, no mitoses/ 10HPFs, no infiltration, Syn?, NeuN?, GFAP-, Olig2-, IDH1-, MIB1 5 %;
R Right, L left, MVP microvascular proliferation
similar to those of the present study, and showed a significantly higher recurrence rate for EVNs, especially in aggressively atypical cases. Most patient deaths occurred in the perioperative period, and we therefore chose recurrence as the prognostic indicator to reflect malignant potential. Multivariate analysis revealed a clear superiority of GTR over STR in both CNs and EVNs, in accordance with previous studies [15, 33]. The analysis also showed that atypia was associated with significantly higher rates of recurrence in CNs, but was not obviously in EVNs. Kane et al. [15] found that the recurrence risk was 2–3-fold higher in atypical compared with typical EVNs. Brat et al. [9] demonstrated cellular proliferation and the histological features of atypia formed the basis for aggressive tumor growth and affected the resectability of tumors, thus increasing the risk of STR and leading to a higher recurrence rate. It was noteworthy that atypical lesions accounted for 66.7 % of recurrent cases of EVNs in the current study. However the sample size was too small to draw conclusions about the prognostic value of atypia in EVNs, and further studies are needed to verify these preliminary results. Although radiotherapy was not independently associated with reduced recurrence in CNs, four typical and one atypical CNs with single STR and postoperative radiotherapy did not recur during follow-up periods of 37–61 months (mean 46.8 months) in our series. Several reports have claimed that postoperative radiotherapy for CNs after STR leads to shrinkage of residual tumors, while radiotherapy after GTR remains controversial, given that most patients have longterm tumor control without radiotherapy. Paek et al. [34] found that radiotherapy effectively controlled the residual CNs after surgery and prevented recurrence with a median follow-up periods of 171 months (range 128–229) and 202 months (range165–227), respectively. Hallock’s et al. [35] found a median relapse-free time of 79.3 months for CNs in a long-term study. Kane et al. [15] demonstrated that adjuvant radiotherapy may improve recurrence rates following STR in EVNs. Radiotherapy is thus generally
123
528
J Neurooncol (2015) 121:521–529
Fig. 2 a–i Histopathologic features of EVNs, showing some atypical histological characteristics. Tumor cells were most often found in sheets (a) or in cell clusters with neuropil islands (b). EVNs were all strongly and diffusely immunoreactive for Syn (c). NeuN stain was positive in a nuclei with less frequent architectural patterns of
neurocytic rosettes (d, arrow). Atypia in EVNs including atypical histological features of high mitotic count (C3 mitoses/10 high-power fields) (e, arrow) and microvascular proliferation (f), along with MIB1 labeling index [3 % (g). EVNs were negative for IDH1 (h) and Olig2 (i)
believed to plays a role in the local control of CNs and EVNs, which may have become evident in the current study with a longer follow-up period.
Acknowledgments This work was supported by the project of National Clinical Specialist Facility.
Conclusions In an overall comparison between CNs and EVNs, CNs are easier to diagnose, having typical clinical, radiological, and histological features, while EVNs are more likely to show malignant biological behavior, with atypia and recurrence. STR and atypia are the main predictors of recurrence in both CNs and EVNs. The optimal treatment choice for CNs and EVNs is surgical total resection. Patients with typical and/or completely-resected CNs or EVNs have favorable clinical outcomes. Further studies with more cases are needed to confirm the findings of this study.
123
Conflict of Interest None of the authors has any personal, financial, or professional conflicts of interest to report.
References 1. Hassoun J, Gambarelli D, Grisoli F et al (1982) Central neurocytoma. An electron-microscopic study of two cases. Acta Neuropathol 56(2):151–156 2. Hassoun J, Soylemezoglu F, Gambarelli D, Figarella-Branger D, von Ammon K, Kleihues P (1993) Central neurocytoma: a synopsis of clinical and histological features. Brain Pathol 3(3):297–306 3. Kim DG, Kim JS, Chi JG et al (1996) Central neurocytoma: proliferative potential and biological behavior. J Neurosurg 84(5):742–747 4. Valdueza JM, Westphal M, Vortmeyer A, Muller D, Padberg B, Herrmann HD (1996) Central neurocytoma: clinical, immunohistologic, and biologic findings of a human neuroglial progenitor tumor. Surg Neurol 45(1):49–56
J Neurooncol (2015) 121:521–529 5. Chang KH, Han MH, Kim DG et al (1993) MR appearance of central neurocytoma. Acta Radiol 34(5):520–526 6. Ashkan K, Casey AT, D’Arrigo C, Harkness WF, Thomas DG (2000) Benign central neurocytoma. Cancer 89(5):1111–1120 7. Schmidt MH, Gottfried ON, von Koch CS, Chang SM, McDermott MW (2004) Central neurocytoma: a review. J Neurooncol 66(3):377–384 8. Sharma MC, Deb P, Sharma S, Sarkar C (2006) Neurocytoma: a comprehensive review. Neurosurg Rev 29(4):270–285 discussion 285 9. Brat DJ, Scheithauer BW, Eberhart CG, Burger PC (2001) Extraventricular neurocytomas: pathologic features and clinical outcome. Am J Surg Pathol 25(10):1252–1260 10. Rodriguez FJ, Mota RA, Scheithauer BW et al (2009) Interphase cytogenetics for 1p19q and t(1;19)(q10;p10) may distinguish prognostically relevant subgroups in extraventricular neurocytoma. Brain Pathol 19(4):623–629 11. Louis DN, Ohgaki H, Wiestler OD et al (2007) The 2007 WHO classification of tumours of the central nervous system. Acta Neuropathol 114(2):97–109 12. Brown DM, Karlovits S, Lee LH, Kim K, Rothfus WE, Brown HG (2001) Management of neurocytomas: case report and review of the literature. Am J Clin Oncol 24(3):272–278 13. Giangaspero F, Cenacchi G, Losi L, Cerasoli S, Bisceglia M, Burger PC (1997) Extraventricular neoplasms with neurocytoma features. A clinicopathological study of 11 cases. Am J Surg Pathol 21(2):206–212 14. Rades D, Fehlauer F, Lamszus K et al (2005) Well-differentiated neurocytoma: what is the best available treatment? Neuro Oncol 7(1):77–83 15. Kane AJ, Sughrue ME, Rutkowski MJ et al (2012) Atypia predicting prognosis for intracranial extraventricular neurocytomas. J Neurosurg 116(2):349–354 16. Swinson BM, Friedman WA, Yachnis AT (2006) Pontine atypical neurocytoma: case report. Neurosurgery 58(5):E990 discussion E990 17. Tatter SB, Borges LF, Louis DN (1995) Correction: central neurocytomas of the cervical spinal cord. J Neurosurg 82(4):706 18. Patil AS, Menon G, Easwer HV, Nair S (2013) Extraventricular neurocytoma, a comprehensive review. Acta Neurochir (Wien) 68(4):e206–e212 19. Yi KS, Sohn CH, Yun TJ, et al (2012) MR imaging findings of extraventricular neurocytoma: a series of ten patients confirmed by immunohistochemistry of IDH1 gene mutation. Acta Neurochir (Wien), 154(11):1973–1979; discussion 1980 20. Ueda F, Suzuki M, Matsui O, Uchiyama N (2007) Automated MR spectroscopy of intra- and extraventricular neurocytomas. Magn Reson Med Sci 6(2):75–81
529 21. von Deimling A, Janzer R, Kleihues P, Wiestler OD (1990) Patterns of differentiation in central neurocytoma. An immunohistochemical study of eleven biopsies. Acta Neuropathol 79(5):473–479 22. Sharma MC, Rathore A, Karak AK, Sarkar C (1998) A study of proliferative markers in central neurocytoma. Pathology 30(4):355–359 23. Soylemezoglu F, Scheithauer BW, Esteve J, Kleihues P (1997) Atypical central neurocytoma. J Neuropathol Exp Neurol 56(5):551–556 24. Robbins P, Segal A, Narula S et al (1995) Central neurocytoma. A clinicopathological, immunohistochemical and ultrastructural study of 7 cases. Pathol Res Pract 191(2):100–111 25. Agarwal S, Sharma MC, Sarkar C et al (2011) Extraventricular neurocytomas: a morphological and histogenetic consideration. A study of six cases. Pathology 43(4):327–334 26. Jay V, Edwards V, Hoving E et al (1999) Central neurocytoma: morphological, flow cytometric, polymerase chain reaction, fluorescence in situ hybridization, and karyotypic analyses Case report. J Neurosurg 90(2):348–354 27. Mut M, Guler-Tezel G, Lopes MB, Bilginer B, Ziyal I, Ozcan OE (2005) Challenging diagnosis: oligodendroglioma versus extraventricular neurocytoma. Clin Neuropathol 24(5):225–229 28. Perry A, Scheithauer BW, Macaulay RJ, Raffel C, Roth KA, Kros JM (2002) Oligodendrogliomas with neurocytic differentiation. A report of 4 cases with diagnostic and histogenetic implications. J Neuropathol Exp Neurol 61(11):947–955 29. Okada M, Yano H, Hirose Y et al (2011) Olig2 is useful in the differential diagnosis of oligodendrogliomas and extraventricular neurocytomas. Brain Tumor Pathol 28(2):157–161 30. Capper D, Reuss D, Schittenhelm J et al (2011) Mutation-specific IDH1 antibody differentiates oligodendrogliomas and oligoastrocytomas from other brain tumors with oligodendroglioma-like morphology. Acta Neuropathol 121(2):241–252 31. Elek G, Slowik F, Eross L, Toth S, Szabo Z, Balint K (1999) Central neurocytoma with malignant course. Neuronal and glial differentiation and craniospinal dissemination. Pathol Oncol Res 5(2):155–159 32. Rades D, Schild SE (2006) Treatment recommendations for the various subgroups of neurocytomas. J Neurooncol 77(3):305–309 33. Vasiljevic A, Francois P, Loundou A et al (2012) Prognostic factors in central neurocytomas: a multicenter study of 71 cases. Am J Surg Pathol 36(2):220–227 34. Paek SH, Han JH, Kim JW et al (2008) Long-term outcome of conventional radiation therapy for central neurocytoma. J Neurooncol 90(1):25–30 35. Hallock A, Hamilton B, Ang LC et al (2011) Neurocytomas: longterm experience of a single institution. Neuro Oncol 13(9):943–949
123
CLINICAL STUDY
A comparative study of intraventricular central neurocytomas and extraventricular neurocytomas Zhongwei Xiong • Jianjian Zhang • Zhengwei Li Jingjing Jiang • Qingdong Han • Shoujia Sun • Xiaolin Wu • Yu Wang • Ting Lei • Jincao Chen
•
Received: 22 February 2014 / Accepted: 6 November 2014 / Published online: 11 November 2014 Ó Springer Science+Business Media New York 2014
Abstract Similar histology and clinical behavior of both intraventricular central neurocytomas (CNs) and extraventricular neurocytomas (EVNs) may argue against the idea that EVNs were the distinct entity to distinguish from CNs in the 2007 World Health Organization classification. To explore respective characteristics and compare similarities and differences in CNs and EVNs, relevant clinical, radiological, operative and pathological data of 49 patients (35 CNs and 14 EVNs) in the Department of Neurosurgery at our hospital from 2005 to 2012 was reviewed and some comparisons between CNs and EVNs were conducted. The factors affecting posttreatment recurrence of CNs and EVNs were assessed by Cox regression analysis. In comparison, CNs showed a more typical clinical manifestation, and radiological and histopathological features, while EVNs demonstrated more malignant biological behavior, with higher MIB-1 index (p = 0.006), higher rate of atypia (p = 0.042), higher recurrence rate (p = 0.028), and shorter time to recurrence (p = 0.049). Subtotal resection
Z. Xiong J. Zhang Q. Han S. Sun X. Wu Y. Wang T. Lei J. Chen (&) Department of Neurosurgery, Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, 1095# Jiefang Avenue, Wuhan 430030, Hubei, People’s Republic of China e-mail: [email protected] Z. Li Department of Neurosurgery, Center Hospital of Wuhan City, Wuhan 430014, Hubei, People’s Republic of China J. Jiang Department of Radiology, Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, 1095# Jiefang Avenue, Wuhan 430030, Hubei, People’s Republic of China
was associated with higher rates of recurrence in both CNs (hazard ratio [HR] 6.16, p = 0.046) and EVNs (HR 5.26, p = 0.045), and atypia was also associated with a higher recurrence rate in CNs (HR 5.03, p = 0.042). CNs were thus easier to diagnose than EVNs, with typical clinical, radiological, and histopathological features, while the latter were more likely to show malignant biological behavior associated with atypia and recurrence. Total surgical resection is the optimal treatment choice for both CNs and EVNs, and patients with either CN or EVN with typical and/or totally resected lesions showed favorable clinical outcomes. Keywords Intraventricular central neurocytomas Extraventricular neurocytomas Central nervous system Tumor
Introduction Intraventricular central neurocytomas (CNs) are well-differentiated tumors of neuronal origin. They were first reported as a new entity with a typical immunohistochemical (IHC) profile and ultrastructural features of neuronal differentiation by Hassoun et al. in 1982 [1]. With increased recognition, CNs displayed some common features, including predominant occurrence in young adults, location chiefly at the septum pellucidum and fornix, presentation with increased intracranial hypertension due to obstructive hydrocephalus, and radiological features of calcification and cystic degeneration [2–8]. CNs may also occur as a periventricular parenchymal mass or even in locations remote from the ventricles, as so-called extraventricular neurocytomas (EVNs). Case reports and series have suggested that EVNs have a poorer prognosis than
123
522
CNs [9, 10], and EVNs were classified as a distinct entity in the 2007 World Health Organization classification of tumors of the CNS [11]. However, similar histology and clinical behavior of both CNs and EVNs in some studies [9, 12] may argue against the idea that EVNs are actually a distinct disease from CNs. Now a few case reports have suggested EVNs appear to exhibit a wider morphological spectrum compared with CNs [9, 13], though little has been reported regarding other clinical and clinic-related aspects of EVNs. The specific features of EVNs, including clinical presentations, radiological features, and clinical outcomes, especially in comparison with CNs have never been systematically explored. In the present study, we compared the clinical features, imaging findings, pathological characteristics, and management and clinical outcomes in patients with CN and EVN, and assessed the factors affecting recurrence in CNs and EVNs.
Materials and methods This retrospective analysis included 49 patients (35 CNs and 14 EVNs) with surgically and histologically proven CN or EVN who visited the Department of Neurosurgery of Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology between 2005 and 2012. Their medical records, radiological findings, surgical records, pathological results and follow-up data were analyzed. Medical records included variable demographic data and information on the main symptoms and the duration of the presentation. Radiological findings were obtained by preoperative computed tomography (CT) of the head, pre- and postoperative cerebral magnetic resonance imaging (MRI), and MRI during the follow-up period. Tumoral volumes were determined by measuring the three major axes (a, b, and c) and calculating abc/2. The degree of enhancement was graded as none, minimal-to-mild, or moderate-tomarked. Cystic degeneration components were defined as areas that appeared hypointense on T1-weighted images, hyperintense on T2-weighted images and hypointense on flair images. MR spectroscopy (MRS) was carried out in the solid tumor part and peritumoural white matter in three patients with CNs and two with EVNs. The extent of removal was classified on the basis of postoperative MRI. Gross total resection (GTR) was indicated by the absence of either contrast enhancement or heterogeneous signals on T1- and T2- weighted images, depending on how the tumor was documented preoperatively. Subtotal resection (STR) was indicated by the presence of any residual mass. The features of CNs and EVNs were studied by IHC staining for synaptophysin (Syn), neuron-specific enolase (NSE), neuron-specific nuclear protein (NeuN), glial fibrillary
123
J Neurooncol (2015) 121:521–529
acidic protein (GFAP), Ki-67 cell proliferation-associated nuclear antigen (with monoclonal antibody MIB-1), oligodendrocyte transcription factor 2 (Olig2) and mutated isocitrate dehydrogenase 1 protein (IDH1). Atypical CNs or EVNs, defined as atypical tumors with MIB-1 labeling index C3 % and atypical histological features, were compared with typical tumors. Atypical histological features included necrosis, microvascular proliferation (MPV), infiltration of surrounding parenchyma, and high mitotic count (C3 mitoses/10 high-power fields). A high MIB-1 labeling index may reflect an elevated proliferation potential, and the atypical histological features were the classic histopathological features of malignancy, and both of which are associated with poorer clinical outcome. Clinical outcome of recurrence was considered as the prognostic indicator reflecting the long-term malignant potential of the tumor. Recurrence in this study referred to both tumor progression after STR, defined as an increase in residual size between two follow-up imaging results, and relapse after GTR. The mean follow up period of CNs and EVNs was 55.2 ± 25.6 months (median, 50 months; range, 18–105 months). The median follow-up period was slightly higher in patients with atypical CNs compared with typical CNs (50 vs 48 months, p = 0.824). Continuous variables are presented with standard errors, and differences between variables were analyzed using independent-samples t test. Differences in categorical variables were analyzed using the v2 or Fisher’s exact tests, as appropriate. Cox regression analysis was carried out on the variables in CNs and EVNs that were found to be associated with recurrence. In all statistical tests, values of p \ 0.05 were considered significant. All descriptive and statistical analyses were performed using SPSS version 13.0 (IBM/SPSS).
Results Clinical and radiological findings The clinical features and neuroimaging characteristics of 35 patients with CN and 14 with EVN are summarized in Tables 1 and 2 respectively. CNs occurred in 19 male and 16 female patients, aged 14–47 years (median 28.2 ± 8.4 years), with a symptom duration of 1 month to 3 years (mean 9.03 ± 9.5 months). In contrast, the sex ratio tended to be more equally even in patients with EVNs, with seven male and seven female patients, and the age of onset tended to be older, ranging from 2 to 63 years (mean 33.14 ± 17.6 years), compared with CNs (p = 0.33). The symptom duration in patients with EVNs ranged from 1 day to 2 years (mean 4.14 ± 6.4 months), which was significantly shorter than in patients with CNs (p = 0.044). EVNs were mainly located in the cerebral hemisphere (71.5 %)
J Neurooncol (2015) 121:521–529
523
Increased ICP
29
82.9
1
7.1
features compared with 50 % with EVNs (p = 0.042). The features of atypia in CNs and EVNs are showed in Tables 5 and 6. Atypical CNs mainly demonstrated atypical histological features of MPV and high mitotic count, along with an MIB-1 index [3 %, while atypical EVNs showed various atypical histological features, together with a high MIB-1 index. Immunohistochemically, both CNs and EVNs were positive for the neuronal markers Syn and NeuN, and all EVNs were negative for Olig2 and IDH1 (Fig. 2). GFAP was positive in 34.3 % of CNs and 50 % of EVNs. The MIB-1 proliferation index in CNs ranged from 1 to 10 % (median 2.47 ± 2.2 %), compared with 1.5–30 % in EVNs (median 9.57 ± 8.1 %) (p = 0.006).
General seizure
1
2.9
5
35.8
Management and outcome
Oppression symptoms
2
8.6
7
50.0
Accidental or other
3
5.7
1
7.1
The extent of resection, adjuvant radiotherapy (RT) and clinical outcome in patients with CNs and EVNs are shown in Table 1. Postoperative chemotherapy was not included in this analysis because only a few patients received standard chemotherapy for a sufficient time. Two patients with CNs died within 1 month postoperatively because of an untreatable brain swelling and secondary bleeding after surgical resection, respectively, and one patient with an EVN died one month postoperatively because of central failure associated with tumor-cell dissemination along the cerebrospinal fluid pathways. One patient with CNs was lost to follow-up. These four patients were excluded from the analysis of factors affecting disease recurrence. The rate of recurrence was significantly higher in patients with EVNs (64.3 %) compared with CNs (29.4 %) (p = 0.028). The median time to recurrence after initial treatment was 37.6 ± 20.2 months (range 18–88 months) for CNs, and was 22.33 ± 8.5 months (range 10–42 months) for EVNs (p = 0.049). Multivariate COX analysis in patients with CNs (Table 4) identified atypia (HR 5.03, p = 0.042) and STR (HR 6.16, p = 0.046) as factors associated with higher risk of recurrence. Given the overall rarity of EVNs, only STR and atypia were analyzed as potential factors affecting recurrence, but STR also remained a predictor of recurrence in EVNs according to Cox regression analysis (HR 5.26, p = 0.045).
Table 1 Summary of the clinical and pathological findings and the outcomes of the CNs and EVNs in our series Variable
CNs(35) No.
Age, Y
EVNs(14) %
28.2 ± 8.4
No.
p value %
33.14 ± 17.6
0.33
Sex Female Male
16 19
Onset time, M
9.03 ± 9.5
45.7 54.3
7 7
50.0 50.0
4.14 ± 6.4
0.044
Main symptoms
Resection Total Radiotherapy IHC results
20/35
57.1
8/14
57.1
22/35
62.9
9/14
64.3
Syn
35/35
14/14
NeuN
34/35
12/14
NSE
5/5
GFAP
12/35
MIB 1 index, %
2.47 ± 2.2
Atypical lesions Outcome
2/2 34.3
7/14
50
9.57 ± 8.1
0.006
7
20
7
50
Recurrence
10/34
29.4
Recurrent time (m)
37.6 ± 20.2
9/14
64.3
Lost
1
2.9
0
0.0
Death
2
5.7
1
7.1
22.33 ± 8.5
0.042
0.028 0.049
0.654
Y = year, M month, ICP intracranial pressure, IHC immunohistochemical, Lost lost to follow-up
and showed variable enhancement (Fig. 1) with less cystic degeneration (p = 0.019) and hydrocephalus (p \ 0.001) than CNs. Particularly on the T2-weighted images, half the EVNs demonstrated peritumoral edema and 60 % of CNs showed flow-void signs (FVS). The MRS findings in three CNs and two EVNs are summarized in Tables 3 and 4. Pathological findings Microscopically, CNs and EVNs were composed mainly of uniform, small, round cells with neuronal differentiation in the neuropil that are strongly immunoreactive for Syn. Seven patients (20 %) with CNs demonstrated atypical
Discussion The main clinical characteristics of CNs and EVNs in this study matched those reported in previous clinical studies [2, 14, 15]. The main symptom of CNs was increased intracranial hypertension due to obstructive hydrocephalus in young adults, while patients with EVNs often displayed general seizures or oppression symptoms due to nerve stimulation or mass effect. Compared with CNs, EVNs
123
524 Table 2 CT and MRI findings in CNs and EVNs
J Neurooncol (2015) 121:521–529
Neuroimaging
CNs(35)
p value
No/total
%
No/total
%
CT
Calcification
17/31
54.8
6/13
46.2
MRI
T1
Hypointensity
5/35
14.3
12/14
85.7
Isointensity
30/35
85.7
2/14
14.3
Hypointensity
9/35
25.7
Isointensity
20/35
57.1
11/14
78.6
Hyperintensity
6/35
17.1
3/14
21.4
No enhancement
4/35
11.4
2/14
14.3
Minimal-mild
20/35
57.2
5/14
35.7
Moderatemarked
11/35
31.4
7/14
50.0
T2
Enhanced
0.442
Tumor volume, cm3
43.42 ± 24.9
Cystic degeneration
34/35
97.1
10/14
71.4
FVS in CNs/Peritumoral edema in EVNs
21/35
60.0
7/14
50.0
Hydrocephalus
29/35
82.9
3/14
21.4 \0.001
Location
9 (R. ventricle)
25.7
4 (R. frontal)
28.6
10 (L. ventricle)
28.6
1 (L. temporal)
FVS Flow-void sign, R right, L left, B bilateral, T the third, P pineal region, F the fourth
showed a slightly older age at onset with no sex predilection, a wider age range, and a significantly shorter duration. In our series, EVNs were usually located in the frontal lobe, followed by the temporal and parietal lobes, and were also found in unusual intracranial locations such as the thalamus and hypothalamus [15], pons [16] and spinal cord [17]. Temporal lobe or frontal lobe lesions caused seizures. Oppression symptoms of headaches and hemiparesis were associated with lesions in the frontal lobe and parietal lobes, and occurred when the lesions were large enough. In contrast with the findings of Patil et al. [18] reported, increased intracranial pressure was rare in EVNs in this study, except in patients with the lesions in the posterior fossa. Increased intracranial pressure is commonly associated with an encephalic volume effect caused by cerebral edema or hydrocephalus, and EVNs usually occurred without peritumoral edema and were located remotely from the ventricles. CNs generally presented with a characteristic neuroradiological appearance. Computed tomography scans typically demonstrated an iso-dense or slightly hyper-dense mass with intratumoral calcification and cystic areas within the lateral ventricle near the foramen of Monro, with
123
EVNs(14)
28.85 ± 34.9
0.107 0.019
7.1
8 (B ventricle)
22.8
2 (R. temporal)
14.3
7 (B, T, ventricle)
20.0
2 (R. parietal-occipital)
14.3
1 (B, P, ventricle)
2.9
2 (F ventricle) 1 (R. parietal)
14.3 7.1
1 (vermis)
7.1
1 (sellar region)
7.1
occasional extension into the third ventricle. MRI findings of typical CNs included hypo- to iso-intensity on T1weighted images and iso- to hyper-intensity on T2weighted images, with mild-moderate enhancement. CNs had a distinctive ‘‘bubbly’’ appearance on MR, and were moderate-enhancing intraventricular lesions with frequent calcification. In contrast, EVNs presented with variable manifestations of circumscribed, sometimes large, complex, and heterogeneous-enhancement masses. EVNs were often partly or mainly cystic, frequently calcified, and sporadically showed a flow-void sign, and half of the EVNs in the present study were associated with peritumoral edema. Yi et al. [19], observed cystic degeneration in 40 % of EVNs, calcification in 30 %, enhancement on contrastenhanced T1-weighted images in 80 % and perilesional edema in 10 %. Such subtle differences in peritumoral edema and cysts in our study could be explained by the histological characteristics of EVNs, which exhibited a wide cell morphological spectrum, low cellularity, and ganglion differentiation [9]. Similar to the MRS results of Ueda [20] in our series, the presence of the N-acetylaspartate (NAA) signal and high MI and/or Gly signals may be the characteristic features of CNs. A combination of
J Neurooncol (2015) 121:521–529
525
Fig. 1 a–d Preoperative MRI and MRS of a patient with CN, a T1weighted image showing isointensity in the cystic tumor in the region of the septum pellucidum. b Hypointensity on T2-weighted image. c Post-contrast T1-weighted image showing moderate-marked enhancement in the solid component after administration of gadolinium. d Automated MR spectra showing small lactate doublet, small N-acetylaspartate resonance, prominent resonance from choline,
detectable creatine resonance, and high MI and/or Gly signals. e– h A 63-year-old man with EVN in the left temporal lobe. e T1weighted image. f T2-weighted image. g Post-contrast T1-weighted image showing a cortically-based infiltrative lesion with cyst and heterogeneous enhancement. h Automated MR spectra showing definite lactate doublet, small N-acetylaspartate resonance, prominent Cho resonance and detectable Cr resonance
prominent Cho resonance and detectable Cr resonance is an ordinary feature of both CNs and EVNs. Consequently, EVNs show variable imaging features compared with CNs. When a patient presents with a cortical based hemispheric lesion with variable contrast enhancement, cystic component, calcifications, remarkable Cho resonance and detectable Cr resonance on MRS, but lacking peritumoral edema, EVNs should thus be included in the differential diagnosis. Histopathologically, typical CNs and EVNs were composed predominantly of uniform small round cells with neuronal differentiation and clear cytoplasms embedded in a neuropil that are strongly immunoreactive for Syn, and which are consistent with current reports [7, 9, 13, 21]. The results of the present study showed that light microscopy of CNs and EVNs commonly revealed fibrillary matrices alternating with anuclear areas and focal calcifications. Tumor cells in EVNs exhibited a wider morphological spectrum than those in CNs, and were arranged in sheets, clusters, ribbons, or rosettes, with neuropils dispersed either in broad zones (neuropil islands). Atypical pathological features of CNs and EVNs in this study usually included high mitotic count and/or MVP, and focal necrosis and infiltration, also in accord with Brat et al.’s
study of patients with 35 EVNs [9]. In the present study, most CNs and EVNs with atypical histological features associated with malignancy also showed high MIB-1 indexes. Sharma et al. [22] also found that the MIB-1 monoclonal antibody index of multiple proliferation markers tended to be higher in CNs with mitosis and necrosis. Soylemezoglu et al. [23] found a MIB-1 labeling index [2 % in 39 % of the specimens of CNs which was significantly correlated with microvascular proliferation (p = 0.0006). Overall, these results suggest an association between MIB-1 index and atypical histological features, and that patients with atypia defined as atypical histological features with MIB-1 index [3, are more likely to have poorer clinical outcomes. CNs can be diagnosed based on their typical clinical, neuroradiological, histomorphological and immunohistochemical presentations. In contrast, it is difficult to distinguish between EVNs and other brain neoplasms because of their heterogeneous expressions. Syn has been recognized as the most reliable diagnostic marker for identifying neuronal differentiation in brain tumors [24], and may thus play a vital role in the diagnosis of EVNs, based on a histopathological appearance of monotonous neoplastic cells with round, regular nuclei embedded within a fine
123
526
J Neurooncol (2015) 121:521–529
Table 3 Metabolites observed in the 1H MR spectra of the CNs and EVNs Patient CNs
EVNs
MI/ Gly
Cho
Cr
NAA
Lac
Lip
1
H
H
H
H
–
–
2
H
H
H
H
H
H
3 4
H H
H H
H H
H H
– H
– H
5
–
H
H
–
–
–
Cho Choline, Cr creatine, Gly glycine, Lac lactate, NAA N-acetylaspartate, lip lipid
neuropil background, strongly immunoreactive for Syn [9, 25]. The main differential diagnoses of EVNs include oligodendroglioma with neurocytic differentiation, oligoastrocytoma with neurocytic differentiation, ganglioglioma, pilocytic astrocytoma, ependymoma and dysembryoblastic neuroepithelial tumor (DNT). Radiologically, EVNs demonstrate radiological features that are likely to be contrast-enhancing and/or cystic, depict a circumscribed, rather than a diffusely-infiltrating mass, while oligodendrogliomas and other EVNs-mimicking gliomas are typically more infiltrative than EVNs. Morphologically, ependymomas show more typical rosettes [9] and DNTs show more mucoid areas and are less cellular than EVNs [26]. Immunohistochemically, oligodendrogliomas can show some Syn positivity, but to a lesser degree or extent
Table 4 Cox Regression
Dependent variable
than EVNs [27, 28]. Positivity for NeuN also supports a diagnosis of EVNs rather than oligodendrogliomas or oligoastrocytomas [25] and Olig2 is useful for differentiating between oligodendrogliomas and EVNs [29]. Further, an absence of IDH1 expression is a powerful indicator for distinguishing between EVNs and EVNs-mimicking gliomas [30]. However, ultrastructural studies are sometimes needed to demonstrate unequivocal neuronal features in EVNs [24]. Moreover, twelve cases of CNs showed the expression of GFAP accounting for a big proportion of CNs in our study, and 6 of these cases showed atypical histopathological features and 5 cases recurred during the follow up. Elek et al’ [31] also found a high incidence of GFAPpositivity, suggesting that GFAP positivity might be associated with a malignant course in CNs. Based on the above clinical data for CNs and EVNs, most patients have achieved total resection and had a favorable prognosis, though a few patients required more than one operation. The Mortality was around 6 % in our study, which was the same as that found by Rades et al. [32] in a larger series of neurocytomas patients. The recurrence rates in the present study was 29.4 % for CNs and 64.3 % for EVNs, compared with 21 % in 126 CN patients reported by Vasiljevic et al. [33]. Kane et al. [15] carried out a literature search including case reports and case series comprising 85 EVNs, and found a recurrence rate of 52 %, including 36 % in patients with typical lesions and 68 % in atypical lesions. These results were
CNs
EVNs
HR (95 % CI)
p value
HR (95 % CI)
p value
Male sex
0.75 (0.18, 3.14)
0.696
Age
1.01 (0.94, 1.10)
0.767
Histological atypia
5.03 (1.06, 23.96)
0.042
5.38 (0.99, 29.21)
0.051
Subtotal resection Tumor volume
6.16 (1.04, 36.68)
0.046
5.26 (1.04, 26.66)
0.045
1.01 (0.98, 1.04)
0.580
HR Hazard rate, CI confidence intervals
Radiological therapy
0.33 (0.07, 1.51)
0.154
Table 5 Atypical features of seven CNs cases
Atypical CNs
Case 1
Case 2 Case 3
Case 4
Case 5
Case 6 Case 7
MIB 1 labeling index (%)
5
4
8
7.5
4
10
5
–
3
2
–
3
–
3
Atypical histological features High mitotic count (C3 mitoses/ 10HPFs)
123
Microvascular proliferation (MVP)
?
–
?
–
–
?
?
Necrosis
–
–
–
–
–
–
–
Infiltration of parenchyma
–
–
–
–
–
–
–
J Neurooncol (2015) 121:521–529 Table 6 The location, pathological features and immunohistochemistry data of 14 EVNs No.
Age
Location
Main pathological features and immunohistochemistry data
1
40
R. temporal
Atypical, sheet-like and cell clusters growth with neuropil islands, 3 mitoses/ 10HPFs, MVP, no infiltration, Syn?, NeuN?, GFAP?, Olig2-, IDH1-, MIB1 7.5 %;
2
13
R. parietal occipita
Atypical, sheet-like and cell clusters growth with neuropil islands, 6 mitoses/ 10HPFs, focal infiltration, Syn?, NeuN?, GFAP?, Olig2-, IDH1-, MIB1 30 %;
3
56
Sellar region
Typical, sheet-like growth with neuropil islands, no mitoses/10HPFs, Syn?, NeuN-, NSE?, GFAP-, LH-, FSH-, GH-, ACTH-, PRL-, TSH-, Olig2, IDH1-, MIB1 1.5 %;
4
4
F. ventricle
Atypical, sheet-like and cell clusters growth with neuropil islands, focal ganglion cells, 3 mitoses/10HPFs, MVP, no infiltration, Syn?, NeuN?, GFAP-, Olig2-, IDH1-, MIB1 7.5 %;
5
32
R. parietal
Atypical, sheet-like and cell clusters growth with neuropil islands, 4 mitoses/ 10HPFs, focal ganglionic cells, no infiltration, Syn?, NeuN?, GFAP?, Olig2-, IDH1-, MIB1 7.5 %;
6
43
R. parietaloccipita
Typical, sheet-like and cell clusters growth with neuropil islands, 2 mitoses/ 10HPFs, Syn?, NeuN-, NSE?, GFAP?, Olig2-, IDH1-, MIB1 2.5 %;
7
29
R. frontal
Atypical, cluster-like growth with neuropil islands, 6 mitoses/10HPFs, MVP, necrosis, focal ganglionic cells, Syn?, NeuN?, GFAP-, Olig2-, IDH1-, MIB1 20 %;
8
37
R. temporal
Typical, sheet-like and cell clusters growth with neuropil islands, 2 mitoses/ 10HPFs, focal ganglionic cells, Syn?, NeuN?, GFAP?, Olig2-, IDH1-, MIB1 7.5 %;
9
38
R. frontal
Typical, sheet-like and cell clusters growth with neuropil islands, no mitoses/10HPFs; Syn?, NeuN?, GFAP-, Olig2-, IDH1-, MIB1 5 %;
10
2
R. frontal
Atypical, sheet-like growth with focal rosettes, 5 mitoses/10HPFs, no infiltration, Syn?, NeuN?, GFAP-, Olig2-, IDH1-, MIB1 10 %;
11
25
F. ventricle
Typical, sheet-like and cell clusters growth with neuropil islands, no mitoses/10HPFs, no infiltration, Syn?, NeuN?, GFAP?, Olig2-, IDH1-, MIB1 5 %;
12
41
Cerebellum vermis
Typical, sheet-like growth, 2 mitoses/ 10HPFs; no infiltration, Syn?, NeuN?, GFAP-, Olig2-, IDH1-, MIB1 5 %;
527 Table 6 continued No.
Age
Location
Main pathological features and immunohistochemistry data
13
63
L. temporal
Atypical, sheet-like and cell clusters growth with neuropil islands, 7 mitoses/ 10HPFs, focal ganglionic cells, MVP, necrosis, no infiltration, Syn?, NeuN?, GFAP?, Olig2-, IDH1-, MIB1 20 %;
14
42
R. frontal
Typical, sheet-like growth, no mitoses/ 10HPFs, no infiltration, Syn?, NeuN?, GFAP-, Olig2-, IDH1-, MIB1 5 %;
R Right, L left, MVP microvascular proliferation
similar to those of the present study, and showed a significantly higher recurrence rate for EVNs, especially in aggressively atypical cases. Most patient deaths occurred in the perioperative period, and we therefore chose recurrence as the prognostic indicator to reflect malignant potential. Multivariate analysis revealed a clear superiority of GTR over STR in both CNs and EVNs, in accordance with previous studies [15, 33]. The analysis also showed that atypia was associated with significantly higher rates of recurrence in CNs, but was not obviously in EVNs. Kane et al. [15] found that the recurrence risk was 2–3-fold higher in atypical compared with typical EVNs. Brat et al. [9] demonstrated cellular proliferation and the histological features of atypia formed the basis for aggressive tumor growth and affected the resectability of tumors, thus increasing the risk of STR and leading to a higher recurrence rate. It was noteworthy that atypical lesions accounted for 66.7 % of recurrent cases of EVNs in the current study. However the sample size was too small to draw conclusions about the prognostic value of atypia in EVNs, and further studies are needed to verify these preliminary results. Although radiotherapy was not independently associated with reduced recurrence in CNs, four typical and one atypical CNs with single STR and postoperative radiotherapy did not recur during follow-up periods of 37–61 months (mean 46.8 months) in our series. Several reports have claimed that postoperative radiotherapy for CNs after STR leads to shrinkage of residual tumors, while radiotherapy after GTR remains controversial, given that most patients have longterm tumor control without radiotherapy. Paek et al. [34] found that radiotherapy effectively controlled the residual CNs after surgery and prevented recurrence with a median follow-up periods of 171 months (range 128–229) and 202 months (range165–227), respectively. Hallock’s et al. [35] found a median relapse-free time of 79.3 months for CNs in a long-term study. Kane et al. [15] demonstrated that adjuvant radiotherapy may improve recurrence rates following STR in EVNs. Radiotherapy is thus generally
123
528
J Neurooncol (2015) 121:521–529
Fig. 2 a–i Histopathologic features of EVNs, showing some atypical histological characteristics. Tumor cells were most often found in sheets (a) or in cell clusters with neuropil islands (b). EVNs were all strongly and diffusely immunoreactive for Syn (c). NeuN stain was positive in a nuclei with less frequent architectural patterns of
neurocytic rosettes (d, arrow). Atypia in EVNs including atypical histological features of high mitotic count (C3 mitoses/10 high-power fields) (e, arrow) and microvascular proliferation (f), along with MIB1 labeling index [3 % (g). EVNs were negative for IDH1 (h) and Olig2 (i)
believed to plays a role in the local control of CNs and EVNs, which may have become evident in the current study with a longer follow-up period.
Acknowledgments This work was supported by the project of National Clinical Specialist Facility.
Conclusions In an overall comparison between CNs and EVNs, CNs are easier to diagnose, having typical clinical, radiological, and histological features, while EVNs are more likely to show malignant biological behavior, with atypia and recurrence. STR and atypia are the main predictors of recurrence in both CNs and EVNs. The optimal treatment choice for CNs and EVNs is surgical total resection. Patients with typical and/or completely-resected CNs or EVNs have favorable clinical outcomes. Further studies with more cases are needed to confirm the findings of this study.
123
Conflict of Interest None of the authors has any personal, financial, or professional conflicts of interest to report.
References 1. Hassoun J, Gambarelli D, Grisoli F et al (1982) Central neurocytoma. An electron-microscopic study of two cases. Acta Neuropathol 56(2):151–156 2. Hassoun J, Soylemezoglu F, Gambarelli D, Figarella-Branger D, von Ammon K, Kleihues P (1993) Central neurocytoma: a synopsis of clinical and histological features. Brain Pathol 3(3):297–306 3. Kim DG, Kim JS, Chi JG et al (1996) Central neurocytoma: proliferative potential and biological behavior. J Neurosurg 84(5):742–747 4. Valdueza JM, Westphal M, Vortmeyer A, Muller D, Padberg B, Herrmann HD (1996) Central neurocytoma: clinical, immunohistologic, and biologic findings of a human neuroglial progenitor tumor. Surg Neurol 45(1):49–56
J Neurooncol (2015) 121:521–529 5. Chang KH, Han MH, Kim DG et al (1993) MR appearance of central neurocytoma. Acta Radiol 34(5):520–526 6. Ashkan K, Casey AT, D’Arrigo C, Harkness WF, Thomas DG (2000) Benign central neurocytoma. Cancer 89(5):1111–1120 7. Schmidt MH, Gottfried ON, von Koch CS, Chang SM, McDermott MW (2004) Central neurocytoma: a review. J Neurooncol 66(3):377–384 8. Sharma MC, Deb P, Sharma S, Sarkar C (2006) Neurocytoma: a comprehensive review. Neurosurg Rev 29(4):270–285 discussion 285 9. Brat DJ, Scheithauer BW, Eberhart CG, Burger PC (2001) Extraventricular neurocytomas: pathologic features and clinical outcome. Am J Surg Pathol 25(10):1252–1260 10. Rodriguez FJ, Mota RA, Scheithauer BW et al (2009) Interphase cytogenetics for 1p19q and t(1;19)(q10;p10) may distinguish prognostically relevant subgroups in extraventricular neurocytoma. Brain Pathol 19(4):623–629 11. Louis DN, Ohgaki H, Wiestler OD et al (2007) The 2007 WHO classification of tumours of the central nervous system. Acta Neuropathol 114(2):97–109 12. Brown DM, Karlovits S, Lee LH, Kim K, Rothfus WE, Brown HG (2001) Management of neurocytomas: case report and review of the literature. Am J Clin Oncol 24(3):272–278 13. Giangaspero F, Cenacchi G, Losi L, Cerasoli S, Bisceglia M, Burger PC (1997) Extraventricular neoplasms with neurocytoma features. A clinicopathological study of 11 cases. Am J Surg Pathol 21(2):206–212 14. Rades D, Fehlauer F, Lamszus K et al (2005) Well-differentiated neurocytoma: what is the best available treatment? Neuro Oncol 7(1):77–83 15. Kane AJ, Sughrue ME, Rutkowski MJ et al (2012) Atypia predicting prognosis for intracranial extraventricular neurocytomas. J Neurosurg 116(2):349–354 16. Swinson BM, Friedman WA, Yachnis AT (2006) Pontine atypical neurocytoma: case report. Neurosurgery 58(5):E990 discussion E990 17. Tatter SB, Borges LF, Louis DN (1995) Correction: central neurocytomas of the cervical spinal cord. J Neurosurg 82(4):706 18. Patil AS, Menon G, Easwer HV, Nair S (2013) Extraventricular neurocytoma, a comprehensive review. Acta Neurochir (Wien) 68(4):e206–e212 19. Yi KS, Sohn CH, Yun TJ, et al (2012) MR imaging findings of extraventricular neurocytoma: a series of ten patients confirmed by immunohistochemistry of IDH1 gene mutation. Acta Neurochir (Wien), 154(11):1973–1979; discussion 1980 20. Ueda F, Suzuki M, Matsui O, Uchiyama N (2007) Automated MR spectroscopy of intra- and extraventricular neurocytomas. Magn Reson Med Sci 6(2):75–81
529 21. von Deimling A, Janzer R, Kleihues P, Wiestler OD (1990) Patterns of differentiation in central neurocytoma. An immunohistochemical study of eleven biopsies. Acta Neuropathol 79(5):473–479 22. Sharma MC, Rathore A, Karak AK, Sarkar C (1998) A study of proliferative markers in central neurocytoma. Pathology 30(4):355–359 23. Soylemezoglu F, Scheithauer BW, Esteve J, Kleihues P (1997) Atypical central neurocytoma. J Neuropathol Exp Neurol 56(5):551–556 24. Robbins P, Segal A, Narula S et al (1995) Central neurocytoma. A clinicopathological, immunohistochemical and ultrastructural study of 7 cases. Pathol Res Pract 191(2):100–111 25. Agarwal S, Sharma MC, Sarkar C et al (2011) Extraventricular neurocytomas: a morphological and histogenetic consideration. A study of six cases. Pathology 43(4):327–334 26. Jay V, Edwards V, Hoving E et al (1999) Central neurocytoma: morphological, flow cytometric, polymerase chain reaction, fluorescence in situ hybridization, and karyotypic analyses Case report. J Neurosurg 90(2):348–354 27. Mut M, Guler-Tezel G, Lopes MB, Bilginer B, Ziyal I, Ozcan OE (2005) Challenging diagnosis: oligodendroglioma versus extraventricular neurocytoma. Clin Neuropathol 24(5):225–229 28. Perry A, Scheithauer BW, Macaulay RJ, Raffel C, Roth KA, Kros JM (2002) Oligodendrogliomas with neurocytic differentiation. A report of 4 cases with diagnostic and histogenetic implications. J Neuropathol Exp Neurol 61(11):947–955 29. Okada M, Yano H, Hirose Y et al (2011) Olig2 is useful in the differential diagnosis of oligodendrogliomas and extraventricular neurocytomas. Brain Tumor Pathol 28(2):157–161 30. Capper D, Reuss D, Schittenhelm J et al (2011) Mutation-specific IDH1 antibody differentiates oligodendrogliomas and oligoastrocytomas from other brain tumors with oligodendroglioma-like morphology. Acta Neuropathol 121(2):241–252 31. Elek G, Slowik F, Eross L, Toth S, Szabo Z, Balint K (1999) Central neurocytoma with malignant course. Neuronal and glial differentiation and craniospinal dissemination. Pathol Oncol Res 5(2):155–159 32. Rades D, Schild SE (2006) Treatment recommendations for the various subgroups of neurocytomas. J Neurooncol 77(3):305–309 33. Vasiljevic A, Francois P, Loundou A et al (2012) Prognostic factors in central neurocytomas: a multicenter study of 71 cases. Am J Surg Pathol 36(2):220–227 34. Paek SH, Han JH, Kim JW et al (2008) Long-term outcome of conventional radiation therapy for central neurocytoma. J Neurooncol 90(1):25–30 35. Hallock A, Hamilton B, Ang LC et al (2011) Neurocytomas: longterm experience of a single institution. Neuro Oncol 13(9):943–949
123
