Purpose: Glioblastoma (GB, WHO grade IV) is the most common primary brain tumor in adults. Survival is typically <1 year but varies between a few months and a couple of years. The aim of the study was to find novel genetic prognostic factors in a well-defined GB series.

Experimental Design: The survival data on 129 GBs were correlated with the results of a detailed analysis of 9 genes. These included 3 genes coding for proteins in the p53 pathway (i.e., TP53, p14ARF, and MDM2), 4 in the Rb1 pathway (i.e., CDKN2A, CDKN2B, RB1, and CDK4), as well as PTEN and epidermal growth factor receptor.

Results: We found that abnormalities in any of the four genes (CDKN2A, CDKN2B, RB1, and CDK4) coding for components of the Rb1 pathway were associated with shorter survival (P = 0.002). In combination with loss of wild-type PTEN, the association was even stronger (P < 0.001), the median survival being 166 days as compared with the group without these abnormalities where the median postoperative survival was 437 days. The survival difference remained statistically significant in Cox’ regression analysis adjusting for age (P = 0.012).

Conclusions: The findings indicate that knowledge of the molecular genetic abnormalities in GBs provides important data in assessing individual patients. As additional advances in our understanding of the molecular genetics and cell biology of gliomas are made, in addition to providing prognostic information, such data may also provide targets for innovative therapy in the individual case.

GB3 is the most malignant form of astrocytic tumor and the most common primary brain tumor in adults. Survival varies from a couple of months to a couple of years, with a median survival of <1 year, despite modern treatment (1, 2). The most important prognostic determinant is age, young patients surviving longer (3, 4). Better performance status (1, 4), extensive surgery, and high-dose postoperative radiation (5, 6) are also associated with longer survival.

More than 12 genes have been identified as involved in the development of human GB. Studies of abnormalities of ≤6 genes and outcome have not produced indisputable correlations with survival (7, 8, 9, 10, 11, 12, 13, 14, 15). A study of pediatric astrocytomas showed an association between loss of wild-type PTEN and poor outcome, but no separate analysis of GB was reported (16).

Aberrations of CDKN2A, CDKN2B, p14ARF, RB1, TP53, PTEN, MDM2, EGFR, and CDK4 genes are seen relatively frequently in GB (17, 18, 19, 20, 21, 22, 23), whereas amplifications of MDM4, CDK6, CCND1, CCND3, and platelet-derived growth factorA-R have been reported in few cases (24, 25, 26, 27, 28).

The p14ARF, MDM2, and TP53 genes code for proteins involved in the p53 pathway (29). The CDKN2A, CDKN2B, CDK4, and RB1 genes code for proteins in the Rb1 pathway (30). The majority of GBs shows abnormalities of both pathways, and generally only one gene in each pathway is affected (19, 31). Loss of wild-type PTEN is one of the most frequent genetic abnormalities in GB (15, 21), and the gene most commonly amplified in GB is EGFR(32, 33, 34). Both of the latter are relatively uncommon in AA.

In an attempt to establish whether knowledge of the presence or absence of abnormalities of the genes most frequently found aberrant in GBs is clinically useful, we have compared genetic data from 129 GBs with patient survival. Here we present both clinical and molecular genetic data indicating that some patterns of mutations are associated with a particularly poor outcome.

Patients and Tumor Tissue.

A total of 130 GB patients operated at the Karolinska Hospital, Stockholm, or Sahlgrenska University Hospital, Gothenburg, between 1989 and 1994 were included. Survival analysis could be assessed in 129. End of follow-up was set to November 1, 2000. Clinical data included age and sex of the patient, tumor localization, duration, and type of symptoms before diagnosis, date of operation, RT given, and date of death. In general, the operations were gross total; no cases with only biopsy were included. Data were missing from a maximum of three cases for any 1 of the 12 clinical parameters. Performance status was not routinely documented in the patients’ files and could therefore not be included in this retrospective study. All diagnoses were revised according to the most recent WHO classification (35). All patients gave informed consent, and ethics committee approval was obtained at all sites.

Tumor tissue was obtained from the first GB operation. Deletions, mutations, and amplifications of genes were identified by comparing tumor cell DNA with the individuals’ WBC DNA. The details of the methods used in the analysis of the Rb1 pathway genes (RB1, CDKN2A, CDKN2B, and CDK4) and the p53 pathway genes (TP53, p14ARF, and MDM2) have been described in detail previously (19, 31). This is also true for the analysis of PTEN(21) and EGFR genes (33, 36).

Briefly, tumor suppressor genes were first analyzed for allele number by microsatellite analysis and multiplex PCR followed by mutation screening using single-strand conformational polymorphism or DGGE and sequencing. Amplification was established using both microsatellite analysis and Southern Blotting with densitometry. For any genetic parameter, data were missing from a maximum of four cases.

Definition of “Genetic Abnormality” and Abnormality of Pathways.

The molecular information was restricted to genetic data. All genetic data were categorized as either “normal” or “abnormal.”

The tumor suppressor genes CDKN2A, CDKN2B, RB1, p14ARF, and PTEN were considered “abnormal” when there was no wild-type allele present, either attributable to homozygous deletion or deletion of one allele and mutation of the other. Hemizygous deletion and retention of one wild-type allele was categorized as “normal.”

TP53 was considered abnormal whenever a somatic mutation was identified in one or more alleles. As the protein functions as a tetramer, one mutated monomer may disrupt tetramer function (37). A separate analysis of cases with no wild-type TP53 was also carried out. Hemizygous deletion of TP53 without detectable mutation of the retained allele was categorized as “normal.” For EGFR, CDK4, and MDM2, gene amplification (more than four copies) was categorized as abnormal (19, 33).

When categorizing the pathways, if at least one of the genes coding for a protein involved in the p53 pathway (p14ARF, MDM2, and TP53) or the Rb1 pathway (CDKN2A, CDKN2B, RB1, and CDK4) was abnormal (as defined above), the pathway was categorized as “abnormal.” Only when all genes studied in a pathway were categorized as normal was the pathway categorized as normal.

Definition of “Primary” and “Secondary” GB.

Patients with no previous histological diagnosis of astrocytoma of lower grade were defined as having primary GB, as all patients with a suspect lesion on computed tomography or magnetic resonance imaging were routinely biopsied. The median duration of symptoms in this group was 2 months (range: 0–21).

All patients in the secondary GB group had a previous histological diagnosis of astrocytoma (WHO grade II) or anaplastic astrocytoma (WHO grade III). The interval between primary tumor diagnosis and the secondary GB varied between 14 months and 30 years (median: 39 months).

Statistical Analysis.

The starting point for patient survival was the date of the first GB operation. Of the 130 tumors analyzed, 122 were primary GB, and 8 were secondary GB. Among these 8 patients, one was first operated for an astrocytoma (WHO grade II) that recurred as a GB and was resected after 6 years and recurred again and was reoperated 2 years later. We had only access to tissue from the second GB operation, so no comparisons between genetic data and survival were carried out, excluding this single patient from all survival analysis.

To study the potential influence of single clinical factors on patient survival, univariate analysis using Wilcoxon-Gehan statistic (38) was performed, except when analyzing the age factor. Then, Cox’s regression analysis was used considering age as a continuous variable. Analyses of genetic factors are given both univariately (Wilcoxon-Gehan statistic) and multivariately adjusting for age (Cox’s regression analysis). For the factors showing significant association with survival in univariate analysis, Cox’s multivariate regression analysis was undertaken, with all such factors forced into the model. Where appropriate, a trend test was performed (39). When comparing age distribution between two groups, we used Student’s t test. Statistical Package for the Social Sciences for Windows, release 11.0.5, was used for all statistical analyses.

Individual Patients’ Clinical Data.

The clinical data are summarized in Table 1. The 2 youngest patients were alive at the end of follow-up with survival times of 6.7 and 7.5 years. The 2 patients with infratentorial tumors survived 185 and 431 days, respectively. Patient age was inversely correlated to survival (P < 0.001), and when divided into four age groups (<50, 50–59, 60–69, and ≥70 years), the older group in any comparison had a shorter survival.

After surgery, 66 patients were treated with postoperative RT (23 received 37–47 Gy, and 39 received 50–56 Gy). One patient received only 24 Gy. For 3 patients, the RT-notes were missing. Fifty-one patients received no RT. The 66 postoperatively irradiated patients had a median survival of 357 days (mean age 51 years), whereas the median survival of the nonirradiated patients was 156 days (mean age 62 years). The survival difference was highly significant, even when adjusted for age (P < 0.001).

The 6 patients who developed a secondary GB from an anaplastic astrocytoma (AA, WHO grade III) had all received RT (37–54 Gy) after the AA operation. Their median survival after the GB operation was 248 days (range: 175-1823 days). Two patients developed a GB by progression from an astrocytoma (WHO grade II). One received 37 Gy after the GB operation and the other radiosurgery twice. They survived 1037 and 1165 days.

Forty-four patients were treated with chemotherapy with varying p.o. schedules, the majority receiving 1-(2-chloroethyl)-3-cyclohexyl-1-nitrosourea on disease progression. Of these, 31 (70%) had also received RT.

Abnormalities of Single Genes or Single Pathways and Survival.

A summary of the types of genetic abnormalities affecting each gene is included in Table 2a. Each single gene or pathway abnormality (as defined above) was tested in a univariate manner for influence on survival (Table 2b). The only gene significantly associated with survival in univariate analysis was PTEN (P = 0.007). The PTEN gene normal group had a median postoperative survival of 289 days compared with 198 days for the abnormal group. Using Cox’s regression analysis to adjust for age, the survival difference was borderline significant (P = 0.078). The mean age of the 60 patients with abnormal PTEN (57 years) was higher than the 65 patients with normal PTEN (54 years), although this was not statistically significant (P = 0.174). The CDK4 gene showed no significant association with survival in univariate analysis (P = 0.299). However, when adjusted for age, the abnormal group showed significantly shorter survival (P = 0.019).

The normal TP53 group (one or two wild-type alleles only) had a nonsignificantly shorter survival than tumors with abnormal TP53. When an additional analysis was performed categorizing only tumors with loss of both wild-type TP53 alleles as abnormal, the abnormal group had the shorter survival.

When the pathways are analyzed, the group with abnormalities of any of the genes (CDKN2A, CDKN2B, RB1, and CDK4) coding for proteins involved in the Rb1 pathway had a significantly shorter median survival than the group with a normal Rb1 pathway (P = 0.002). When adjusted for age, this difference was still significant (P = 0.033). The group with abnormal p53 pathway had a significantly shorter survival than the normal group (P = 0.008), but when adjusted for age, this difference was not significant (P = 0.345). It is important to note that 96% of the patients with an abnormal Rb1 pathway had also an abnormal p53 pathway.

Combinations of Abnormalities.

We then combined pair-wise the PTEN gene, the EGFR gene, the p53 pathway, and the Rb1 pathway status. This gave us six pairs, each pair divided into four groups because of the four possible combinations (both factors normal; one normal, the other abnormal; and vice versa; both abnormal). The median postoperative survival data and results from statistical analysis of the survival differences between the paired groups, using Wilcoxon-Gehan statistic for univariate analysis and Cox’s regression analysis adjusting for age, are shown in Table 3.

In all but one of the six pairs (PTEN gene/EGFR gene), the median postoperative survival in the abnormal/abnormal group was shorter than that in the normal/normal group. For two of these pairs (the Rb1 pathway/PTEN gene and Rb1 pathway/p53 pathway pairs), the two “normal/abnormal” pairs showed intermediate survival values. The survival differences in each of these five pairs were significant in univariate analysis. The greatest survival differences were observed within the Rb1 pathway/PTEN gene pair (P < 0.001) and statistically significant even when adjusted for age (P = 0.012). A trend test, further exploring the trend from long survival in the normal/normal group to short survival in the abnormal/abnormal group, gave P < 0.001. When survival data for the Rb1 pathway/EGFR pair was analyzed adjusting for age using Cox’s regression, the survival differences were borderline significant (P = 0.044) but showed no continuous trend from better to worse prognosis between the groups.

In Table 4, the consistency of the association of the Rb1 pathway/PTEN gene pair with survival, in clinically defined subgroups, is evaluated further. Similar survival differences, as described above for all 129 GB, were seen within all subgroups.

Multivariate Analysis.

To further explore the indepen-dence of the associations between Rb1 pathway and PTEN abnormalities with survival, Cox’s multivariate regression analysis was performed. All factors showing significant association with survival in univariate analysis (age, postoperative RT, left or right tumor location, Rb1 pathway, and PTEN) were included. As shown in Table 5, all these factors, with the exception of PTEN, showed an independent impact on survival.

Long Survivors.

Five patients survived >3 years. These include the 2 youngest patients and 3 patients aged between 35 and 45 years. Two of the eight secondary GB are found among these 5 long-term survivors, and these 2 had a clinical history before the GB operation of 11 and 30 years, respectively. Both had abnormalities of p53 and Rb1 pathways in their GBs, and they died of their disease after 1165 and 1823 days, respectively. The other 3 were primary GBs with duration of symptoms of between 0.5 and 3 months. All 3 of them had no detected abnormalities of the genes examined in this study, but we have additional data on all 3 showing other clonal genetic abnormalities affecting presently unknown genes. One of these 3 has died of his tumor surviving 1175 days. The remaining 2 were alive at the close of the study.

Primary and Secondary GB.

The 122 primary GBs (as defined above) had preoperative symptoms ranging from 1 week to 21 months (median: 2 months), 34% having had symptoms ≥3 months. Their mean age (58 years) and median survival (223 days) were comparable with the patients with <3 months history (56 years; 247 days). In addition, frequencies of mutations in these two subgroups were similar (Table 6). Therefore, they were combined when comparing with the secondary GBs.

Only 8 patients had clear evidence of progression to secondary GB. The mean age was 39 years, and median survival after GB diagnosis was 339 days. Although this was longer than the “primary’ GBs, the difference in survival did not reach statistical significance (P = 0.175).

As shown in Table 6, there were clear differences in the incidences of mutations of some genes between primary and secondary GB (TP53, EGFR, and PTEN). Because of the low numbers of secondary GB, no statistical analyses were applied to these data.

This study has focused on a relatively large series of a single, histologically well-defined tumor entity, classified according to the recent WHO criteria. We have reviewed the Swedish Cancer Register data and found that ∼20% of newly diagnosed GBs in our catchment area are included. The material was found to be representative regarding age and sex, although the mean age was slightly lower than the mean for all cases in the register, probably because of the fact that resection was necessary for inclusion. Patients with high age and/or complicating diseases, other than GB, were thus automatically excluded. An analysis of the material shows survival figures to be similar to other reports, and age is inversely correlated with survival, both well-established findings in other GB series. Taken together, these data indicate that the material is comparable with that used in many previous studies of prognostic indicators. Therapy was not uniform, reflecting the lack of effective treatment. We have in our analysis attempted to correct for this by examining relatively homogenous therapy groups. Although GBs, in contrast to some other glial tumors, such as oligodendrogliomas, rarely show a measurable response to RT or chemotherapy, RT is generally accepted as prolonging survival. This is also found in this study, where in multivariate analysis, RT was second to age (Table 5). Overall, survival with GB is poor, and the tumors are difficult to excise completely, making “recurrence-free survival” hard to assess. Therefore, we have chosen survival as our end point.

The aim of this study was to identify novel genetic prognostic factors, either single genes or combinations of genes in a series consisting of only GBs. Adjusted for age, no single gene was significantly associated with survival, although loss of wild-type PTEN showed borderline significance (P < 0.078). This is perhaps not surprising considering the number of previous reports on this issue lacking conclusive results. It is unlikely that such a simple association exists considering the complexity of the cellular machinery. We then explored the possibility of an association between combinations of genetic abnormalities and prognosis. When choosing relevant combinations, we found it most appropriate to link together genes coding for proteins involved in the same cellular mechanism rather than random combinations. We also wished to examine whether combinations of aberrant cellular pathways could have synergistic or additive effects. If so, we would expect a spectrum of survival figures from the best for the normal/normal group, intermediate for the normal/abnormal groups with the abnormal/abnormal group showing shortest survival, i.e. a clear tendency toward either better or worse prognosis. In univariate analysis of single genes or pathways, the combination of PTEN and Rb1 pathway data showed median survival figures best fitting with such an additional effect (Table 3). This finding was also tested in the treatment subgroups, and the tendencies were the same for all (Table 4). In multivariate analysis, including all factors showing significant association with survival in univariate analysis, disruption of the Rb1 pathway, but not loss of wild-type PTEN, showed independent impact on survival (Table 5). Also included were age and RT, which are well-known predictors of outcome, whereas location of the tumor is not. Our finding that a right-sided tumor is associated with short survival was surprising, and we have found no other indication of that in the literature.

It should be noted that the vast majority (101 of 105) of the tumors with an aberrant Rb1 pathway also had a demonstrable aberrant p53 pathway. An analysis of the patients with aberrant p53 and Rb1 pathways with and without wild-type PTEN gave similar findings. Experimental studies have demonstrated that disruption of the Rb1 pathway can only occur simultaneously with or after disruption of the p53 pathway (40, 41). Disruption of the Rb1 pathway alone leads invariably to p53 pathway-induced apoptosis (42). In the present study, only 3 genes in the p53 pathway were analyzed. This pathway is not completely understood, but it is certain that many other genes are involved, and it is conceivable that one or another of these may be aberrant in the four tumors that showed no abnormalities of p14ARF, MDM2, or TP53. In addition, epigenetic phenomena, such as methylation, may be operative.

The patient group identified in this study with the worst survival data had disruption of the Rb1, p53, and PTEN/Akt pathways. An explanation as to why this combination is associated with an aggressive tumor phenotype is relatively easy to propose. A dysfunctional Rb1 pathway would lead to inappropriate progression from G1 into S phase of the cell cycle. Combining this with loss of the normal p53 pathway and the activation of Akt secondary to the loss of wild-type PTEN would disrupt the cellular self-eradicating affect of apoptosis at two levels. A recent study of the stromal and epithelial components of breast carcinomas showed mutual exclusive loss of the TP53 and PTEN wild-type genes within each of these two compartments (43). In our study, 27 (22%) of the GBs had TP53 mutation and loss of wild-type PTEN, in concordance with previous reports on GB (9, 44).

Five of the GBs were classified as normal for all nine genes analyzed. We have examined all our data on these cases, and the histology of a fraction of the piece studied shows GB in all cases; in four of them, other clonal genetic aberrations have been identified, excluding the possibility of the study of nontumor tissue in these cases. We are sure that additional analysis will identify genetic abnormalities in the remaining case.

Analysis of the number of genetic abnormalities found in each case revealed no associations with survival (data not shown). One might expect an inverse correlation between survival and number of mutations, but this was not seen. This supports the idea that the pattern of mutations, rather than the number of “hits,” dictates prognosis.

As genetic and other molecular data have accumulated on GBs and other tumors, attempts have been made to identify more reliable predictors of prognosis and response to therapy. However, most findings remain controversial or inconclusive. Studies of GB and prognosis have addressed clinical, histopathological, proteomic, and genetic factors, either singly or in combination, often with different end points. Most studies describe gene abnormalities and their potential association to survival tested singly. Many focus on either tumor suppressor genes or oncogenes or various combinations of these groups (8, 9, 10, 11, 12, 15). Raffel et al.(16) correlated the status of the PTEN, CDK4, CDKN2A/p14ARF, TP53, MDM2, and EGFR genes and patient survival in a series of “pediatric malignant astrocytomas” (both malignancy grade III and IV). They reported PTEN mutation to be significantly associated with shorter survival.

There has been considerable interest in the association between response to chemotherapy and genetic data (allelic loss from 1p and 19q) in anaplastic oligodendroglioma (WHO grade III; Refs. 45 and 46). No such findings have been clearly described in GB or other astrocytic tumors.

Many studies of GB have focused on characteristics of primary and secondary GB. Although histopathologically indistinguishable, several clinical and genetic differences have been observed. Primary GBs are clinically defined as having developed after a short period of symptoms, typically <3 months (35). In our series, median time of symptoms was 2 months for primary GB and 41 months for secondary GB. Our findings also support the observation that secondary GB occurs in younger individuals, the mean age being 39 years as compared with the 56.5 years of the patients with primary GB. There have been proposals that TP53 mutations characterize secondary GB, whereas EGFR amplification is a characteristic of primary GB. Our findings support the observation that TP53 mutations are more common but by no means exclusive to secondary GB (47). We have reported previously that two-thirds of astrocytomas and anaplastic astrocytomas have TP53 mutations but in almost all cases no disruption of the Rb1 pathway. The vast majority of GBs have both p53 and Rb1 pathway disruption (19). Thus, Rb1 pathway disruption would appear to be a progression factor. The fact that coamplification of the CDK4 and MDM2 genes (12q13–15) or codeletion of the CDKN2A and p14ARF genes (9p12) disrupt both the Rb1 pathway and the p53 pathway by a minimal number of genetic events may explain the high frequency of these abnormalities in primary GB. Secondary GBs, on the other hand, develop from astrocytomas or anaplastic astrocytomas, two-thirds of which have TP53 mutations. They then must develop further individual mutations of the genes coding for components of the Rb1 pathway to progress to GB.

In our small series of secondary GB, there are no cases with EGFR amplification confirming previous reports (8, 48, 49). We also find a higher frequency of loss of wild-type PTEN in primary GB (49). However, we observed a similar frequency of MDM2 amplification and loss of wild-type CDKN2A in primary and secondary GB, a finding that does not confirm previous reports (50, 51, 52).

We have only begun the molecular dissection of solid tumors. This is a relatively small study, despite the fact it is the most comprehensive to date. It is difficult to find homogeneously treated groups of GB outside of trials in a situation where there is no successful treatment. However, much more genetic and molecular data will become available in the future. Micro-array and other technologies are being developed and will make feasible the documentation of a broad spectrum of genetic and other characteristics of GBs that will undoubtedly assist in prognostication and stratification for therapy. Such data will also form the basis for the logical design of patient-specific therapy.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1

Supported by grants from the Swedish Cancer Society, Stockholm’s Cancer Society, King Gustaf V. Jubilee Fund, CAMPOD, CRUK, the Ludwig Institute for Cancer Research, and the Jacqueline Seroussi Memorial Foundation for Cancer research.

3

The abbreviations used are: GB, glioblastoma; EGFR, epidermal growth factor receptor; CDK, cyclin-dependent kinase; Rb, retinoblastoma; RT, radiotherapy; DGGE, denaturing gradient gel electrophoresis.

Table 1

Summary of clinical data

No. of casesMedian survival (days)P                  a
Overall postoperative survival (Deceased 127; alive 2)  129 240  
Sex Male 79 (61%) 274 0.477 
 Female 50 (39%) 219  
Age (mean = 56; range 12–81) <50 39 431  
 50–59 29 274 <0.001b 
 60–69 46 177  
 70− 15 119  
Local Frontal 30 245  
 Temporal 36 219  
 Parietal 19 305 0.652 
 Occipital 243  
 >1 cerebral lobe 34 228  
 Infratentorial 308  
Side Left 64 305 0.039 
 Right 63 211  
Progression from lower grade  339 0.175 
“Primary” glioblastoma  122 222.5  
Postoperative radiotherapy Yes 66 357 <0.001 
 No 51 156  
No. of casesMedian survival (days)P                  a
Overall postoperative survival (Deceased 127; alive 2)  129 240  
Sex Male 79 (61%) 274 0.477 
 Female 50 (39%) 219  
Age (mean = 56; range 12–81) <50 39 431  
 50–59 29 274 <0.001b 
 60–69 46 177  
 70− 15 119  
Local Frontal 30 245  
 Temporal 36 219  
 Parietal 19 305 0.652 
 Occipital 243  
 >1 cerebral lobe 34 228  
 Infratentorial 308  
Side Left 64 305 0.039 
 Right 63 211  
Progression from lower grade  339 0.175 
“Primary” glioblastoma  122 222.5  
Postoperative radiotherapy Yes 66 357 <0.001 
 No 51 156  
a

Univariate analysis of postoperative survival using Wilcoxon-Gehan.

b

Cox’s regression analysis was used considering age as a continuous variable.

Table 2A

Summary of gene status in all cases

CDKN2ACDKN2BCDK4RB1TP53p14ARFMDM2PTENEGFR
Number of cases with wild type+ wild type 40 42  76 79 40  12  
Number of cases with deletion+ wild type 20 21  33 19  53  
Number of cases with mutation+ wild type  28   
Number of cases with deletion+ mutation  13 17  53  
Number of cases with homozygous deletion 65 66  67   
Number of cases with amplification   20    13  48 
CDKN2ACDKN2BCDK4RB1TP53p14ARFMDM2PTENEGFR
Number of cases with wild type+ wild type 40 42  76 79 40  12  
Number of cases with deletion+ wild type 20 21  33 19  53  
Number of cases with mutation+ wild type  28   
Number of cases with deletion+ mutation  13 17  53  
Number of cases with homozygous deletion 65 66  67   
Number of cases with amplification   20    13  48 
Table 2B

Median postoperative survival in cases with and without individual gene abnormalities and pathway abnormalitiesa

Rb1 pathwayp53 pathwayPTENEGFR
CDKN2ACDKN2BCDK4RB1Rb1 pathwaybTP53p14ARFMDM2p53 pathwayc
Median survival in days in cases without gene abnormalitiesa (= “normal”) 256 272 249 236 431 224 240 248 431 289 223 
(n =) (60) (63) (109) (109) (23) (84) (59) (116) (15) (65) (78) 
Median survival in days in cases with gene abnormalitiesa (= “abnormal”) 227 226 186 279 221 254 235 185 223 198 304 
(n =) (69) (66) (20) (16) (105) (45) (70) (13) (114) (60) (48) 
Wilcoxon-Gehan P0.162 0.102 0.299 0.850 0.002 0.584 0.158 0.397 0.008 0.007 0.558 
Cox’s regressiondP0.651 0.741 0.019 0.166 0.033 0.829 0.565 0.091 0.345 0.078 0.511 
Rb1 pathwayp53 pathwayPTENEGFR
CDKN2ACDKN2BCDK4RB1Rb1 pathwaybTP53p14ARFMDM2p53 pathwayc
Median survival in days in cases without gene abnormalitiesa (= “normal”) 256 272 249 236 431 224 240 248 431 289 223 
(n =) (60) (63) (109) (109) (23) (84) (59) (116) (15) (65) (78) 
Median survival in days in cases with gene abnormalitiesa (= “abnormal”) 227 226 186 279 221 254 235 185 223 198 304 
(n =) (69) (66) (20) (16) (105) (45) (70) (13) (114) (60) (48) 
Wilcoxon-Gehan P0.162 0.102 0.299 0.850 0.002 0.584 0.158 0.397 0.008 0.007 0.558 
Cox’s regressiondP0.651 0.741 0.019 0.166 0.033 0.829 0.565 0.091 0.345 0.078 0.511 
a

For definitions of gene abnormalities, see “Materials and Methods.”

b

Any abnormality (as defined a) affecting genes in the Rb1 pathway.

c

Any abnormality (as defined a) affecting genes in the p53 pathway.

d

Cox’s regression analysis performed to adjust for the age factor.

Table 3

Univariate and Cox’s regression analysisa of combined genetic abnormalitiesb and median survival in days

p53 pathwaybEGFRPTEN
“Normal”“Abnormal”“Normal”“Abnormal”“Normal”“Abnormal”
Rb1 pathwayb “Normal” (n =) 516 (10) 277 (13) 465 (17) 200 (4) 437 (11) 252 (12) 
 “Abnormal” (n =) 323 (4) 215 (101) 201 (61) 311 (43) 279 (54) 166 (48) 
 Wilcoxon-Gehan P = 0.006  P = 0.002  P < 0.001  
 Cox Regression P = 0.189  P = 0.044  P = 0.012  
PTEN “Normal” (n =) 394 (8) 280 (57) 277 (39) 339 (25)   
 “Abnormal” (n =) 483 (6) 180 (54) 179 (39) 300 (20)   
 Wilcoxon-Gehan P = 0.001  P = 0.085    
 Cox Regression P = 0.109  P = 0.293    
EGFR “Normal” (n =) 453 (11) 211 (67)     
 “Abnormal” (n =) 194 (3) 311 (45)     
 Wilcoxon-Gehan P = 0.005      
 Cox Regression P = 0.099      
p53 pathwaybEGFRPTEN
“Normal”“Abnormal”“Normal”“Abnormal”“Normal”“Abnormal”
Rb1 pathwayb “Normal” (n =) 516 (10) 277 (13) 465 (17) 200 (4) 437 (11) 252 (12) 
 “Abnormal” (n =) 323 (4) 215 (101) 201 (61) 311 (43) 279 (54) 166 (48) 
 Wilcoxon-Gehan P = 0.006  P = 0.002  P < 0.001  
 Cox Regression P = 0.189  P = 0.044  P = 0.012  
PTEN “Normal” (n =) 394 (8) 280 (57) 277 (39) 339 (25)   
 “Abnormal” (n =) 483 (6) 180 (54) 179 (39) 300 (20)   
 Wilcoxon-Gehan P = 0.001  P = 0.085    
 Cox Regression P = 0.109  P = 0.293    
EGFR “Normal” (n =) 453 (11) 211 (67)     
 “Abnormal” (n =) 194 (3) 311 (45)     
 Wilcoxon-Gehan P = 0.005      
 Cox Regression P = 0.099      
a

Cox’s regression analysis performed to adjust for the age factor.

b

For definitions, see “Materials and Methods.”

Table 4

Combination of Rb1 pathway and PTEN abnormalities and survival in clinically defined subgroups

Group selectedGenetic dataaNo. of casesPostoperative survivalStatistical analysis
Rb1 pathwayPTENMedianLongest survivalWilcoxon-GehanCox’s regressionb
All cases (n = 125) “Normal” “Normal” 11 437 >2738 P < 0.001 P = 0.012 
 “Normal” “Abnormal” 12 252 914   
 “Abnormal” “Normal” 54 279 1823   
 “Abnormal” “Abnormal” 48 166 888   
Cases with survival <30 days excluded (n = 120) “Normal” “Normal” 11 437 >2738 P = 0.001 P = 0.011 
 “Normal” “Abnormal” 12 252 914   
 “Abnormal” “Normal” 52 282 1823   
 “Abnormal” “Abnormal” 45 179 888   
“Primary” GB (n = 119) “Normal” “Normal” 10 600 >2738 P = 0.002 P = 0.013 
 “Normal” “Abnormal” 12 252 914   
 “Abnormal” “Normal” 50 270 729   
 “Abnormal” “Abnormal” 47 163 888   
Cases with postoperative RT (n = 66) “Normal” “Normal” 841 >2449 P = 0.021 P = 0.031 
 “Normal” “Abnormal” 489 914   
 “Abnormal” “Normal” 30 365 1165   
 “Abnormal” “Abnormal” 24 301 888   
Cases not given RT (n = 50) “Normal” “Normal” 318 >2738 P = 0.022 P = 0.132 
 “Normal” “Abnormal” 187 232   
 “Abnormal” “Normal” 20 175 474   
 “Abnormal” “Abnormal” 21 107 329   
Group selectedGenetic dataaNo. of casesPostoperative survivalStatistical analysis
Rb1 pathwayPTENMedianLongest survivalWilcoxon-GehanCox’s regressionb
All cases (n = 125) “Normal” “Normal” 11 437 >2738 P < 0.001 P = 0.012 
 “Normal” “Abnormal” 12 252 914   
 “Abnormal” “Normal” 54 279 1823   
 “Abnormal” “Abnormal” 48 166 888   
Cases with survival <30 days excluded (n = 120) “Normal” “Normal” 11 437 >2738 P = 0.001 P = 0.011 
 “Normal” “Abnormal” 12 252 914   
 “Abnormal” “Normal” 52 282 1823   
 “Abnormal” “Abnormal” 45 179 888   
“Primary” GB (n = 119) “Normal” “Normal” 10 600 >2738 P = 0.002 P = 0.013 
 “Normal” “Abnormal” 12 252 914   
 “Abnormal” “Normal” 50 270 729   
 “Abnormal” “Abnormal” 47 163 888   
Cases with postoperative RT (n = 66) “Normal” “Normal” 841 >2449 P = 0.021 P = 0.031 
 “Normal” “Abnormal” 489 914   
 “Abnormal” “Normal” 30 365 1165   
 “Abnormal” “Abnormal” 24 301 888   
Cases not given RT (n = 50) “Normal” “Normal” 318 >2738 P = 0.022 P = 0.132 
 “Normal” “Abnormal” 187 232   
 “Abnormal” “Normal” 20 175 474   
 “Abnormal” “Abnormal” 21 107 329   
a

For definitions, see “Materials and Methods.”

b

Cox’s regression analysis performed to adjust for the age factor.

Table 5

Cox’s multivariate regression analysis with respect to survival

FactorWald statisticdfRHaCIbPs
Age10c 29.0 1.71 1.41–2.08 <0.001 
RTd 10.0 0.51 0.34–0.77 0.002 
Sidee 8.4 1.81 1.21–2.69 0.004 
Rb1 pathwayf 4.3 1.52 1.02–2.26 0.036 
PTEN                  g 1.7 1.30 0.88–1.93 0.189 
FactorWald statisticdfRHaCIbPs
Age10c 29.0 1.71 1.41–2.08 <0.001 
RTd 10.0 0.51 0.34–0.77 0.002 
Sidee 8.4 1.81 1.21–2.69 0.004 
Rb1 pathwayf 4.3 1.52 1.02–2.26 0.036 
PTEN                  g 1.7 1.30 0.88–1.93 0.189 
a

RH, relative hazards.

b

CI, confidence interval.

c

Increasing age, per 10-year increment.

d

Postoperative radiotherapy given vs. no radiotherapy.

e

Right vs. left sided tumor.

f

Rb1 pathway “abnormal” vs. “normal” (as defined in “Materials and Methods”).

g

PTEN “abnormal” vs. “normal” (as defined in “Materials and Methods”).

Table 6

Clinical and genetic features of “primary” and “secondary” glioblastomaa

“Primary GB”“Secondary GB”
Duration of symptoms (months)b  <3 3–21 16–363 
Number of cases  78 44  8 
Age (years) Mean 56 58 39 
Survival (days) Median 247 223 339 
Frequency of gene abnormalitya Rb1 pathwaya 64/77 (83%) 35/44 (80%) 7/8 (88%) 
 CDKN2A 40/78 (51%) 24/44 (55%) 5/8 (63%) 
 CDKN2B 39/78 (50%) 24/44 (55%) 3/8 (38%) 
 RB1 10/77 (13%) 6/42 (14%) 1/7 (14%) 
 CDK4 14/78 (18%) 5/44 (11%) 1/8 (13%) 
 p53 pathwaya 68/78 (87%) 39/44 (89%) 8/8 (100%) 
 TP53 26/78 (33%) 15/44 (34%) 5/8 (63%) 
 MDM2 8/78 (10%) 4/44 (9%) 1/8 (13%) 
 p14                  ARF 42/78 (54%) 24/44 (55%) 4/8 (50%) 
 EGFR 29/76 (38%) 19/44 (43%) 0/7 (0%) 
 PTEN 39/77 (51%) 20/42 (48%) 1/7 (14%) 
“Primary GB”“Secondary GB”
Duration of symptoms (months)b  <3 3–21 16–363 
Number of cases  78 44  8 
Age (years) Mean 56 58 39 
Survival (days) Median 247 223 339 
Frequency of gene abnormalitya Rb1 pathwaya 64/77 (83%) 35/44 (80%) 7/8 (88%) 
 CDKN2A 40/78 (51%) 24/44 (55%) 5/8 (63%) 
 CDKN2B 39/78 (50%) 24/44 (55%) 3/8 (38%) 
 RB1 10/77 (13%) 6/42 (14%) 1/7 (14%) 
 CDK4 14/78 (18%) 5/44 (11%) 1/8 (13%) 
 p53 pathwaya 68/78 (87%) 39/44 (89%) 8/8 (100%) 
 TP53 26/78 (33%) 15/44 (34%) 5/8 (63%) 
 MDM2 8/78 (10%) 4/44 (9%) 1/8 (13%) 
 p14                  ARF 42/78 (54%) 24/44 (55%) 4/8 (50%) 
 EGFR 29/76 (38%) 19/44 (43%) 0/7 (0%) 
 PTEN 39/77 (51%) 20/42 (48%) 1/7 (14%) 
a

For definitions, see “Materials and Methods.”

b

Time from debut of any symptoms attributable to the disease to operation of GB.

We thank the Departments of Neurosurgery and Pathology at Karolinska and Sahlgrenska University Hospitals and the many other hospitals in Sweden who helped us obtain follow-up data.

1
Curran W. J., Jr, Scott C. B., Horton J., Nelson J. S., Weinstein A. S., Fischbach A. J., Chang C. H., Rotman M., Asbell S. O., Krisch R. E., et al Recursive partitioning analysis of prognostic factors in three Radiation Therapy Oncology Group malignant glioma trials.
J. Natl. Cancer Inst. (Bethesda)
,
85
:
704
-710,  
1993
.
2
Barker F. G., II, Prados M. D., Chang S. M., Gutin P. H., Lamborn K. R., Larson D. A., Malec M. K., McDermott M. W., Sneed P. K., Wara W. M., Wilson C. B. Radiation response and survival time in patients with glioblastoma multiforme.
J. Neurosurg.
,
84
:
442
-448,  
1996
.
3
Burger P. C., Green S. B. Patient age, histologic features, and length of survival in patients with glioblastoma multiforme.
Cancer (Phila.)
,
59
:
1617
-1625,  
1987
.
4
Salminen E., Nuutinen J. M., Huhtala S. Multivariate analysis of prognostic factors in 106 patients with malignant glioma.
Eur. J. Cancer
,
32A
:
1918
-1923,  
1996
.
5
Bleehen N. M., Stenning S. P. A Medical Research Council trial of two radiotherapy doses in the treatment of grades 3 and 4 astrocytoma. The Medical Research Council Brain Tumour Working Party.
Br. J. Cancer
,
64
:
769
-774,  
1991
.
6
Walker M. D., Strike T. A., Sheline G. E. An analysis of dose-effect relationship in the radiotherapy of malignant gliomas.
Int. J. Radiat. Oncol. Biol. Phys.
,
5
:
1725
-1731,  
1979
.
7
Burton E. C., Lamborn K. R., Forsyth P., Scott J., O’Campo J., Uyehara-Lock J., Prados M., Berger M., Passe S., Uhm J., O’Neill B. P., Jenkins R. B., Aldape K. D. Aberrant p53, mdm2, and proliferation differ in glioblastomas from long-term compared with typical survivors.
Clin. Cancer Res.
,
8
:
180
-187,  
2002
.
8
Galanis E., Buckner J., Kimmel D., Jenkins R., Alderete B., O’Fallon J., Wang C. H., Scheithauer B. W., James C. D. Gene amplification as a prognostic factor in primary and secondary high-grade malignant gliomas.
Int. J. Oncol.
,
13
:
717
-724,  
1998
.
9
James C. D., Galanis E., Frederick L., Kimmel D. W., Cunningham J. M., Atherton-Skaff P. J., O’Fallon J. R., Jenkins R. B., Buckner J. C., Hunter S. B., Olson J. J., Scheithauer B. W. Tumor suppressor gene alterations in malignant gliomas: histopathological associations and prognostic evaluation.
Int. J. Oncol.
,
15
:
547
-553,  
1999
.
10
Kraus J. A., Glesmann N., Beck M., Krex D., Klockgether T., Schackert G., Schlegel U. Molecular analysis of the PTEN, TP53 and CDKN2A tumor suppressor genes in long-term survivors of glioblastoma multiforme.
J. Neuro-Oncol.
,
48
:
89
-94,  
2000
.
11
Newcomb E. W., Bhalla S. K., Parrish C. L., Hayes R. L., Cohen H., Miller D. C. bcl-2 protein expression in astrocytomas in relation to patient survival and p53 gene status.
Acta Neuropathol. (Berl)
,
94
:
369
-375,  
1997
.
12
Olson J. J., Barnett D., Yang J., Assietti R., Cotsonis G., James C. D. Gene amplification as a prognostic factor in primary brain tumors.
Clin. Cancer Res.
,
4
:
215
-222,  
1998
.
13
Simmons M. L., Lamborn K. R., Takahashi M., Chen P., Israel M. A., Berger M. S., Godfrey T., Nigro J., Prados M., Chang S., Barker F. G., II, Aldape K. Analysis of complex relationships between age, p53, epidermal growth factor receptor, and survival in glioblastoma patients.
Cancer Res.
,
61
:
1122
-1128,  
2001
.
14
Smith J. S., Tachibana I., Passe S. M., Huntley B. K., Borell T. J., Iturria N., O’Fallon J. R., Schaefer P. L., Scheithauer B. W., James C. D., Buckner J. C., Jenkins R. B. PTEN mutation, EGFR amplification, and outcome in patients with anaplastic astrocytoma and glioblastoma multiforme.
J. Natl. Cancer Inst. (Bethesda)
,
93
:
1246
-1256,  
2001
.
15
Zhou X. P., Li Y. J., Hoang-Xuan K., Laurent-Puig P., Mokhtari K., Longy M., Sanson M., Delattre J. Y., Thomas G., Hamelin R. Mutational analysis of the PTEN gene in gliomas: molecular and pathological correlations.
Int. J. Cancer
,
84
:
150
-154,  
1999
.
16
Raffel C., Frederick L., O’Fallon J. R., Atherton-Skaff P., Perry A., Jenkins R. B., James C. D. Analysis of oncogene and tumor suppressor gene alterations in pediatric malignant astrocytomas reveals reduced survival for patients with PTEN mutations.
Clin. Cancer Res.
,
5
:
4085
-4090,  
1999
.
17
Collins V. P. Gene amplification in human gliomas.
Glia
,
15
:
289
-296,  
1995
.
18
He J., Olson J. J., James C. D. Lack of p16INK4 or retinoblastoma protein (pRb), or amplification-associated overexpression of cdk4 is observed in distinct subsets of malignant glial tumors and cell lines.
Cancer Res.
,
55
:
4833
-4836,  
1995
.
19
Ichimura K., Bolin M. B., Goike H. M., Schmidt E. E., Moshref A., Collins V. P. Deregulation of the p14ARF/MDM2/p53 pathway is a prerequisite for human astrocytic gliomas with G1-S transition control gene abnormalities.
Cancer Res.
,
60
:
417
-424,  
2000
.
20
Reifenberger G., Liu L., Ichimura K., Schmidt E. E., Collins V. P. Amplification and overexpression of the MDM2 gene in a subset of human malignant gliomas without p53 mutations.
Cancer Res.
,
53
:
2736
-2739,  
1993
.
21
Schmidt E. E., Ichimura K., Goike H. M., Moshref A., Liu L., Collins V. P. Mutational profile of the PTEN gene in primary human astrocytic tumors and cultivated xenografts.
J. Neuropathol. Exp. Neurol.
,
58
:
1170
-1183,  
1999
.
22
Rasheed B. K., McLendon R. E., Herndon J. E., Friedman H. S., Friedman A. H., Bigner D. D., Bigner S. H. Alterations of the TP53 gene in human gliomas.
Cancer Res.
,
54
:
1324
-1330,  
1994
.
23
Ueki K., Ono Y., Henson J. W., Efird J. T., von Deimling A., Louis D. N. CDKN2/p16 or RB alterations occur in the majority of glioblastomas and are inversely correlated.
Cancer Res.
,
56
:
150
-153,  
1996
.
24
Buschges R., Weber R. G., Actor B., Lichter P., Collins V. P., Reifenberger G. Amplification and expression of cyclin D genes (CCND1, CCND2 and CCND3) in human malignant gliomas.
Brain Pathol.
,
9
:
435
-442, discussion 432–433 
1999
.
25
Costello J. F., Plass C., Arap W., Chapman V. M., Held W. A., Berger M. S., Su Huang H. J., Cavenee W. K. Cyclin-dependent kinase 6 (CDK6) amplification in human gliomas identified using two-dimensional separation of genomic DNA.
Cancer Res.
,
57
:
1250
-1254,  
1997
.
26
Riemenschneider M. J., Buschges R., Wolter M., Reifenberger J., Bostrom J., Kraus J. A., Schlegel U., Reifenberger G. Amplification and overexpression of the MDM4 (MDMX) gene from 1q32 in a subset of malignant gliomas without TP53 mutation or MDM2 amplification.
Cancer Res.
,
59
:
6091
-6096,  
1999
.
27
Smith J. S., Wang X. Y., Qian J., Hosek S. M., Scheithauer B. W., Jenkins R. B., James C. D. Amplification of the platelet-derived growth factor receptor-A (PDGFRA) gene occurs in oligodendrogliomas with grade IV anaplastic features.
J. Neuropathol. Exp. Neurol.
,
59
:
495
-503,  
2000
.
28
Westermark B., Heldin C. H., Nister M. Platelet-derived growth factor in human glioma.
Glia
,
15
:
257
-263,  
1995
.
29
Sherr C. J. The Pezcoller lecture: cancer cell cycles revisited.
Cancer Res.
,
60
:
3689
-3695,  
2000
.
30
Weinberg R. A. The retinoblastoma protein and cell cycle control.
Cell
,
81
:
323
-330,  
1995
.
31
Ichimura K., Schmidt E. E., Goike H. M., Collins V. P. Human glioblastomas with no alterations of the CDKN2A (p16INK4A, MTS1) and CDK4 genes have frequent mutations of the retinoblastoma gene.
Oncogene
,
13
:
1065
-1072,  
1996
.
32
Ekstrand A. J., James C. D., Cavenee W. K., Seliger B., Pettersson R. F., Collins V. P. Genes for epidermal growth factor receptor, transforming growth factor α, and epidermal growth factor and their expression in human gliomas in vivo.
Cancer Res.
,
51
:
2164
-2172,  
1991
.
33
Liu L., Ichimura K., Pettersson E. H., Goike H. M., Collins V. P. The complexity of the 7p12 amplicon in human astrocytic gliomas: detailed mapping of 246 tumors.
J. Neuropathol. Exp. Neurol.
,
59
:
1087
-1093,  
2000
.
34
Sauter G., Maeda T., Waldman F. M., Davis R. L., Feuerstein B. G. Patterns of epidermal growth factor receptor amplification in malignant gliomas.
Am. J. Pathol.
,
148
:
1047
-1053,  
1996
.
35
Kleihues P., Cavanee W. K. .
WHO Classification of Tumours, Pathology and Genetics of the Nervous system
, IARC Press  
2000
.
36
Liu L., Ichimura K., Pettersson E. H., Collins V. P. Chromosome 7 Rearrangements in Glioblastomas; Loci Adjacent to EGFR Are Independently Amplified.
J. Neuropathol. Exp. Neurol.
,
57
:
1138
-1145,  
1998
.
37
Levine A. J. p53, the cellular gatekeeper for growth and division.
Cell
,
88
:
323
-331,  
1997
.
38
Gehan E. A. A generalized two-sample Wilcoxon test for doubly censored data.
Biometrika
,
52
:
650
-653,  
1965
.
39
Peto R., Pike M. C., Armitage P., Breslow N. E., Cox D. R., Howard S. V., Mantel N., McPherson K., Peto J., Smith P. G. Design and analysis of randomized clinical trials requiring prolonged observation of each patient. II. Analysis and examples.
Br. J. Cancer
,
35
:
1
-39,  
1977
.
40
Morgenbesser S. D., Williams B. O., Jacks T., DePinho R. A. p53-dependent apoptosis produced by Rb-deficiency in the developing mouse lens.
Nature (Lond.)
,
371
:
72
-74,  
1994
.
41
Debbas M., White E. Wild-type p53 mediates apoptosis by E1A, which is inhibited by E1B.
Genes Dev.
,
7
:
546
-554,  
1993
.
42
Wu X., Levine A. J. p53 and E2F-1 cooperate to mediate apoptosis.
Proc. Natl. Acad. Sci. USA
,
91
:
3602
-3606,  
1994
.
43
Kurose K., Gilley K., Matsumoto S., Watson P. H., Zhou X. P., Eng C. Frequent somatic mutations in PTEN and TP53 are mutually exclusive in the stroma of breast carcinomas.
Nat. Genet.
,
32
:
355
-357,  
2002
.
44
Kato H., Kato S., Kumabe T., Sonoda Y., Yoshimoto T., Han S. Y., Suzuki T., Shibata H., Kanamaru R., Ishioka C. Functional evaluation of p53 and PTEN gene mutations in gliomas.
Clin. Cancer Res.
,
6
:
3937
-3943,  
2000
.
45
Bauman G. S., Ino Y., Ueki K., Zlatescu M. C., Fisher B. J., Macdonald D. R., Stitt L., Louis D. N., Cairncross J. G. Allelic loss of chromosome 1p and radiotherapy plus chemotherapy in patients with oligodendrogliomas.
Int. J. Radiat. Oncol. Biol. Phys.
,
48
:
825
-830,  
2000
.
46
Cairncross J. G., Ueki K., Zlatescu M. C., Lisle D. K., Finkelstein D. M., Hammond R. R., Silver J. S., Stark P. C., Macdonald D. R., Ino Y., Ramsay D. A., Louis D. N. Specific genetic predictors of chemotherapeutic response and survival in patients with anaplastic oligodendrogliomas.
J. Natl. Cancer Inst. (Bethesda)
,
90
:
1473
-1479,  
1998
.
47
Watanabe K., Tachibana O., Sata K., Yonekawa Y., Kleihues P., Ohgaki H. Overexpression of the EGF receptor and p53 mutations are mutually exclusive in the evolution of primary and secondary glioblastomas.
Brain Pathol.
,
6
:
217
-223, discussion 223–224 
1996
.
48
Reifenberger J., Ring G. U., Gies U., Cobbers L., Oberstrass J., An H. X., Niederacher D., Wechsler W., Reifenberger G. Analysis of p53 mutation and epidermal growth factor receptor amplification in recurrent gliomas with malignant progression.
J. Neuropathol. Exp. Neurol.
,
55
:
822
-831,  
1996
.
49
Tohma Y., Gratas C., Biernat W., Peraud A., Fukuda M., Yonekawa Y., Kleihues P., Ohgaki H. PTEN (MMAC1) mutations are frequent in primary glioblastomas (de novo) but not in secondary glioblastomas.
J. Neuropathol. Exp. Neurol.
,
57
:
684
-689,  
1998
.
50
Biernat W., Kleihues P., Yonekawa Y., Ohgaki H. Amplification and overexpression of MDM2 in primary (de novo) glioblastomas.
J. Neuropathol. Exp. Neurol.
,
56
:
180
-185,  
1997
.
51
Biernat W., Tohma Y., Yonekawa Y., Kleihues P., Ohgaki H. Alterations of cell cycle regulatory genes in primary (de novo) and secondary glioblastomas.
Acta Neuropathol. (Berl)
,
94
:
303
-309,  
1997
.
52
Nakamura M., Watanabe T., Klangby U., Asker C., Wiman K., Yonekawa Y., Kleihues P., Ohgaki H. p14ARF deletion and methylation in genetic pathways to glioblastomas.
Brain Pathol.
,
11
:
159
-168,  
2001
.