вход по аккаунту



код для вставкиСкачать
Pleomorphic Xanthoastrocytoma
What Do We Really Know about It?
Caterina Giannini, M.D., Ph.D.1
Bernd W. Scheithauer, M.D.1
Peter C. Burger, M.D.2
Daniel J. Brat, M.D., Ph.D.2
Peter C. Wollan, Ph.D.3
Bolek Lach, M.D.4
Brian P. O’Neill, M.D.5
Department of Pathology, Mayo Clinic, Rochester, Minnesota.
Department of Pathology, Johns Hopkins Hospital, Baltimore, Maryland.
Department of Biostatistics, Mayo Clinic, Rochester, Minnesota.
Department of Pathology, Ottawa Civic Hospital,
Ottawa, Canada.
Department of Neurology, Mayo Clinic, Rochester, Minnesota.
Presented in part at the 73rd annual meeting of the
American Association of Neuropathologists, Pittsburgh, Pennsylvania, June 11–16, 1997.
Address for reprints: Bernd W. Scheithauer, M.D.,
Department of Pathology and Laboratory Medicine,
Mayo Clinic, 200 1st Street S.W., Rochester, MN
Received July 30, 1998; revision received December 3, 1998; accepted December 23, 1998.
© 1999 American Cancer Society
BACKGROUND. Pleomorphic xanthoastrocytomas (PXA) may recur and demonstrate aggressive clinical behavior with a mortality rate between 15% and 20%. To
the authors’ knowledge, no histopathologic features currently are known to reliably
predict recurrence or tumor progression.
METHODS. The study was based on 71 cases with available information regarding
clinical and therapeutic data and follow-up. Diagnostic features included cellular
pleomorphism, giant and/or xanthic cells, eosinophilic granular bodies, desmoplasia, and leptomeningeal involvement. The mitotic index (MI), the presence of
necrosis, and endothelial proliferation were recorded in all primary resection
RESULTS. The study included 35 females and 36 males, age 26 6 16 years (mean 6
standard deviation). Approximately 98% of tumors were supratentorial, with 49%
in the temporal lobe. Seizures were the presenting symptoms in 71% of patients.
Extent of tumor removal was macroscopic total resection in 68% of cases and
subtotal resection (STR) in 32% of cases. Postoperative radiotherapy, alone or with
chemotherapy, was administered in 29% and 12.5% of cases, respectively. The
recurrence free survival rates (RFS) were 72% at 5 years and 61% at 10 years,
whereas overall survivals rates (OS) were 81% at 5 years and 70% at 10 years. In
univariate analysis, the extent of resection was the single factor associated most
strongly with RFS (P 5 0.003), followed by MI (P 5 0.007) and atypical mitoses (P 5
0.04). Necrosis was not found to be significant. The extent of resection and MI were
confirmed as independent predictors of RFS by multivariate analysis. MI (P 5
0.001), atypical mitoses (P 5 0.02), and necrosis (P 5 0.04) were associated with OS
by univariate analysis. In multivariate analysis, only MI was an independent
predictor of survival. Information regarding MIB-1 labeling index and the use of
adjuvant therapy was too limited to explore their prognostic significance confidently.
CONCLUSIONS. The study confirms that PXA is an astrocytic tumor with a relatively
favorable prognosis. MI and extent of resection appear to be the main predictors
of RFS and OS. Given the slow growth of the tumor, more studied cases and longer
periods of follow-up will be essential to confirm our findings regarding prognostic
factors affecting this unusual tumor. Cancer 1999;85:2033– 45.
© 1999 American Cancer Society.
KEYWORDS: glioma, pleomorphic xanthoastrocytoma, natural history, pathology,
prognostic factors, neurooncology.
leomorphic xanthoastrocytoma (PXA), a new addition to the recent 1993 World Health Organization (WHO) classification of tumors of the central nervous system, was first described by Kepes and
coauthors in 1979.1,2 This morphologically distinctive cerebral glioma
occurred primarily in young subjects, showed a predilection to su-
CANCER May 1, 1999 / Volume 85 / Number 9
perficial growth, and had a relatively favorable prognosis. Prior to that time, such tumors were often classified among giant cell glioblastomas given their
ominous histologic features, which include marked
cellular pleomorphism, nuclear atypia, and the presence of bizarre, multinucleate giant cells. Due to the
rich reticulin network that characterizes superficial
portions of PXA, some had initially been considered
mesenchymal tumors and were designated fibrous
xanthomas or xanthosarcomas by the same author.3
Finally, some had been included in the now defunct
category of “monstrocellular sarcoma.” Unlike giant
cell glioblastomas and sarcomas, however, PXAs are
associated with much longer recurrence free and overall survivals.
Since its initial description, over 100 PXAs have
been reported, most as single cases or small series.4 – 67
Consequently, data regarding its natural history and
prognostic factors are fragmentary. Nonetheless studies generally confirm the clinical and pathologic observations of the original series. It is of note that PXA
may recur and that 15–20% undergo progressive anaplastic transformation. Such tumors are less pleomorphic and are more diffusely infiltrative.62,68 The 1993
WHO classification of brain tumors recognizes the
occurrence of Grade 3 (anaplastic) transformation in
PXAs but does not provide minimal criteria permitting
their identification.1 Indeed, a careful review of published cases leaves the reader without a definition
permitting prospective identification of “anaplastic
PXA,” an aggressive variant prone to recurrence
and/or a fatal outcome. Furthermore, although there
is agreement that surgical resection is the treatment of
choice, the role of adjuvant therapies is presently unclear.3
Herein, we studied a large series of patients with
PXA in order to better define its clinicopathologic features and natural history. Our findings were compared
with an analysis of previously reported, adequately
documented cases. Having evaluated clinical data at
initial presentation and the various pathologic features generally thought to signify aggressive behavior,
we found that mitotic index IMI) was the main histologic predictor of recurrence and survival. Use of the
term “PXA with anaplastic features” is recommended,
followed by a clear statement of just which features,
such as increased mitotic activity, necrosis, or other
indicate correlating with potentially more aggressive
clinical behavior were found. We purposely chose not
to use the term “anaplastic (Grade 3I) PXA,” because
that designation may provoke inappropriately aggressive treatment. Our results, although based upon the
largest series of patients presently available, might still
not prove to be definitive. More cases and longer
follow-up will be necessary to confirm or refute these
Patient Identification
The 71 cases studied were derived from the database
of the Mayo Clinic tissue registry and from the consultation files of two of the authors (B.W.S., P.C.B.).
Their inclusion was based on availability of original
hematoxylin and eosin (H&E)-stained sections and/or
representative paraffin embedded tissue obtained at
primary surgery. Histologic features upon which the
diagnosis of PXA was made included the original criteria of Kepes et al. as well as those of the 1993 WHO
classification of tumors of the nervous system.1,2 Thus,
PXA is a moderately cellular astrocytoma composed of
cells of varying size and shape, giant and lipidized
examples being characteristic. Often superficial in location and partly occupying the leptomeninges, PXA
invariably shows some infiltration of underlying brain
parenchyma and exhibits a tendency to perivascular
(Virchow-Robin) space involvement. Reticulin stains
demonstrate a meshwork surrounding individual or
clusters of tumor cells, especially in superficial and
leptomeningeal portions of the tumor. Small numbers
of lymphocytes and plasma cells are usually present.
Glial fibrillary acidic protein (GFAP) immunoreactivity
is seen within the cytoplasm of greatly varying numbers of tumor cells. Although it was not stressed in the
original description of PXA or in the WHO monograph,
granular bodies of varying size, texture, and eosinophilia are a regular feature of this tumor.69
Clinical Features
Data recorded included patient age and gender, presenting symptoms e.g., seizure or increased intracranial pressure duration of symptoms, tumor location,
neuroimaging appearance (e.g., presence of a cyst),
and mode of treatment, including type and extent of
surgery as well as adjuvant radiotherapy and/or chemotherapy.
Pathologic Studies
In all cases, routinely processed tissues included original and/or newly prepared 5 mm, H&E-stained sections. All were separately examined by two neuropathologists (C.G., B.W.S.), the diagnosis of PXA being
confirmed independently. Specifically recorded were
the presence of: cellular pleomorphism with multinucleate giant cells, spindle cells, small cells, and vacuolated (xanthic) cells; granular bodies, whether coarse
and intensely eosinophilic or delicate and pale;
Natural History and Prognosis of PXA/Giannini et al.
Pleomorphic Xanthoastrocytoma: Clinical Summary
Current series
Total number
Male:female ratio
Presenting symptoms
Seizures (%)
Seizure duration (yrs)2
Location (%)
.1 lobe, including temporal
Frontal, parietal, occipital
Cystic (%)
Follow-up (yrs)2
26 6 16
20 6 13
Literature review
7.6 6 8.5
5 6 4.8
6.2 6 8.4
5.7 6 5.4
Patient age distribution at initial diagnosis. The tumor affects
predominantly children and young adults, 62% of whom are less than age 25
years. The distribution of previously reported cases is quite similar.
No. indicates the number of patients for whom information was available when this is different from
the total number.
Data represent mean 6 standard deviation.
Rosenthal fibers; reticulin pattern and desmoplasia;
calcification; inflammatory cell aggregates; growth
pattern, including leptomeningeal extension and parenchymal infiltration; mitotic activity and MI; endothelial proliferation; and necrosis with or without
pseudopalisading. The MI was defined as the maximum number of mitotic figures present in 10 high
power fields (HPF) at 3400 magnification. All counts
were performed meticulously by one of us (C.G.), with
particular efforts made to distinguish karyorrhectic
debris from mitotic figures. Also recorded was the
number of HPF examined before encountering the
first mitosis. At high magnification (3400), a single
microscopic field of the Olympus BH-2 microscope
(Olympus, Tokyo, Japan) fitted with a D plan 40 (0.65,
160/0.17) objective represented an area of 0.12 mm2.
Immunohistochemical staining for GFAP, in most
cases performed at the time of the original diagnosis,
was repeated only in the cases wherein paraffin blocks
and/or unstained slides were available. A detailed
study of the immunophenotype of PXA is the subject
of a separate publication.70
In 29 cases, immunostains for MIB-1 antigen were
performed on 6 mm paraffin sections mounted on
silanized slides using MIB-1 antisera (Immunotech,
Westbrook, ME; monoclonal, dilution 1:100). To reduce variation, all specimens were stained simultaneously in a single batch, the procedure having been
described in detail previously.71 MIB-1 labeling indices (MIB-1 LI) were expressed as percentage of nuclear area stained in areas of maximum labeling.
Statistical Analysis
Recurrence free survival and overall survival were calculated from the date of first surgery to the date of
second surgery or to death or last follow-up, respectively. Recurrence free survival and overall survival
distributions, formulated by using the Kaplan–Meier
method, were examined for equality with log-rank
tests. To assess the strength of association between
recurrence free/overall survival and the potential
prognostic factors studied, Cox proportional hazards
modeling was used both individually (univariate
model) and jointly (multivariate model). The variables
examined included patient age and gender; tumor
site; presence or absence of a cyst; presenting symptoms (seizure vs. other); extent of tumor resection
(gross or subtotal resection [GTR] vs. STR); presence
or absence of parenchymal infiltration, a small cell
component, mitoses, and necrosis; MI 0, ,5, or $5
mitoses per 10 HPF; presence or absence of atypical
mitoses; number of HPFs examined before encountering the first mitotic figure; and MIB-1 LI.
Literature Review
The 121 cases subject to analysis were derived from a
review of cases reported to May 1997. For comparison
purposes, these data are presented side by side with
those derived from the study of our patient population. The data are incomplete in some instances, and
the actual prevalence of a given feature is presented in
association with the number of cases in which the
information was available (denoted in parentheses).
CANCER May 1, 1999 / Volume 85 / Number 9
FIGURE 2. Histologic features of pleomorphic xanthoastrocytoma (PXA).
Typically superficial in location, it variably involves leptomeninges, infiltrates
underlying parenchyma, and extends
into perivascular spaces (A). Cellular
pleomorphism is the rule and includes
spindle cells (B) as well as mononucleated and multinucleated giant cells, occasionally with vacuolated cytoplasm
(C). Granular bodies, intensely eosinophilic or pale, are an almost constant
finding (D). Reticulin staining surrounds
individual or clustered tumor cells (E);
glial fibrillary acidic protein staining varies in extent but is often intense (F). In
some cases, mitotic figures and/or necrosis are present (G,H). D and H are
reproduced with permission from Giannini and Scheithaur.69
Clinical Findings
Clinical data regarding our study population are summarized in Table 1. The 71 patients included 35 females and 36 males. Their ages varied considerably
(mean 6 standard deviation [SD], 26 6 16 years; median, 22 years; range, 5– 82 years). Figure 1 demonstrates age distributions at diagnosis compared with
that of previously reported cases.
Presenting symptoms were known in 56 patients
(79%) and consisted primarily of seizures (71%), often
of long duration (mean 6 SD, 7.6 6 9.2 years; median,
3 years). The tumors were supratentorial in 98% of the
cases, one cerebellar example having been reported
previously.46 The most common single location was
the temporal lobe (49%), followed by the parietal
(17%), frontal (10%), and occipital (7%) lobes. Of the 9
of 71 tumors that involved more than one lobe, the
contiguous temporal lobe was affected in 7 (10% of
total). Of the 56 cases in which neuroimaging data
were available, 48% of the tumors had a cystic component.
Natural History and Prognosis of PXA/Giannini et al.
Pleomorphic Xanthoastrocytoma: Frequency of Histologic Features
Current series
Multinucleated giant cells
Xanthic cells
Granular bodies
Rosenthal fibers
Nuclear inclusions
Lymphocyte collections
GFAP reactivity
GFAP: glial fibrillary acidic protein.
No. indicates the number of patients for whom information was available when this is different from
the total number.
Pleomorophic Xanthoastrocytoma: Histologic Findings at Initial
Current series
,5 3 10 HPF
$5 3 10 HPF
Endothelial proliferation
MIB-1 labeling index
Literature review
HPF: high-power fields.
No. indicates the number of patients for whom information was available when this is different from
the total number.
Histologic Findings
The characteristic morphologic features of PXA are
illustrated in Figure 2. The frequency with which these
were identified in the present series is illustrated in
Tables 2 and 3.
Despite meticulous examination of all available
tissues, 26 tumors (37%) revealed no mitotic figures.
Low level mitotic activity (,5 mitoses per 10 HPF) was
identified in 32 tumors (45%), more than half having
only 1 per 10 HPF. Levels of $5 mitoses per 10 HPF
were noted in 13 cases (18%), the index being distributed uniformly between 5 per 10 HPF and 10 per 10
Of the 45 tumors that exhibited mitoses, the first
mitotic figure was found in the first HPF examined in
9 cases (20%); in 22 cases (49%), it was found in the
first 25 HPF; in 24 cases (53%), it was found in the first
50 HPF; in 32 cases (71%), it was found in the first 100
HPF; and, in 43 cases (96%), it was found in the first
200 HPF. A significant negative correlation was observed between the MI and the number of HPF
counted before encountering the first mitosis (Spearman correlation coefficient, 20.487; P 5 0.017). Nonetheless, in some cases with mitotic indices ,5 mitoses
per 10 HPF (3 of 9), the first mitosis was found in the
first HPF examined, whereas, in 1 case with a high MI
(9 mitoses per 10 HPF), the first mitosis was found
only after study of 122 HPF. These results reflect regional heterogeneity of mitotic activity in PXA.
With regard to proliferative activity and its comparability between the three groups (no mitotic activity; ,5 mitoses per 10 HPF; $5 mitoses per 10 HPF)
comprising the study set, we compared the total number of HPF and found that the respective mean, SD,
and range values were 181 6 84 (41–344), 254 6 126
(35–507), and 239 6 88 (100 –381). The numbers of
HPF examined were significantly lower in the group
without mitoses than in that with ,5 mitoses per 10
HPF (P 5 0.03), whereas there were no significant
differences between the group without mitoses and
that with $5 mitoses per 10 HPF and between the
groups with #5 and $5 mitoses per 10 HPF(P . 0.05).
This finding, together with the fact that there were a
few cases in both the group without mitoses (3 of 26;
11%) and the one with ,5 mitoses per 10 HPF (3 of 32;
9%) in which less than 100 HPF were assessable
(whereas, in all tumors with $5 mitoses per 10 HPF,
more than 100 HPF were examined), suggests that a
modest underestimation of both the number of cases
showing mitotic activity as well as of the MI itself may
have taken place.
Although endothelial proliferation was not observed in any initial resection specimen, necrosis was
noted in 8 cases (11%). Our results are summarized
and compared with published data (Table 3). Surprisingly, published information regarding mitotic activity
is scant; only two tumors were subject to a formal
mitotic count.26,43
MIB-1 LIs performed in such a way as to reduce
methodological variation (see above) were available in
29 cases (41%), and the results were as follows:
mean 6 SD, 1.9 6 3.1%; median, 0.6%. In 23 cases
(79%), the labeling index was low (#2%). It exceeded
2% only in 6 instances (21%), the range being from
2.9% to 14%.
CANCER May 1, 1999 / Volume 85 / Number 9
Histologic and Follow-Up Data for 9 Patients with Pleomorphic Xanthoastrocytoma
Histologic features
Recurrence 4 yrs
Recurrence 4 yrs
Alive and well 0.4 yr
Recurrence 15 yrs
No follow-up
Recurrence 20 yrs
Alive 1 yr
Recurrence 1 yr
No follow-up
Recurrence 10 yrs
Presented with hemorrhage; died
Recurrence 1 yr
Recurrence 1 yr
Recurrence 2 yrs
Died 8 mos later from
intratumoral hemorrhage
Recurrence 6.7 yrs
Alive with disease 3.1 yrs
Recurrence 0.9 yr
Alive 1.4 yrs
Recurrence 1 yr
Alive 1 mo follow-up
MI: mitotic index (number of mitoses per 10 high-power fields; EP: endothelial proliferation; P: primary tumor; R1, R2, R3: first, second, and third recurrence.
Histologic Findings at Recurrence
We had the opportunity to review specimens obtained
at multiple points in time in 9 of 12 patients who
underwent surgery for recurrence on one (7 cases) or
more occasions (2 cases). Data regarding histologic
features and follow-up in these cases are summarized
in Table 4. Of the 9 specimens studied, 5 showed
histologic signs of anaplasia, as manifested by increased mitotic activity, and in 2 instances, by the
acquisition of necrosis and/or endothelial proliferation. Four tumors, 2 of which had a low MI (,5 mitoses per 10 HPF) and 2 of which a high MI ($5 mitoses
per 10 HPF), remained histologically unchanged at
Pleomorphic Xanthoastrocytoma: Therapy
Current series
Extent of resection
Gross total removal
Subtotal removal
Biopsy only
Adjuvant therapy
Radiation 1 chemotherapy
No. indicates the number of patients for whom information was available when this is different from
the total number.
Natural History and Prognosis of PXA/Giannini et al.
Pleomorphic Xanthoastrocytoma: Significant Prognostic Factors in
Recurrence Free Survival
P value
Factors and models
FIGURE 3. Recurrence free survival (Kaplan–Meier method). The curve for
the present series (solid line) is compared with that based on previously
reported cases (dashed line): The two are similar.
Univariate analyses
Extent of resection
0, ,5, $5 mitosesa
Atypical mitoses
Multivariate analyses
Extent of resection
0, ,5, $5 mitosesa
Current series
Literature review
ns: Not significant; nd: not done; na: not available.
Number of mitoses per 10 high-power fields.
Pleomorphic Xanthoastrocytoma: Significant Prognostic Factors in
Patient’s Overall Survival
P value
Factors and models
FIGURE 4. Overall survival curves (Kaplan–Meier method) for patients in the
present series (solid line) and for previously reported cases (dashed line) are
quite similar.
Treatment Modalities
Table 5 summarizes our data regarding initial treatment. Extent of surgical removal is based on estimates
of the neurosurgeon. Information regarding adjuvant
therapy is fragmentary. Twenty-eight patients received adjuvant therapy: 19 underwent external beam
radiation therapy (EBRT), 1 received chemotherapy,
and 8 underwent a combination of chemotherapy and
EBRT. Because most of our patients received adjuvant
therapy in their home community, there was no control over total dose of radiation employed, the ports
used to administer the radiation, or the drug used.
Survival Analysis
At termination of data collection, 59 patients (83%)
were alive, a rate that is strikingly similar to the literature value of 83% of the 100 reported patients in
whom follow-up data were available. Overall survival
in our study was 81% at 5 years and 70% at 10 years.
Recurrence free survival was 72% at 5 years and 61% at
10 years. Estimates of recurrence free survival and
overall survival curves based on our data and the those
Univariate analyses
0, ,5, $5 mitosesa
Extent of resection
Atypical mitoses
Multivariate analyses
0, ,5, $5 mitosesa
Current series
Literature review
ns: Not significant; nd: not done; na: not available.
Number of mitoses per 10 high-power fields.
generated from published data were comparable
(Figs. 3, 4).
Results of univariate and multivariate analyses for
recurrence free and overall survivals are summarized
in Tables 6 and 7. These show the P values for all
prognostic factors (see above) that were found to be of
significance in both the univariate and the stepwise
multivariate models. Variables are listed in order of
decreasing P values and are therefore of decreasing
probability. Where they were available, correlative
data based on our literature review also are provided.
Recurrence Free Survival
Extent of resection was the single factor most strongly
associated with recurrence free survival (Table 6, Fig.
5). Taken individually, MI and presence of atypical
mitoses also showed a significant correlation (Fig. 6).
In contrast to what was reported previously in the
CANCER May 1, 1999 / Volume 85 / Number 9
FIGURE 5. Recurrence free survival curves (Kaplan–Meier method) in the
present series based on extent of primary tumor resection. This measure of
survival was significantly longer in patients who underwent gross total resection rather than subtotal resection. In 14 cases, extent of resection could not
be established. (log-rank test; P 5 0.003).
Recurrence free survival curves (Kaplan–Meier method) in the
present series based on mitotic index [0, ,5, and $5 mitoses per 10
high-power fields (HPF)]. There are significant differences in this measure of
survival among patient groups (log-rank test; P 5 0.007). Two group comparisons demonstrate a significant difference in recurrence free survival between
tumors with 0 and $5 mitoses per 10 HPF [P 5 0.002; relative risk (RR), 7.5],
whereas the difference between tumors with ,5 and $5 mitoses per 10 HPF
approached significance (P 5 0.06).
Overall survival curves (Kaplan–Meier method) in the present
series based on mitotic index [0, ,5, and $5 mitoses per 10 high-power fields
(HPF)]. Significant differences in survival were noted among these patient
groups (log-rank test; P 5 0.001). Two group comparisons demonstrate a
significant difference in overall survival between tumors with ,5 versus $5
mitoses per 10 HPF [P 5 0.007; relative risk (RR), 6.4] and those with 0 versus
$5 mitoses per 10 HPF (P 5 0.005; relative risk, 23). The difference between
tumors with 0 versus ,5 mitoses per 10 HPF was not significant (ns).
literature and confirmed by our analysis of previously
reported cases (Table 6), in our series necrosis was not
significantly associated with recurrence free survival.
In the multivariate model, extent of resection and MI
were the only independent predictors.
Overall Survival
The MI (0, ,5, and $5 mitoses per 10 HPF) was the
single factor that was associated strongly with overall
survival (Table 7, Fig. 7); taken individually, the presence of atypical mitoses and of necrosis also were
significantly associated (Fig. 8). We noted no significant difference in overall survival based on extent of
resection (Fig. 9). In this multivariate model, MI was
the only independent predictor of overall survival.
Overall survival curves (Kaplan–Meier method) in the present
series based on the presence of necrosis. Survival was significantly longer in
patients with tumors that did not show necrosis at first resection compared
with those in which this parameter was present.(log-rank test; P 5 0.04).
Table 8 summarizes the data regarding extent of
resection (GTR vs. STR), MI (0, ,5, and $5 mitoses per
10 HPF), presence of necrosis, and MIB-1 LI (,2 vs.
$2) in patients who did not have tumor recurrence, in
those who had tumor recurrence but did not die, and
in patients who died.
The course of disease in patients who died and the
histologic features of their tumors were as follows:
Four patients experienced continuous tumor progression and died 0.3–1.6 years after initial surgery. Of
these, the tumors in 3 patients exhibited high mitotic
activity, 1 with necrosis at primary resection; the tumor in the fourth patient showed low mitotic activity
and no necrosis. Two patients whose tumor showed
high mitotic activity and necrosis at initial surgery
despite GTR followed by radiation and chemotherapy,
Natural History and Prognosis of PXA/Giannini et al.
Overall survival curves (Kaplan–Meier method) in the present
series based on extent of primary tumor resection. There was not a significant
difference in survival between patients who underwent gross total removal and
who underwent only subtotal resection (log-rank test; P 5 0.08). In 14 patients,
data regarding the extent of resection were not available.
experienced recurrences at 0.6 years and 0.8 years and
died at 2.5 years and 1.8 years, respectively. None of
these 6 patients underwent further surgery or autopsy.
Three patients whose tumors showed low mitotic activity and no necrosis at primary resection experienced one or multiple recurrences and eventually died
between 2.1 years and 4.8 years after primary surgery.
A recurrent tumor could be examined in only 1 patient
and showed histologic progression (see case 14, Table
4). One patient whose tumor exhibited necrosis but a
low MI experienced a recurrence after 10 years and
died postoperatively (see case 57, Table 4). Yet another
patient with a high MI but no necrosis at first surgery
died 5.7 years after primary surgery of a contralateral,
postirradiation anaplastic oligodendroglioma. Finally,
1 patient was well and free of seizures at 20.5 years,
when he died in a motor vehicle accident.
PXA is a new addition in the 1993 WHO classification
of tumors of the central nervous system.1 Although it
is rare and accounts for less than 1% of all astrocytic
neoplasms,72 it is the subject of numerous articles,
many reporting only single cases or small series. Since
the original description by Kepes et al. of 12 cases in
1979,2 their clinical and pathologic description of a
neoplasm with relatively favorable prognosis despite
ominous histologic features generally has been confirmed.4 – 67 In many cases, its superficial location, relative circumscription, and cyst/mural nodule architecture facilitate a gross total removal. These
architectural features, therefore, are considered major
determinants of the often favorable outcome.62,68
The role of histologic features and of tumor grade
in predicting biological behavior has received less at-
tention, although increased mitotic activity and necrosis have been associated with increased aggressiveness.62 For that reason, we studied a large group of
patients with PXA and undertook an analysis of previously published cases in which adequate information regarding morphology and outcome had been
provided. From the analysis of the data, it became
apparent that these two patient populations, one
mainly representing the combination of two large
North American consult practices and the other a
compilation of case reports and small series from the
international neuropathology community, were remarkably similar in their characteristics and closely
mirrored the original experience of Kepes et al.2 Numerous similarities, including age distribution, with a
high incidence in children and young adults, equal
frequency of occurrence in both males and females,
seizures as the most common presenting sign, predilection for the temporal lobe, and a high frequency of
cystic lesions, all contribute to making these two
groups of patients comparable. Furthermore, frequencies of GTR as well as of radiation therapy were quite
similar. Thus, it was not surprising that both recurrence free and overall survival curves were essentially
superimposable. Our study confirmed that PXA, with
its 70% 10-year survival rate, behaves significantly better than those lesions with which it was and continues
to be mistaken. Nonetheless, unlike other astrocytic
tumors of favorable prognostic type e.g., pilocytic astrocytoma and subependymal giant cell astrocytoma,
PXA is associated with a higher frequency of recurrence, anaplastic transformation, and death.
The question remained: Could one predict which
tumors would recur and/or behave more aggressively
based on morphologic and clinical information available at time of diagnosis? We carefully analyzed our
cases as well as published PXA cases in an effort to
extract as much useful information as possible regarding prognostic parameters of proven significance in
ordinary, diffuse astrocytic tumors1,73 and PXA,62,72
specifically, mitotic activity, endothelial proliferation,
and necrosis. Table 3 shows that, although the presence of necrosis has been recorded frequently in the
literature, information regarding mitotic activity is often barely mentioned.
We found that, in univariate analysis, the MI appeared to be predictive of both recurrence free and
overall survival. In contrast, the presence of necrosis
correlated significantly with overall survival, but not
with recurrence free survival. This last finding is at
odds with what was reported previously. Our analysis
of previously reported cases (Tables 6, 7), showed that
necrosis was a significant predictor of both recurrence
CANCER May 1, 1999 / Volume 85 / Number 9
Pleomorphic Xanthoastrocytoma: Summary Data by Group
Extent of surgery (%)a
Mitoses (%)b
MIB-1 LI (%)
No recurrence
32 (70)
3 (23)
4 (33)
9 (20)
3 (23)
6 (50)
20 (43)
4 (31)
2 (17)
21 (46)
7 (54)
4 (33)
5 (11)
2 (15)
6 (50)
4 (9)
0 (0)
4 (33)
14 (82)
5 (100)
4 (57)
3 (18)
3 (43)
GTR: gross total resection; STR: subtotal resection; LI: labeling index.
Information was available in only 57 of 71 cases.
Number of mitoses per 10 high-power fields.
and survival. Given the paucity of published data, the
significance of the MI could not be evaluated formally
in cases from the literature. Furthermore, in a multivariate analysis of our series, MI emerged as the only
histologic finding that, together with extent of surgical
resection, could independently predict tumor recurrence and overall survival. Such a comparison of the
relative weight of MI versus necrosis could not be
performed in previously reported cases.
We propose the designation “PXA with anaplastic
features” to denote PXA, featuring high mitotic activity
($5 mitoses per 10 HPF) with or without accompanying
necrosis. At present, it is unclear whether such tumors
should be termed anaplastic or Grade 3 (of 4). Use of the
term “anaplastic” might be found objectionable when
applied to a tumor that behaves less aggressively than
ordinary anaplastic astrocytoma and in which resectability more often affects the outcome. It is even less clear
just when, if ever, a PXA should be considered Grade 4 or
be equated with glioblastoma, as was suggested in the
1993 WHO blue book.1 We fully concur with Pahapill and
coauthors, who stress that necrotic PXAs are not and
should not be termed glioblastoma.62 Indeed, PXA appears to be yet another example of an astrocytic tumor
distinct from ordinary diffuse astrocytomas and to which
criteria used in grading such astrocytomas1,73 should not
be applied.
Five of the 9 cases in which we reviewed multiple
resection specimens acquired histologic anaplastic
features, as manifested by increased mitotic activity
and, in 2 cases, necrosis and/or endothelial proliferation (Table 4). It should be pointed out, however, that
histologic malignancy in PXA does not correlate with
prognosis as reliably as in patients with ordinary, diffuse astrocytomas. Indeed, the clinical course of patients with histologically malignant PXAs was often
more favorable and less precipitous than that of patients with fibrillary astrocytomas whose tumors
showed the same features.42,62 That these tumors are
inherently different is also supported by molecular
genetic studies, which suggest that the genetic events
underlying PXA formation and progression differ significantly from those involved in diffuse astrocytoma
tumorigenesis.64 These provide further support for
maintaining a separation between anaplastic PXA and
diffuse astrocytomas of Grade 3 (anaplastic) and
Grade 4 (glioblastoma).
Our study also confirmed the importance of extent of primary resection in the prevention of tumor
recurrence. However, it is noteworthy that the extent
of surgical resection did not appear to significantly
affect long term survival. Because 1) the extent of
resection was assessed by review of operative reports,
a method known to be imperfect compared to immediate postoperative magnetic resonance imaging evaluation, and 2) even this limited information was unavailable in nearly 20% of cases, caution must be used
in evaluating these data. A possible benefit to complete surgical resection, which did not reach statistical
significance in our series, was suggested in the literature review by Macauley et al.42 The more recent review of Pahapill clearly demonstrated a beneficial effect of complete resection on survival in patients with
PXAs that lacked necrosis,62 although survival was not
influenced by the extent of resection in cases featuring
necrosis. We conclude that, whenever possible, GTR
remains the goal of primary treatment. Furthermore,
in cases in which a GTR is achieved, we recommend a
“wait and see” approach followed, if necessary, by
additional surgery for residual and/or recurrent tumor.68 At present, the role of adjuvant radiotherapy
and/or chemotherapy remains uncertain. Our own
data could not provide precise guidelines, since only
19 patients underwent adjuvant EBRT, and 1 patient
received chemotherapy alone. One might extrapolate
from current management recommendations to employ EBRT in those patients in whom there is postop-
Natural History and Prognosis of PXA/Giannini et al.
erative residual tumor74 and who are thought to have
“anaplastic” features.75
We recognize that, in a slowly growing tumor such
as PXA, one with an inherently good prognosis, more
cases and longer periods of follow-up will be essential to
confirm or refute our findings. Nonetheless, given the
infrequency of these tumors72 and the difficulties inherent in obtaining homogeneous data, particularly long,
uniform follow-up on large patient numbers, we believe
that our data provide some guidelines. Performing a
prospective, randomized clinical trial of treatment modalities is practically impossible in a single center. Therefore, it becomes of importance, as suggested by Pahapill
et al., to make the best of the available, albeit imperfect
data and to follow as many patients as possible for long
periods of time.62 In fact, a “registry ” approach to such
uncommon tumors may be the only way for the knowledge base to expand. The neurooncology community
may need to assume the responsibility of creating such
databases. The establishment of some mechanism to
announce, facilitate data transfer, maintain, and analyze
this information will become a challenge for the future.
It is our goal to follow our patient cohort and to prospectively add each new case coming to our attention. The
formation of such a PXA “registry,” to which cases would
be gratefully received, will, hopefully, shed light on the
natural history of this unusual tumor and on factors
influencing its prognosis.
Kleihues P, Burger PC, Scheithauer BW. Histological typing
of tumours of the central nervous system. World Health
Organization. 2nd ed. New York: Springer-Verlag, 1993:
11– 4.
Kepes JJ, Rubinstein LJ, Eng LF. Pleomorphic xanthoastrocytoma: a distinctive meningocerebral glioma of young subjects with relatively favorable prognosis. A study of 12 cases.
Cancer 1979;44:1839 –52.
Kepes JJ, Kepes M, Slowik F. Fibrous xanthomas and xanthosarcomas of the meninges and the brain. Acta Neuropathol 1973;23:187–99.
Kuhajda FP, Mendelsohn G, Taxy JB, Long DM. Pleomorphic
xanthoastrocytoma: report of a case with light and electron
microscopy. Ultrastruct Pathol 1981;2:25–32.
Sakai H, Kawano N, Okada K, Tanabe T, Yagishita S. A case
of pleomorphic xanthoastrocyoma. No Shinkei Geka 1981;9:
1519 –24.
Heyerdahl Strom E, Skullerud K. Pleomorphic xanthoastrocytoma: report of 5 cases. Clin Neuropathol 1983;2:
188 –91.
Jones MC, Drut R, Raglia G, Pleomorphic xanthoastrocytoma: a report of two cases. Pediatr Pathol 1983;1:459 – 67.
Maleki M, Robitaille Y, Bertrand G. Atypical xanthoastrocytoma presenting as a meningioma. Surg Neurol 1983;20:
235– 8.
Kawano N. Pleomorphic xanthoastrocytoma: pathological
viewpoints on pleomorphic xanthoastrocytoma. Noshuyo
Byori 1983;1:85–90.
Weldon-Linne CM, Victor TA, Groothuis DR, Vick NA. Pleomorphic xanthoastrocytoma: ultrastructural and immunohistochemical study of a case with a rapidly fatal outcome
following surgery. Cancer 1983;52:2055– 63.
Bukeo T, Nakakura S, Nishimoto A, Tabucki K. Pleomorphic
xanthoastrocytoma (Kepes). No Shinkei Geka 1985;13:773–7.
Goldring S, Rich KM, Picker S. Experience with gliomas in
patients presenting with a chronic seizure disorder. Clin
Neurosurg 1985;3:15– 42.
Gomez JG, Garcia JH, Colon LE. A variant of cerebral glioma
called pleomorphic xanthoastrocytoma: case report. Neurosurgery 1985;16:703– 6.
Palma L, Maleci A, Di Lorenzo N, Lauro GM. Pleomorphic
xanthoastrocytoma with 18-year survival: case report. J Neurosurg 1985;63:808 –10.
Pasquier B, Kojder I, Labat F, Keddari E, Pasquier D, Stoebner P, et al. Le xanthoastrocytome du sujet jeune: revue de
la littérature à propos de deux observations d’évolution
discordante. Ann Pathol 1985;5:29 – 43.
Grant JW, Gallager PJ. Pleomorphic xanthoastrocytoma: immunohistochemical methods for differentiation from fibrous histiocytomas with similar morphology. Am J Surg
Pathol 1986;10:336 – 41.
Iwaki T, Fukui M, Kondo A, Matsushima T, Takeshita I.
Epithelial properties of pleomorphic xanthoastrocytomas
determined in ultrastructural and immunohistochemical
studies. Acta Neuropathol 1987;74:142–50.
MacKenzie JM. Pleomorphic xanthoastrocytoma in a 62year-old male. Neuropathol Appl Neurobiol 1987;13:481–7.
Severi P, Gambini C, Andrioli GC, Boccardo M. Pleomorphic
xantoastrocytoma. Pathologica 1987;79:507–12.
Gaskill SJ, Marlin AE, Saldivar V. Glioblastoma multiforme
masquerading as a pleomorphic xanthoastrocytoma. Child
Nerv Syst 1988;4:237– 40.
Paulus W, Peiffer J. Does the pleomorphic xanthoastrocytoma exist? Acta Neuropathol 1988;76:245–52.
Stuart G, Appleton DB, Cooke R. Pleomorphic xanthoastrocytoma: report of two cases. Neurosurgery 1988;22:422.
Ametani T, Gi H, Yamamoto S, Tsuda T, Kako M, Fukumitsu
T, et al. Pleomorphic xanthoastrocytoma: report of a case
and review of the literature. Nisseki Igaku Jpn Red Cross Med
J 1989;41:199 –206
Kepes JJ, Rubinstein LJ, Ansbacher L, Schreiber DJ. Histopathological features of recurrent pleomorphic xanthoastrocytomas: further corroboration of the glial nature of this
neoplasm. Acta Neuropathol 1989;78:585–93.
Whittle IR, Gordon A, Misra BK, Shaw JF, Steers AJW. Pleomorphic xanthoastrocytoma. J Neurosurg 1989;70:463– 8.
Sugita Y, Kepes JJ, Shigemori M, Kuramoto S, Reifenberger
G, Kiwit JCW, Wechsler W. Pleomorphic xanthoastrocytoma
with desmoplastic reaction: angiomatous variant. Report of
two cases. Clin Neuropathol 1990;9:271– 8.
Allegranza A, Ferraresi S, Bruzzone M, Giombini S. Cerebromeningeal pleomorphic xanthoastrocytoma. Report on four
cases: clinical, radiologic and pathological features (including a case with malignant evolution). Neurosurg Rev 1991;
Hosokawa Y, Tsuchihashi Y, Okabe H, Toyama M, Namura
K, Kuga M, et al. Pleomorphic xanthoastrocytoma: ultrastructural, immunohistochemical, and DNA cytofluorometic study of a case. Cancer 1991;68:853–9.
CANCER May 1, 1999 / Volume 85 / Number 9
29. Kawano N. Pleomorphic xanthoastrocytomas (PXA) in Japan: its clinicopathologic features and diagnostic clues.
Brain Tumor Pathol 1991;8:5–10.
30. Kros JM, Vecht CJ, Stefanko SZ. The pleomorphic xanthoastrocytoma and its differential diagnosis: a study of five cases.
Hum Pathol 1991;22:1128 –35.
31. Matsuhisa T, Maeshiro T, Kunishio K, Shigematsu H, Tsuno
K, Mishima N, et al. Proliferative potential of pleomorphic
xanthoastrocytoma: determined with nucleolar organizer
regions and monoclonal antibodies of Ki-67 and anti-DNA
polymerase a. Brain Tumor Pathol 1991;8:55– 8.
32. Sawyer JR, Roloson GJ, Chadduck WM, Boop FA. Cytogenetic findings in a pleomorphic xanthoastrocytoma. Cancer
Genet Cytogenet 1991;55:225–30.
33. Furuta A, Takahashi H, Ikuta F, Onda K, Takeda N, Tanaka
R. Temporal lobe tumor demonstrating ganglioglioma and
pleomorphic xanthoastrocytoma components. J Neurosurg
34. Kawano N. Pleomorphic xanthoastrocytoma: some new observations. Clin Neuropathol 1992;11:323– 8.
35. Lindboe CF, Cappelen J, Kepes JJ. Pleomorphic xanthoastrocytoma as a component of a cerebellar ganglioglioma:
case report. Neurosurgery 1992;31:353–5.
36. Rippe DJ, Boyko OB, Radi M, Worth R., Fuller GN. MRI of
temporal lobe pleomorphic xanthoastrocytoma. J Comput
Assist Tomogr 1992;16:856 –9.
37. Sawyer JR, Thomas EL, Roloson GJ, Chadduck WM, Boop
FA. Telomeric associations evolving to ring chromosomes in
a recurrent pleomorphic xanthoastrocytoma. Cancer Genet
Cytogenet 1992;60:188 –9.
38. Tien RD, Cardenas CA, Rajagopalan S. Pleomorphic xanthoastrocytoma of the brain: MR findings in six patients. AJR
39. Yoshino MT, Lucio R. Pleomorphic xanthoastrocytoma.
AJNR 1992; 13:1330 –2.
40. Zorzi F, Facchetti F, Baronchelli C, Cani E. Pleomorphic
xanthoastrocytoma: an immunohistochemical study of
three cases. Histopathology 1992;20:267–9.
41. Lipper MH, Eberhard DA, Phillips CD, Vezina LG, Cail WS.
Pleomorphic xanthoastrocytoma, a distinctive astroglial tumor: neuroradiologic and pathologic features. AJNR 1993;
14:1397– 404.
42. Macaulay RJB, Jay V, Hoffman HJ, Becker LE. Increased
mitotic activity as a negative prognostic indicator in pleomorphic xanthoastrocytoma. J Neurosurg 1993;79:761– 8.
43. Ozek MM, Sav A, Pamir MN, Ozer AF, Ozek E, Erzen C. Pleomorphic xanthoastrocytoma associated with von Recklinghausen neurofibromatosis. Child Nerv Syst 1993;9:39–42.
44. Thomas C, Golden B. Pleomorphic xanthoastrocytoma: report of two cases and brief review of the literature. Clin
Neuropathol 1993;12:97–101.
45. Herpers MJHM, Freling G, Beuls EAM. Pleomorphic xanthoastrocytoma in the spinal cord: case report. J Neurosurg
1994;80:564 –9.
46. Wasdahl DA, Scheithauer BW, Andrews BT, Jeffrey RA Jr.
Cerebellar pleomorphic xanthoastrocytoma: case report.
Neurosurgery 1994;35:947–51.
47. Bicik I, Raman R, Knightly JJ, Di Chiro G, Fulham MJ. PETFDG of pleomorphic xanthoastrocytoma. J Nucl Med 1995;
48. De Jesús O, Rifkinson N. Pleomorphic xanthoastrocytoma
with invasion of the tentorium and falx. Surg Neurol 1995;
49. Petropoulou K, Whiteman MLH, Altman NR, Bruce J, Morrison G. CT and MRI of pleomorphic xanthoastrocytoma:
unusual biologic behavior. J Comput Assist Tomogr 1995;19:
860 –5.
50. Sakamoto T, Sakakibara Y, Hayashi T, Yamashita K. Recurrence of pleomorphic xanthoastrocytoma six years after. No
Shinkei Gika 1995;23:941–5.
51. Cervoni L, Salvati M, Santoro A, Celli P. Pleomorphic xanthoastrocytoma: some observations. Neurosurg Rev 1996;19:
13– 6.
52. Daita G, Yonemasu Y, Muraoka S, Nakai H, Maeda T. A case
of anaplastic astrocytoma transformed from pleomorphic
xanthoastrocytoma. Brain Tumor Pathol 1991;8:63– 6.
53. Glasser RS, Rojiani AM, Mickle JP, Eskin TA. Delayed occurrence of cerebellar pleomorphic xanthoastrocytoma after
supratentorial pleomorphic xanthoastrocytoma removal.
J Neurosurg 1995;82:116 – 8.
54. Katoh M, Aida T, Sugimoto S, Suwamura Y, Abe H, Isu T, et
al. Immunohistochemical analysis of giant cell glioblastoma. Pathol Int 1995;45:275– 82.
55. Kordek R, Biernat W, Sapieja W, Alwasiak J, Liberski PP.
Pleomorphic xanthoastrocytoma with a glangliomatous
component: an immunohistochemical and ultrastructural
study. Acta Neuropathol 1995;89:194 –7.
56. Li YS, Ramsay DA, Fan YS, Armstrong RF, Del Maestro RF.
Cytogenic evidence that a tumor suppressor gene in the
long arm of chromosome 1 contributes to glioma growth.
Cancer Genet Cytogenet 1995;84:46 –50.
57. Haga S, Morioka T, Nishio S, Fukui M. Multicentric pleomorphic xanthoastrocytomas: case report. Neurosurgery
58. Kubo O, Sasahara A, Tajika Y, Kawamura H, Kawabatake H,
Takakura K. Pleomorphic xanthoastrocytoma with neurofibromatosis type 1: case report. Brain Tumor Pathol 1996;13:
79 – 83.
59. Lach B, Duggal N, DaSilva VF, Benoit BG. Association of
pleomorphic xanthoastrocytoma with cortical dysplasia and
neuronal tumors. Cancer 1996;78:2551– 63.
60. Lee TT, Landy HJ, Bruce JH. Arteriovenous malformation
associated with pleomorphic xanthoastrocytoma. Acta Neurochir 1996;138:590 –1.
61. Levy RA, Allen R, McKeever P. Pleomorphic xanthoastrocytoma presenting with massive intracranial hemorrhage.
AJNR 1996;17:154 – 6.
62. Pahapill PA, Ramsay DA, Del Maestro RF. Pleomorphic xanthoastrocytoma: case report and analysis of the literature
concerning the efficacy of resection and the significance of
necrosis. Neurosurgery 1996;38:822–9.
63. Pasquier B, Bost F, Peoc’h M, Barnoud R, Pasquier D. Données neuropathologiques dans I’épilepsie partielle pharmacorésistante. Ann Pathol 1996;16:174 – 81.
64. Paulus W, Lisle DK, Tonn JC, Wolf HK, Roggendorf W, Reeves
SA, Louis DN. Molecular genetic alterations in pleomorphic
xanthoastrocytoma. Acta Neuropathol 1996;91:293–7.
65. Powell SZ, Yachnis AT, Rorke LB, Rojiani AM, Eskin TA.
Divergent differentiation in pleomorphic xanthoastrocytoma. Am J Surg Pathol 1996;20:80 –5.
66. Van Roost D, Kristof R, Zentner J, Wolf HK, Schramm J. Clinical, radiological, and therapeutic features of pleomorphic
xanthoastrocytoma: report of three patients and review of
the literature. J Neurol 1996;60:690 –2.
Natural History and Prognosis of PXA/Giannini et al.
67. Tonn JC, Paulus W, Warmuth-Metz M, Schachenmayr W,
Sörenson N, Roosen K. Pleomorphic xanthoastrocytoma:
report of six cases with special consideration of diagnostic
and therapeutic pitfalls. Surg Neurol 1997;47:162–9.
68. Kepes JJ. Pleomorphic xanthostrocytoma: the birth of a diagnosis and a concept. Brain Pathol 1993;3:269 –74.
69. Giannini C, Scheithauer BW. Classification and grading of
low grade astrocytic tumors in children. Brain Pathol 1997;
70. Giannini C, Lopes MBS, Scheithauer BW, Hirose T, Vandenberg SR, Kros M. Immunophenotype of pleomorphic xanthoastrocytoma. J Neuropathol Exp Neurol 1998;57:501.
71. Giannini C, Scheithauer BW, Burger PC, Christensen MR, Wollan PC, Sebo TJ, et al. Cellular proliferation in pilocytic and
diffuse astrocytomas. J Neuropathol Exp Neurol (in press).
72. Kepes JJ, Louis DN, Paulus W. Pleomorphic xanthoastrocytoma. In: Kleihues P, editor. Pathology and genetics of tumours of the nervous system. Lyon: W.K. Cavenee International Agency for Research on Cancer, 1997:34 – 6.
73. Daumas-Duport C, Scheithauer BW, O’Fallon J, Kelly P.
Grading of astrocytomas. A simple and reproducible
method. Cancer 1988;62:2152– 65.
74. Berger MS, Deliganis AV, Dobbins J, Keles GE. The effect of
extent of resection on recurrence in patients with low grade
cerebral hemisphere gliomas. Cancer 1995;75:2785–7.
75. Dinapoli RP, Brown LD, Arusell RM, Barle JD, O’Fallon JR,
Buckner JC, Scheithauer BW, Krook JE, Tsechetter LK, Maier
JA. Phase III comparative evaluation of pCNU and carmustine combined with radiation therapy in high grade glioma.
J Clin Oncol 1993;11:1316 –21.
Без категории
Размер файла
558 Кб
Пожаловаться на содержимое документа