Inflammatory Pseudotumors of the Central
Nervous System: Report of 3 Cases and a
Literature Review
MARTIN HÄUSLER, MD, LARS SCHAADE, MD,
VINCENT T. RAMAEKERS, MD, MARTIN DOENGES, MD,
GERHARD HEIMANN, AND BERND SELLHAUS, MD
Inflammatory pseudotumors (IPs), mostly benign lesions characterized by fibrotic ground tissue and polyclonal mononuclear infiltrate, may affect all organ systems. IPs originating in the central
nervous system (IP-CNS) are very rare, and their distinct histopathologic features are poorly characterized. Three otherwise healthy patients (age 8, 15, and 17 years) presented with focal neurologic
symptoms (seizures, n ⴝ 2; headaches, n ⴝ 1), corresponding to a left
temporal, left occipital, and left frontal IP, respectively, extending
from meningeal structures into brain tissue. After resection, no recurrence was observed in patient 1 during 5 years of follow-up,
whereas patient 2 developed a rapidly progressive local recurrence
and a second intracerebral lesion despite antiviral, immunosuppressive, antibiotic, and radiation therapy. In patient 3, who also showed
local recurrences, sequential histopathologic investigations revealed
transformation to a semimalignant fibrohistiocytic tumor. In this
patient, anaplastic lymphoma kinase (ALK) expression was also positive, whereas it was negative in patient 1. A detailed literature analysis
confirmed that most IP-CNS arise from dural/meningeal structures
(n ⴝ 34). Intraparenchymatous (n ⴝ 7), mixed intraparenchymatous/
meningeal (n ⴝ 4), and intraventricular lesions (n ⴝ 7) or IP extend-

ing per continuitatem from intracerebral to extracerebral sites (n ⴝ
5) were rare. The recurrence rate was 40% within 2 years in general.
It was increased after incomplete resection and in female patients
(multivariate Cox regression model, P < 0.02). Although rare, IPCNS are important differential diagnoses among tumor-like intracranial lesions. Their potential risk of malignant transformation and
high risk of recurrence necessitate close follow-up, especially when
resection is incomplete. Prospective multicenter trials are needed to
optimize classification and treatment of this rare inflammatory lesion.
HUM PATHOL 34:253-262. Copyright 2003, Elsevier Inc. All rights
reserved.
Key words: inflammatory pseudotumor, brain neoplasm, inflammatory disease.
Abbreviations: ALK, anaplastic lymphoma kinase; CNS, central
nervous system; IP, inflammatory pseudotumor; TCR, T-cell receptor; EBV, Epstein-Barr virus; PCR, polymerase chain reaction; CSF,
cerebrospinal fluid MRI, magnetic resonance imaging; EMA, epithelial membrane antigen, SMA, smooth muscle antigen; LCA, leukocyte
common antigen; VIM, vimentin; RR, relative risk.

Inflammatory pseudotumors (IPs), or inflammatory myofibroblastic tumors, represent a group of lesions that are histopathologically characterized by a
collagenous stroma of different vascularization, spindle
cells, and a polyclonal mononuclear infiltrate.1,2 In contrast to the lung, mesentery, omentum, retroperitoneum, genitourinary tract, or upper respiratory tract,1-3
IPs affecting primarily the central nervous system (IPCNS) are very rare, with fewer than 60 reported cases.
According to Tresser et al,4 who reviewed the clinical
features of 26 patients with IP-CNS described as of
1995, nonspecific complaints such as headaches, seizures, pareses, vision disturbance, and ataxia may be
the most common symptoms.4 From their data, these
authors also calculated a recurrence rate of 12.5%
when surgical resection was not followed by radiation
therapy.4
Histopathologically, IPs show heterogeneous features, which explains the variety of synonymous expressions, including fibrous xanthoma, xanthomatous

pseudotumor, plasma cell granuloma, plasma cell/histiocytoma complex, pseudosarcomatous fibromixoid
tumors, and inflammatory myofibrohistiocytic proliferation.5 Nevertheless, major subgroups have been identified among IPs not affecting the CNS (IP-nCNS),
including a predominant myxoid/vascular pattern resembling granulation tissue, a compact spindle cellular
and a hypocellular fibrotic pattern in nonpulmonary
IP-nCNS, and an organizing pneumonia pattern with
central hyalinization, a fibrous histiocytoma-like pattern, and a lymphoplasmacytic pattern in pulmonary
IP-nCNS.1,2 Although basically considered benign lesions, IP-nCNS may invade vessels1,6 or, very rarely, even
undergo malign transformation.1,7 In 2 pulmonary
cases, for example, transformation to malignant fibrous
histiocytomas was reported.7 For IP-CNS, in contrast, a
similar classification is not yet available.
In this report we focus on histopathologic characteristics of IP-CNS, on their intracranial growth patterns, and on the clinical outcomes. With reference to
3 new cases and a detailed literature analysis, we suggest
that although it shares characteristics with IP-nCNS,
IP-CNS may be a distinct disease entity that carries a
high risk of local recurrence in cases of incomplete
resection, and may undergo malign transformation.

From the Department of Pediatrics; Institute of Medical Microbiology, Division of Virology; Department of Neuroradiology, and
Institute of Neuropathology, University Hospital, RWTH Aachen,
Germany. Accepted for publication November 21, 2002.
Address correspondence and reprint requests to M. Häusler,
Department of Pediatrics, University Hospital, RWTH Aachen, Pauwelsstr. 30, 52074 Aachen, Germany.
Copyright 2003, Elsevier Inc. All rights reserved.
0046-8177/03/3403-0009$30.00/0
doi:10.1053/hupa.2003.35

PATIENTS AND METHODS
Detailed clinical records were available for the 3 patients
with IP-CNS who had been treated at the University Hospital

253

HUMAN PATHOLOGY

Volume 34, No. 3 (March 2003)

RWTH Aachen between 1996 and 2001. One patient was
suffering from an additional pulmonary IP. To begin, 7-␮mthick sections from formalin-fixed and paraffin-embedded
tissue specimens were stained for hematoxylin and eosin and
Elastica-van Gieson according to routine methods. Immunohistochemistry was also performed according to routine
methods with peroxidase/DAB or alkaline phosphatase/
RED. Antibodies were directed against leukocyte common
antigen (LCA), B cells (CD20), T cells (CD45RO), kappa and
lambda immunoglobulin chains, epithelial membrane antigen (EMA), vimentin (VIM), desmin, smooth muscle antigen
(SMA), S-100 antigen, Ki67/MIBI, and anaplastic lymphoma
kinase (ALK). All antibodies were purchased from Dako
(Hamburg, Germany). Polymerase chain reaction (PCR) Tcell receptor (TCR) investigation for ␥ chain rearrangement
was performed as reported previously.8
PCR investigation for the presence of viral genome was
performed on cerebrospinal fluid (CSF) (patient 2) and paraffin-embedded brain (patients 1 and 3) or pulmonary tissue
(patient 1). Nucleic acid was extracted from CSF using the
QIAgen Blood Kit (Qiagen, Hilden, Germany) and from paraffin-embedded tissues with the QIAgen Tissue Kit. PCR for detection of Epstein-Barr virus (EBV) and human herpes virus type
6 and 8 DNA was performed according to previously published
protocols.9 PCR for detection of herpes simplex virus type 1 and
2 DNA (target, polymerase gene) and varicella zoster virus DNA
(target, open reading frame [ORF] 29) was performed on a
LightCycler instrument (Roche Biochemicals, Mannheim, Germany) applying the fluorescence resonance energy transfer
technology with 2 fluorophore-labeled hybridization probes.
Protocols have repeatedly passed national and international external quality control cycles. The detection limit was 1000 to
2000 genomes /mL of CSF or 200 to 400 genomes in brain
tissue. Detection of JC virus DNA from CSF was performed by
the Institute of Virology and Immunobiology, University of
Würzburg a national reference laboratory for polyoma viruses.
Evaluation of intrathecal synthesis of antibodies to neurotropic
viruses was performed as described previously.10
The literature review was based on a PUB MED and WEB
OF SCIENCE search for the key words “inflammatory pseudotumor”, “plasma cell granuloma,” “(inflammatory) myofibroblastic tumor,” and further references cited in previously published articles. Reports on orbital pseudotumors extending into
the cranial cavity were excluded, because they represent a welldescribed, distinct entity.11 Publications on isolated pachymeningitis, lymphocyte-rich meningiomas, and cases positive for
EMA expression indicative of meningiomas were also excluded.
The relationship between distinct patient characteristics
and tumor recurrence was analyzed using the Cox regression
model.12 The variables showing a significant relationship with
recurrence on univariate analysis were included in a multivariate forward stepwise analysis to determine factors independently related to recurrence (probability of stepwise entry
or removal, 0.05 and 0.1, respectively). Kaplan-Meier plots
were constructed for the recurrence rate in general and for
variables related with recurrence (log-rank statistics; P ⬍ 0.05
considered significant). The Statistical Package for Social Sciences software was used for data analysis.

RESULTS
Patient Reports

FIGURE 1. T2-weighted MRI image of patient 1, performed 4
years after resection of the temporal lesion, showing 2 subcortical residual signal intense lesions.

he suffered a right-sided seizure due to a round, calcified non– contrast-enhanced lesion within the left temporal lobe. During surgery, a solid, not well-delineated
tumor infiltrating the surrounding brain tissue was
found and resected incompletely. After surgery, no
radiation or chemotherapy was performed. Postsurgical
magnetic resonance imaging (MRI) revealed a small
residual left temporal lesion and 2 left subcortical intraparenchymatous contrast-enhanced lesions without
surrounding edema. Because of the neurotopography
of the subcortical lesions and the good postsurgical
seizure control, these lesions were not resected. During
5 years of follow-up, the brain lesions did not grow, the
contrast medium uptake disappeared, and no pulmonary recurrence occurred (Fig 1). Clinically, the patient
shows an attention deficiency disorder and a disturbed
short-term memory. Laboratory tests for autoimmune
and immunodeficiency disorders proved normal.
Patient 2

Patient 1

In this boy, a left thoracic IP had necessitated
left-sided pneumonectomy at age 2 years. Six years later

This 15-year-old male patient presented with a
right-sided epileptic seizure, corresponding to a contrast-enhanced left occipital lesion (Fig 2A). During
surgery, this lesion was found to infiltrate the brain, and

254

INFLAMMATORY PSEUDOTUMORS OF THE CNS (Häusler et al)

FIGURE 2. Patient 2. (A) T1-weighted, contrast-enhanced MRI image performed before surgery ( T1-weighted), showing a large
contrast-enhanced lesion with central necrosis and perifocal edema. (B), T1-weighted, contrast-enhanced MRI taken 6 months
later showing extensive progressive brain destruction with an additional right hemispheric intracerebral lesion.

an incomplete resection was performed. Postsurgically,
thoracic imaging, bone marrow, and microbiologic investigations were normal. CSF analysis revealed a selective intrathecal synthesis of IgG antibodies to herpes
simplex virus but not to further neurotropic viruses.
After surgery, the patient suffered persistent right-sided
hemianopsia without further neurologic or cognitive
deficits, and no additional treatment was initiated. A
control brain MRI scan taken 2-1/2 months later
showed a residual cystic structure within the former
resection area, surrounded by a new contrast-enhanced
inflammatory process now involving 1/3 of the left
hemisphere and showing perifocal edema. Spinal lesions were absent. At that time, no more intrathecal
anti– herpes simplex virus antibody synthesis was
present. During the subsequent 2 months, progressive
aphasia with recurrent focal seizures occurred, rapidly
followed by right-sided hemiparesis and a stuporic
state. Simultaneously the lesion showed local progression that did not respond to steroid treatment, longterm acyclovir therapy, a 2-week virostatic course of
zidovudine, cidofovin, and ribavirin as well as antibiotic
treatment empirically performed in the course of steroid-related infectious complications (cefotiam, metronidazole, ciprofloxacin, and clindamycin). Six months

after the diagnosis, after local radiation with a total of
20 Gy in daily 2-Gy fractions, a new right-hemispheric
lesion occurred (Fig 2B). At 1-1/2 years after manifestation, the patient showed spastic tetraparesis and unimproved mental status. Cranial computed tomography
revealed extensive residual lesions with global brain
atrophy, widening of inner CSF spaces, left-occipital
porencephaly, and extensive calcifications within the
areas of former inflammation. Repeated thoracic and
abdominal imaging revealed normal findings, as did
blood analysis for malign hematologic or autoimmune
diseases.
Patient 3

This 17-year-old female presented with left frontal
headaches. A brain MRI scan revealed a large (6 ⫻ 6 ⫻
3.5 cm) left frontal lesion with perifocal edema, homogeneous gadolinium uptake, and an angiographically
verified blood supply via the frontal meningeal and the
superficial temporal arteries. At surgery, the lesion was
found to be of fibrous consistency, attached to the
hypervascularized dura, and extending into the white
matter. According to macroscopic criteria, a complete
resection was performed. Two years after primary sur-

255

HUMAN PATHOLOGY

Volume 34, No. 3 (March 2003)

gery, an asymptomatic 1 ⫻ 1 cm left-frontal intracerebral lesion adjacent to the former resection cavity was
detected. Again, surgical resection was performed.
Three years after the second operation, a new 1 ⫻ 1 cm
left frontal local recurrence occurred that also necessitated resection. Repeated thoracic imaging ruled out
pulmonary lesions.
Histopathologic Findings
Patient 1

The circumscribed pulmonary lesion represented
a highly vascularized inflammatory pseudotumor composed of spindle cells with collagenous stroma and a
dense lymphoplasmacellular infiltrate. The spindle
cells were arranged in a parallel fashion showing regular and normochromic nuclei (MIB-1 index, ⬍1%).
Immunohistochemically, they coexpressed VIM, SMA,
and actin but were negative for desmin and ALK. The
inflammatory cells were of polyclonal origin (T, B, and
plasma cells) and scattered throughout the lesion, occasionally forming aggregates or showing perivascular
accumulation. In some areas, lymphocytes were suspected to transmigrate small vessel walls and intravascular IP tissue was detected within the lesion (Fig 3A).
The superficial components of the brain mass consisted
of thickened meninges with an accumulation of fibroblasts producing collagen fibers and a mononuclear
infiltrate (B, T, and plasma cells), but no calcifications
or germinal centers. Focally, mononuclear cells extended from the meninges along Virchow-Robin spaces
into deeper brain structures, but without invading surrounding brain tissue (Fig 3B). At other sites, the superficial cortical brain structures adjacent to the inflammatory areas were penetrated by fibrotic septs, also
arising from the meningeal lesion. Focally, the glioneuronal tissue adjacent to these septs contained psammomatous bodies, whereas inflammatory cells were not
present (Fig 3C). Some of the small vessel walls showed
circular fibrotic thickening. The MIB-1 index of the
fibroblast population was ⬍1%. The cerebral fibroblasts were negative for ALK expression and, in contrast
to the pulmonary lesion, did not express actin or SMA
either.
Patient 2

Histopathology revealed thickened meninges due
to multiple layers of collagenous bundles. Lymphoplasmocellular infiltrates were prominent, especially in inner meningeal layers and surrounding blood vessels.
The superficial cortical structures were destroyed by
extensive inflammation consisting of B and T lymphocytes, plasma cells, and macrophages. Germinal centers
were not observed (Fig 4). The inflammatory cells extended within Virchow-Robin spaces from the inflamed
meningeal structures into the adjacent cortex, from
where lymphocytes diffusely infiltrated and destroyed
surrounding brain tissue. Fibrotic thickening with a
decreased intraluminal diameter was found in small
meningeal veins. The MIB-1 index of the fibroblasts was

FIGURE 3. Patient 1, all hematoxylin and eosin. (A) Pulmonary
venous vessel located within the IP lesion, showing intraluminal
IP tissue (arrow). (B) Glioneuronal tissue penetrated by fibrovascular septs, containing a mononuclear infiltrate. The inflammatory cells are prominent around vessels but do not infiltrate
brain tissue. (C) Numerous psammomatous bodies focally
present in glioneuronal tissue adjacent to fibrotic septs.

256

INFLAMMATORY PSEUDOTUMORS OF THE CNS (Häusler et al)

FIGURE 4. Patient 2, all hematoxylin and eosin. (A) Extensively thickened and fibrotic meninges containing a marked mononuclear infiltrate(*) that is most prominent adjacent to the brain surface. The infiltrate extends within Virchow-Robin spaces (arrow)
and diffusely infiltrates brain tissue. Focal intraparenchymatous hemorrhages are present. (B) Extensive perivascular mononuclear
inflammation extending into surrounding brain tissue.

FIGURE 5. Patient 3. (A) Primary lesion. IP lesion consisting of
fibrotic (*), xanthomatous (arrowhead), and markedly inflamed areas. (B) First recurrence. Occurrence of multinucleated Touton giant cells (arrow). (C) Second recurrence. Aggressive replacement of brain tissue (*) by fibrohistiocytic cells
(all hematoxylin and eosin).

257

HUMAN PATHOLOGY

Volume 34, No. 3 (March 2003)

⬍1%. PCR investigation of TCR-gamma chains revealed no monoclonal T cell proliferation.
Patient 3

The findings in the first lesion (Fig 5A) were consistent with the diagnosis of an IP-CNS, again showing
thickened fibrotic meninges and fibrohistiocytic areas
admixed with infiltrated cerebral cortex. Scattered or
focally aggregated mononuclear cells (B and T cells,
plasma cells, and macrophages) and xanthomatous
cells were found throughout the lesion, with the density
highest within the meninges. Anaplastic elements or
mitoses were absent, as were germinal centers and
meningothelial differentiation of spindle cells. Few
multinucleated cells were seen. Emperipolesis was absent. The first recurrence (Fig 5B) showed similar tissue
components, but also contained focal accumulations of
multinucleated cells with centrally arranged nuclei, resembling Touton giant cells. The nuclei of the fibrohistiocytic cells were more prominent, and mitoses were
also found (MIB index, 1% to 3%). The adjacent cortical tissue was severely infiltrated and replaced by the
tumor. Xanthomatous cells were abundant and indicative of tissue absorption. A fibrohistiocytic tumor with a
high Touton giant cell component was suspected. The
second recurrence (Fig 5C) still showed a marked polyclonal mononuclear infiltrate. The fibrohistiocytic tissue component (CD68 positive) had increased, however. It showed an increasing number of mitoses (MIB
index, focally 10%) and aggressively destroyed brain
tissue. Pleomorphic and multinucleated giant cells
were abundant. A fibrohistiocytic tumor of low malignancy not otherwise specified (NOS) was now diagnosed. In the course of the disease, the fibrohistiocytic
component became increasingly positive for VIM and
SMA, whereas ALK expression was present from disease
onset. Staining for S-100 antigen was positive in reactive
astroglial cells but negative in the fibrohistiocytes.
In all patients, stains for mycobacterial, bacterial,
and fungal diseases proved negative. No morphological
signs of viral diseases, such as inclusion bodies, were
found. Immunohistochemistry ruled out EMA expression on spindle cells. The staining for kappa and
lambda chains revealed polyclonal B cell infiltrates.
PCR Investigations
PCR investigations failed to detect viral DNA from
EBV, herpes simplex virus, or type 6 and 8 human
herpes viruses in paraffin-embedded brain and pulmonary tissue of patient 1 or in brain tissue of patients 1
and 3. In patient 2, CSF PCR proved negative for mycobacteria, herpes simplex virus, varicella zoster virus,
JC virus, EBV, and type 6 and type 8 human herpesviruses.
Review of the Literature
A list of the works reviewed is available from the
authors on request.

Intracranial Localization of the Brain Lesions

According to the anatomical distribution, 5 groups
of lesions could be identified: (A) intraparenchymatous
lesions (n ⫽ 7); (B) meningeal lesions (n ⫽ 34; including 7 with hypophyseal/hypothalamic dysfunction due
to compression of diencephalic structures and 13 with
infiltration of brain tissue); (C) mixed intraparenchymatous and meningeal lesions (n ⫽ 4); (D) intraventricular lesions (n ⫽ 7, in 5 presumably deriving from
the choroid plexus); and (E) IP continuously extending between the cranial cavity and the sphenoidal sinus
(n ⫽ 2), right temporal muscle (n ⫽ 1), pterygopalatinate fossa (n ⫽ 1), or right nasal cavity (n ⫽ 1).
Additional pulmonary IPs were present in 6 cases, presenting either before (n ⫽ 3) or simultaneously with
(n ⫽ 3) the CNS manifestations.
Histopathologic Findings

The most frequent ground tissue pattern reported
in the literature resembled the compact spindle cellular or hypocellular fibrotic pattern of nonpulmonary
IP-nCNS, containing a mononuclear infiltrate dominated by plasma cells 1 (Table 1). In 8 other patients, a
hyalinized fibrous ground tissue pattern with numerous
plasma cells was found, reminiscent of the organizing
pneumonia-type pattern of pulmonary IP2 A myxoid
ground tissue was reported in only 1 case. None of the
further histopathologic features, such as the presence
of calcifications or of giant cells, was restricted to a
single entity. The immunohistochemical data of 9 previously published and the 3 present cases are summarized in Table 2.
Therapy and Outcome

None of the previously published patients died
from the IP, whereas 3 died from surgical or postsurgical complications. Treatment consisted of biopsy (n ⫽
16), or complete (n ⫽ 28), or partial (n ⫽ 11) resection, in some cases combined with corticosteroid treatment (n ⫽ 11) and/or radiation therapy (n ⫽ 13).
Follow-up was reported for 40 previously published
cases, 11 of them suffered local intracranial recurrences; these were combined with distant intracerebral
recurrences in 2 cases and/or extracerebral recurrences in 3 cases. The extracerebral recurrences affected mediastinal structures (n ⫽ 2), lung (n ⫽ 1),
carotic arteries (n ⫽ 1), and nasal sinus (n ⫽ 1). One
other patient showed an isolated extracerebral subglottic recurrence (Table 1). Kaplan-Meier analysis of these
40 previously published cases and of the present 3
patients suggests that approximately 40% of the patients will suffer recurrences within 2 years of primary
surgery (Fig 6). Only data on the intracerebral distribution of the lesions, patient age, basic ground tissue
components, the quantity of plasma cells, treatment,
and gender were available from nearly all these 43
patients, so that only these variables were included in
the Cox regression analysis. In the univariate model (B,
regression coefficient; SE, standard error; RR, risk ra-

258

INFLAMMATORY PSEUDOTUMORS OF THE CNS (Häusler et al)

TABLE 1. Histopathologic Data of 57 Patients With IP-CNS and of the Present 3 Cases*
Previous cases
Group A
(n ⫽ 7)

Group B
(n ⫽ 34)

4

24
6

Ground tissue
Fibrous
Hyaline
Myxoid
Plasma cell density
High
Medium
Low
Russel bodies
Germinal centers
Granulocytes
Calcifications
Total
Psammoma
Giant cells
Multinucleated
Foreign body
Touton
Necrosis
Perivascular inflammation
Xanthomatous cells
Foamy cells

Group C
(n ⫽ 4)

New cases
Group D
(n ⫽ 7)

Group E
(n ⫽ 5)

5
1

4
1

2
2
3
4
2

2
3

Case 1

Case 2

Case 3
1st res.

Case 3
2nd res.

Case 3
3rd res.

⫹⫹⫹

⫹

⫹⫹⫹

⫹⫹⫹

⫹⫹

⫹

⫹

⫹

⫹

⫹

⫹

⫹⫹

⫹⫹⫹

⫹
⫹⫹⫹
⫹⫹

⫹
⫹
⫹

⫹

1
7
3
2
3

1

4
1

20
10
2
19
7
9

3
1
2

1
2

2

4
2

3
2

1
2
1
4
4
1
3

2

⫹
⫹

1

⫹

3
1

⫹
⫹⫹
⫹

1

⫹

Abbreviation: res., resection.
*Detailed description of the groups discussed in text.

tio; CI, 5% to 95% confidence interval; P, significance
level) incomplete resection (B ⫽ 1.807; P ⫽ 0.019) and
female gender (B ⫽ 1.407; P ⫽ 0.018) were found to be
associated with a higher risk of recurrence. In the
multivariate model, this was confirmed for both incomplete resection (B ⫽ 1.958; SE ⫽ 0.781; RR ⫽ 7.087;
CI ⫽ 1.535 to 32.72; P ⫽ 0.012) and female gender
(B ⫽ 1.598; SE ⫽ 0.615; RR ⫽ 4.945; CI ⫽ 1.483 to
16.493; P ⫽ 0.009) (Fig 6). Seven patients with local
(n ⫽ 6) or extracerebral (n ⫽ 1) recurrence were
treated with radiation (n ⫽ 3), steroids (n ⫽ 2), meth-

otrexate (n ⫽ 1), azathioprine (n ⫽ 1), and/or repeated surgery (n ⫽ 2). Among these, a clinical cure
was reported in 1 patient with resurgery only.
DISCUSSION
IP-CNS represent a diagnostic and clinical challenge. Due to their nonspecific neuroradiologic and
clinical findings,4 the correct diagnosis depends on the
postsurgical histopathologic examinations and on the

TABLE 2. Immunohistochemical Findings in Stromal Fibroblasts

IP-CNS
Figarella et al, 1990
Chang et al, 1991
Sitton et al, 1992
Chan et al, 1994
Al Sarray et al, 1995
Tresser et al, 1996
Bramwit et al, 1997
Diederichs et al, 1998
Hansen et al, 2001
Present study case 1
Present study case 2
Present study case 3
Total
IP-nCNS
Coffin et al
Cook et al

Vimentin

Smooth muscle actin

Desmin

Anaplastic
lymphoma kinase

n.d.
Positive
Negative
Positive
Positive
Positive
Positive
n.d.
Positive
Positive
Positive
Positive
9/10 (90%)

Positive
n.d.
n.d.
n.d.
Positive
Positive
focally positive
Positive
Positive
Negative
n.d.
Positive
7/8 (88%)

Negative
n.d.
n.d.
n.d.
Negative
Positive
Negative
n.d.
Positive
Negative
n.d.
Negative
2/7 (29%)

n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
Negative
n.d.
Positive
1/2 (50%)

112/114 (98%)
n.d.

94/117 (80%)
n.d.

63/117 (54%)
n.d.

n.d.
44/73 (60%)

Abbreviation: n.d., not done.

259

HUMAN PATHOLOGY

Volume 34, No. 3 (March 2003)

FIGURE 6. Increased recurrence rate after incomplete tumor
resection as shown by Kaplan-Meier analysis (n ⫽ 43; including
the present 3 cases).

pathologist’s awareness of this rare disease. The possibility of IP should be considered in all intracranial
tumor-like lesions composed of a lymphoplasmacellular
infiltrate and a collagenous stroma of varying density,
regardless of additional features such as giant cells,
calcifications, or the composition of the lymphocyte
infiltrate. In contrast with IP-nCNS, where immunohistochemistry is of diagnostic impact (Table 2),3 immunohistochemical data are available from only few
patients with IP-CNS, so that its importance remains
uncertain.4,13-19
With respect to differential diagnosis, classical CNS
tumors, such as astrocytomas and primitive neuroectodermal tumors, are easy to identify, whereas further
diseases may be more difficult to distinguish. Lymphoplasma cell-rich meningiomas, for instance, show fibrous ground tissue as well as mononuclear infiltrate,
but the finding of meningeal whorls and positive EMA
staining rules out IP-CNS. A lymphoma should be considered when the inflammatory component dominates,
necessitating further analyses to rule out monoclonality. The intracranial extension of an orbital pseudotumor, which shows similar histopathologic features,
should be ruled out by neuroimaging.11 Sinus histiocytosis with massive lymphadenopathy (Rosai-Dorfman
disease) may show an intracranial manifestation without nodal involvement.20 As in patient 3, in whom it
might have been inferred from the further histopathological findings, it should be precluded by the absence
of S-100 antigen expression on histiocytes and of emperipolesis.20 Finally, as seen from patient 3, who initially showed a typical IP-CNS, conversion to primary
malignant fibrohistiocytic tumors can be ruled out only
during follow-up studies.
This malignant transformation is also relevant to
the hitherto unanswered questions of whether IP are
primarily neoplastic lesions, whether malignant transformation occurs in distinct subentities, or whether
patients with malignant transformation suffer from

completely different diseases, so that for this subgroup
the term “inflammatory” might be misleading and
should not be used any longer. Neoplastic transformation, (e.g., aggressive growth and additional distant
lesions) have also been reported in IP-nCNS.1-3,6,21 In
patient 3, malignant transformation was observed after
5 years of follow-up, a time period that was available for
only 3 of the 57 previously reported patients. Therefore, the risk of transformation may be higher than
suggested by the literature. Also, an increasing number
of studies are describing clonal cytogenetic changes in
stromal fibroblasts,22,23 suggesting a neoplastic nature
in a considerable proportion of lesions. Some 60% of
the inflammatory myofibroblastic tumors not affecting
the CNS may express ALK , a characteristic that is well
known from anaplastic large cell lymphomas, where it is
explained by the t(2;5)(p23;q35) translocation at the
ALK gene locus.24, 25 From disease onset, ALK expression was also recorded in patient 3, whereas it was
absent from pulmonary and CNS tissue of patient 1.
Further studies on larger patient groups are needed to
evaluate whether ALK expression characterizes a more
aggressive subgroup of IP-CNS.
Apart from neoplastic transformation, the simultaneous occurrence of multiple lesions might also be
caused by systemic (auto) immune mechanisms or infections with synchronous manifestations at predisposed sites. This postinfection hypothesis is supported
by the isolation of viral DNA from IP-nCNS lesions
(EBV, human herpesvirus 8)26,27 and the presence of
clinical or laboratory signs of systemic inflammation in
15% to 30% of the previous cases.3 In IP-CNS the
association with viral disease (EBV) has been reported
only once,28 and all of our PCR investigations proved
negative. But these negative findings do not rule out a
viral etiology; the viral load may have been too low,
viruses not investigated might be involved, and viruses
may have triggered inflammatory or even neoplastic
processes before disappearing from brain tissue. Patient 2, for instance, showed a transient intrathecal
synthesis of IgG antibodies to the herpes simplex virus,
which was one reason for antiviral treatment. Because
of the frequent association with hypergammaglobulinemia and autoimmune diseases,3,5,29 autoimmune pathomechanisms have also been discussed in the literature,
but could not be confirmed in the present cases either.
Vascular changes are common in IP-CNS. Circular
fibrosis of small veins, for instance, was found in 6
previously published and in 2 of the present cases.13,19,30,31
Whereas its etiology remains unknown, this fibrosis may
reduce perfusion and enhance tissue destruction. Few
patients with IP-nCNS showed invasion of hepatic or
pulmonary veins by IP tumor tissue, leading to local
thromboembolism.6,32 In our cases, pulmonary venous
invasion by IP tissue was recorded in patient 1, and
hematogeneous propagation from the lung to the brain
may have led to the cerebral manifestation in this patient. Although growth of IP tissue within cerebral vessels has not been described in the literature, the histopathologic findings in patients 1 and 2 suggest
migration of mononuclear cells through vessel walls,

260

INFLAMMATORY PSEUDOTUMORS OF THE CNS (Häusler et al)

and vascular spread might account for the intracranial
spread observed in patient 2. The third pattern of
vascular involvement is characterized by the accumulation of mononuclear cells within Virchow-Robin spaces.
It was recorded in 10 previous case and in all 3 new
cases with IP-CNS.4,13,33-38 On the one hand, this may
mirror a reactive process; macrophages reside in Virchow-Robin spaces and, after contact with foreign antigens, such as viral antigens derived from deeper brain
structures, they might initiate an inflammatory cascade
with attraction of further mononuclear cells.39 On the
other hand, mononuclear cells also might migrate
along Virchow-Robin spaces from inflamed to noninflamed tissue, recruiting an increasing amount of brain
tissue.
Among the clinical and histopathologic characteristics, incomplete surgical resection and, astonishingly,
female gender were found to be related to poorer
outcome (Fig 6). However, the analyzed data were not
collected prospectively, so that the presence of further
prognostic factors cannot be ruled out, and even the
relationship to different surgical approaches and to
gender remains to be confirmed in larger prospective
investigations. However, the high overall risk of recurrence and the deleterious consequences of progressive
brain destruction merit frequent postsurgical brain imaging and most likely a radical surgical approach in the
event of recurrence. Although suggested by single case
reports, neither immunosuppression nor radiation
therapy has hitherto been convincingly shown to be of
prognostic benefit.
IP-CNS comprise a poorly characterized and rare
group of combined fibrosing and inflammatory lesions
of unknown etiology and with unsatisfactory treatment
strategies. Collaborative prospective studies are needed
to improve their histopathologic classification, to clarify
their etiology, and to develop more effective treatment
modalities.
REFERENCES
1. Coffin CM, Watterson J, Priest JR, et al: Extrapulmonary
inflammatory myofibroblastic tumor: A clinicopathologic and immunohistochemical study of 84 cases. Am J Surg Pathol 19:859-872, 1995
2. Matsubara O, Tan-Liu NS, Kenney RM, et al: Inflammatory
pseudotumors of the lung: Progression from organizing pneumonia
to fibrous histiocytoma or to plasma cell granuloma in 32 cases. HUM
PATHOL 19:807-814, 1988
3. Coffin CM, Humphrey PA, Dehner LP: Extrapulmonary inflammatory myofibroblastic tumor: A clinical and pathological survey.
Semin Diagn Pathol 15:85-101, 1998
4. Tresser N, Rolf C, Cohen M: Plasma cell granulomas of the
brain: Pediatric case presentation and review of the literature. Childs
Nerv Syst 12:52-57, 1996
5. Biselli R, Boldrini R, Ferlini C, et al: Myofibroblastic tumours:
Neoplasias with divergent behaviour. Ultrastructural and flow cytometric analysis. Pathol Res Pract 195:619-632, 1999
6. Warter A, Satge D, Roeslin N: Angioinvasive plasma cell granulomas of the lung. Cancer 59:435-443, 1987
7. Spencer H: The pulmonary plasma cell/histiocytoma complex. Histopathology 8:903-916, 1984
8. Diss TC, Watts M, Pan LX, et al: The polymerase chain
reaction in the demonstration of monoclonality in T cell lymphomas.
J Clin Pathol 48:1045-1050, 1995

9. Bai X, Hosler G, Rogers BB, et al: Quantitative polymerase
chain reaction for human herpesvirus diagnosis and measurement of
Epstein-Barr virus burden in posttransplant lymphoproliferative disorder. Clin Chem 43:1843-1849, 1997
10. Felgenhauer K, Reiber H: The diagnostic significance of
antibody specificity indices in multiple scleosis and herpes virusinduced diseases of the nervous system. Clin Investg 70:28-37, 1992
11. Garner A: Pathology of pseudotumours of the orbit: A review. J Clin Pathol 26:639-648, 1973
12. Hosmer DWJ, Lemeshow S, Barnett V, et al (eds): Applied
survival analysis: Regression modeling of time to event data. New
York: Wiley, 1999
13. Figarella-Branger D, Gambarelli D, Perez-Castillo M, et al:
Primary intracerebral plasma cell granuloma: A light, immunocytochemical, and ultrastructural study of one case. Neurosurgery 27:142147, 1990
14. Chan YF, White J, Brash H: Metachronous pulmonary and
cerebral inflammatory pseudotumor in a child. Pediatr Pathol 14:805815, 1994
15. Diederichs L, Kostovic N, Harzheim A, et al: Intracranial
plasma-cell granuloma. Rofo Fortschr Geb Rontgenstr Neuen
Bildgeb Verfahr 168:205-207, 1998
16. Sitton JE, Harkin JC, Gerber MA: Intracranial inflammatory
pseudotumor. Clin Neuropathol 11:36-40, 1992
17. al-Sarraj S, Wasserberg J, Bartlett R, et al: Inflammatory
pseudotumour of the central nervous system: Clinicopathological
study of one case and review of the literature. Br J Neurosurg 9:57-66,
1995
18. Bramwit M, Kalina P, Rustia-Villa M: Inflammatory pseudotumor of the choroid plexus. Am J Neuroradiol 18:1307-1309, 1997
19. Chang Y, Horoupian DS, Lane B, et al: Inflammatory
pseudotumor of the choroid plexus in Sjogren’s disease. Neurosurgery 29:287-290, 1991
20. Andriko JW, Morrison A, Colegial CH, et al: Rosai-Dorfman
disease isolated to the central nervous system: a report of 11 cases.
Mod Pathol 14:172-178, 2001
21. Maier HC, Sommers SC: Recurrent and metastatic pulmonary fibrous histiocytoma/plasma cell granuloma in a child. Cancer
60:1073-1076, 1987
22. Snyder CS, Dell Aquila M, Haghighi P, et al: Clonal changes
in inflammatory pseudotumor of the lung. Cancer 76:1545-1549,
1995
23. Su LD, Atayde-Perez A, Sheldon S, et al: Inflammatory myofibroblastic tumor: Cytogenetic evidence supporting clonal origin.
Mod Pathol 11:364-368, 1998
24. Cook JR, Dehner LP, Collins MH, et al: Anaplastic lymphoma kinase (ALK) expression in the inflammatory myofibroblastic
tumor. Am J Surg Pathol 25:1364-1371, 2001
25. Pulford K, Lamant L, Morris SW, et al: Detection of anaplastic lymphoma kinase (ALK) and nucloeolar protein nucleophosmin
(MPM)-ALK proteins in normal and neonplastic cells with the monoclonal antibody ALK1. Blood 89:1394-1404, 1997
26. Arber DA, Kamel OW, van de Rijn M, et al: Frequent presence of the Epstein-Barr virus in inflammatory pseudotumor. HUM
PATHOL 26:1093-1098, 1995
27. Gomez-Roman JJ, Ocejo-Vinyals G, Sanchez-Velasco P, et al:
Presence of human herpesvirus-8 DNA sequences and overexpression
of human IL-6 and cyclin D1 in inflammatory myofibroblastic tumor
(inflammatory pseudotumor). Lab Invest 80:1121-1126, 2000
28. Fukunaga A, Yoshida K, Otani M, et al: Plasma cell granuloma extending from the extracranial to the intracranial space associated with Epstein-Barr virus infection. Neurol Med Chir (Tokyo)
38:292-296, 1998
29. Coffin CM, Dehner LP: Fibroblastic-myofibroblastic tumors
in children and adolescents: A clinicopathologic study of 108 examples in 103 patients. Pediatr Pathol 11:569-588, 1991
30. Gartman JJ, Jr., Powers SK, Fortune M: Pseudotumor of the
sellar and parasellar areas. Neurosurgery 24:896-901, 1989
31. Le Marc’hadour F, Fransen P, Labat-Moleur F, et al: Intracranial plasma cell granuloma: A report of four cases. Surg Neurol
42:481-488, 1994
32. Broughan TA, Fischer WL, Tuthill RJ: Vascular invasion by
hepatic inflammatory pseudotumor: A clinicopathologic study. Cancer 71:2934-2940, 1993

261

HUMAN PATHOLOGY

Volume 34, No. 3 (March 2003)

33. Makino K, Murakami M, Kitano I, et al: Primary intracranial
plasma-cell granuloma: A case report and review of the literature.
Surg Neurol 43:374-378, 1995
34. Eimoto T, Yanaka M, Kurosawa M, et al: Plasma cell granuloma (inflammatory pseudotumor) of the spinal cord meninges:
Report of a case. Cancer 41:1929-1936, 1978
35. Wilner HI, Vinas FC, Duffy C, et al: Magnetic resonance
imaging of intracranial plasmocytic granuloma. Skull Base Surg
9:265-270, 1999

36. Kudesia S, Shankar SK, Das S, et al: Intracranial plasma cell
granuloma: A report of four cases. Neuro India 46:142-146, 1998
37. Ferrer I, Garcia Bach M, Aparicio MA, et al: Plasma cell
granuloma of the hypothalamic region. Acta Neurochir (Wien) 99:
152-156, 1989
38. Breidahl WH, Robbins PD, Ives FJ, et al: Intracranial plasma
cell granuloma. Neuroradiology 38 (Suppl 1):S86-S89, 1996
39. Esiri MM, Gay D: Immunological and neuropathological
significance of the Virchow-Robin spaces. J Neurol Sci 100:3-8, 1990

262