The first two authors contributed equally to this article.
Prognostic impact of TP53 mutation status for adult patients with supratentorial World Health Organization Grade II astrocytoma or oligoastrocytoma
A long-term analysis
Article first published online: 7 JUL 2004
Copyright © 2004 American Cancer Society
Volume 101, Issue 5, pages 1028–1035, 1 September 2004
How to Cite
Ständer, M., Peraud, A., Leroch, B. and Kreth, F. W. (2004), Prognostic impact of TP53 mutation status for adult patients with supratentorial World Health Organization Grade II astrocytoma or oligoastrocytoma. Cancer, 101: 1028–1035. doi: 10.1002/cncr.20432
- Issue published online: 18 AUG 2004
- Article first published online: 7 JUL 2004
- Manuscript Accepted: 13 MAY 2004
- Manuscript Revised: 10 APR 2004
- Manuscript Received: 31 OCT 2003
- Deutsche Forschungsgemeinschaft. Grant Number: PE 758/1-1
- Deutsche Krebshilfe
- low-grade glioma;
- genetic alteration;
The goal of the current study was to retrospectively assess the prognostic impact of TP53 mutation status and P53 expression/accumulation on long-term outcome for adult patients with supratentorial World Health Organization (WHO) Grade II astrocytoma or oligoastrocytoma.
The authors revisited a previously published short-term data set containing information on 159 consecutive patients who were treated between 1991 and 1998. Each patient was screened for TP53 mutations and P53 overexpression/accumulation. The reference point for all analyses was the date of surgical treatment, and the date of last follow-up examination was August 2002. Overall survival, progression-free survival, postrecurrence survival, and time to malignant transformation were estimated using the Kaplan–Meier method, and potential prognostic factors were evaluated using the multivariate proportional hazards model.
The median follow-up duration for survivors was 80.4 months (standard deviation, 33.0 months). TP53 mutations, which were present in 49.1% of all tumors, occurred preferentially in gemistocytic tumors (P < 0.05). In addition, the TP53 status of the primary tumor was predictive of the TP53 status of the recurrent tumor in all cases of disease recurrence. The 5-year overall and progression-free survival rates were 77.5% and 43.2%, respectively, and the risk of malignant transformation at 5 years postsurgery was 32.7%. Unfavorable prognostic factors with respect to survival duration included older age (≥ 50 years; P < 0.002), gemistocytic subtype (P < 0.01), and positive TP53 mutation status (P < 0.05), all of which were also negatively associated with progression-free survival (P < 0.05, P < 0.001, and P < 0.003, respectively). In contrast, positive TP53 mutation status was the only significant predictor of a reduction in time to malignant transformation (P < 0.03). P53 overexpression/accumulation did not exhibit prognostic relevance in any of the multivariate models constructed in the current study.
TP53 mutations are common early events in the pathogenesis of WHO Grade II astrocytoma or oligoastrocytoma. In the current study, positive TP53 mutation status (but not P53 overexpression/accumulation) was found to be an independent unfavorable predictor of survival, progression-free survival, and time to malignant transformation. The therapeutic implications of these findings have yet to be determined. Cancer 2004. © 2004 American Cancer Society.
Uncertainty remains regarding the prognostic significance of TP53 mutations for adult patients with supratentorial World Health Organization (WHO) Grade II astrocytoma or oligoastrocytoma. The limited size and heterogeneous nature of existing study cohorts, the lack of standardized treatment strategies, differences in techniques for assessing TP53 mutation status, and differences in histologic classification systems all represent major obstacles in the retrospective evaluation of available data.1–13 In the largest known series to date (which was recently reported on by our group), each tumor was reviewed for histologic type according to the WHO 2000 classification system, and TP53 status was analyzed using direct DNA sequencing.14 That study failed to reveal any overall prognostic impact associated with TP53 status, and only a highly selected subpopulation (patients with a hotspot mutation at codon 175 of the TP53 gene) was found to have fared significantly poorer in terms of time to progression and time to malignant transformation. Nonetheless, due to the relatively short duration of follow-up (median, 49 months) in that study, the possibility that TP53 mutation status had late effects on overall prognosis could not be ruled out. Furthermore, the short-term analysis of prognostic factors in that study was based on a relatively small number of events, and thus, its statistical power was limited.
To rectify these shortcomings, in August 2002, we revisited the data set used in that previous study. The original large, homogeneously treated population of adults with supratentorial WHO Grade II astrocytoma or oligoastrocytoma was reevaluated, with the median follow-up duration having increased to > 80 months.
MATERIALS AND METHODS
Inclusion Parameters, Treatment Protocol, and Patient Follow-Up
One hundred fifty-nine patients with supratentorial WHO Grade II astrocytoma or oligoastrocytoma (de novo in 129 cases) treated at the Department of Neurosurgery, Klinikum Grosshadern, Ludwig-Maximilians-Universität (Munich, Germany), between January 1991 and December 1998 were included in the current analysis. All tumor specimens were histologically reviewed at the German National Brain Tumor Reference Center (Bonn, Germany) in accordance with the WHO 2000 classification system, as has been described previously.14 Patient demographic data are summarized in Table 1. Microsurgical resection was the most commonly used treatment method. For the 25 patients with surgically inaccessible tumors, however, stereotactic biopsy (BX) was used to establish histopathologic diagnoses. Extent of resection was determined on the basis of the intraoperative opinion of the surgeon and/or on the basis of any relevant description in the surgical records. Conventionally fractionated limited-field radiotherapy (tumor dose, 50–60 grays) was used as second-line therapy for patients with documented tumor progression or malignant transformation. Only five patients with large residual tumor masses after subtotal resection (STR) and seven patients who underwent BX were treated with early external-beam radiotherapy. In addition, eight patients with eloquently located, small, circumscribed tumors underwent interstitial radiosurgery with iodine-125.
|Primary/de novo (n = 129)||Recurrent (n = 30)||All types (n = 159)|
|Age (yrs)a||36.9 ± 12.4||39.6 ± 11.9||37.4 ± 12.3|
|Gender (M:F ratio)||1:0.93||1:2.75||1:1.12|
|Duration of symptoms (mos)a||23.0 ± 45.5||39.6 ± 11.9||27.5 ± 45.5|
|Epileptic seizures (%)||107 (82.9)||26 (86.7)||133 (83.6)|
|Focal neurologic defects (%)||81 (62.8)||22 (73.3)||103 (64.8)|
|Signs of increased intracranial pressure (%)||37 (28.7)||9 (30.0)||46 (28.9)|
|GTR||64 (49.6)||10 (33.3)||74 (46.6)|
|STR||44 (34.1)||16 (53.4)||60 (37.7)|
|BX||21 (16.3)||4 (13.3)||25 (15.7)|
After treatment, clinical evaluation and magnetic resonance imaging (MRI) were performed at 6-month intervals and at the time of the last follow-up examination (August 2002). Any tumor growth observed on MRI after macroscopic (‘gross’) total resection (GTR) was considered indicative of recurrent disease, and any increase of > 25% in tumor volume as assessed on MRI after STR or BX was considered indicative of progressive disease. Malignant transformation was considered to have occurred 1) when Grade III/IV disease was histologically diagnosed after surgery or BX or 2) when multilocular tumor appearance or contrast enhancement of an initially nonenhancing lesion was observed in combination with rapid tumor growth.
Detection of TP53 Mutations and P53 Overexpression/Accumulation
The techniques used to detect TP53 mutations and P53 overexpression/accumulation are described in detail elsewhere.14 In brief, tumor samples from each patient were screened for TP53 mutations in exons 5–8 using polymerase chain reaction–single-strand conformational polymorphism analysis, and samples exhibiting mobility shift were further analyzed by direct DNA sequencing. In addition, P53 expression/accumulation was assessed using immunochemical methods. The monoclonal immunoglobulin G2b mouse anti-human P53 antibody M7001 (Dako, Glostrup, Denmark), which binds to both wild-type and mutated P53, was used in these immunochemical analyses.
The reference point for all statistical analyses was the date of first surgical treatment. Clinical and neuroradiologic evaluations were last performed in August 2002. Study endpoints included death, tumor recurrence or progression, and malignant transformation. Progression-free survival was defined as the time to tumor progression after STR or BX or as the time to tumor recurrence after GTR. Postrecurrence survival was defined as the interval between tumor recurrence or progression and death or last follow-up. Overall survival, progression-free survival, postrecurrence survival, and time to malignant transformation were analyzed using the Kaplan–Meier method.15 Potential prognostic factors were identified using a proportional hazards model,16 and correlations with outcome were assessed using the χ2 test. The prognostic significance of each covariate was initially tested on univariate analysis, and subsequent multivariate analysis was performed whenever possible (i.e., when the number of events was sufficient). Alternative models were compared by computing maximized likelihoods. The optimized model contained only variables that were significantly associated with the endpoint of interest after adjustment for the effects of the other variables included in the model. Each optimized model was validated by verifying that no term in the model could be added or omitted without significantly altering the maximized likelihood. The following variables were tested: age (≥ 50 years vs. < 50 years [categoric] or continuously scaled), gender, Karnofsky performance status (KPS; ≥ 80 vs. < 80), extent of resection (GTR vs. STR/BX), histologic subtype (gemistocytic vs. other; astrocytoma vs. oligoastrocytoma), TP53 mutation status (positive vs. negative), and P53 overexpression/accumulation (yes vs. no). Due to the limited number of events, the prognostic impact of mutations found in hotspot codons could be tested only on univariate analysis. All risk ratios are presented together with P values and 95% confidence intervals.
The current series comprised 159 consecutive patients with supratentorial WHO Grade II astrocytoma or oligoastrocytoma (including 129 patients with de novo tumors). The median follow-up duration for survivors was 80.4 months (standard deviation [SD], 33.0 months), and the average patient age was 37.4 years (SD, 12.3 years) (Table 1). GTR was performed for 74 patients (46.6%), STR was performed for 60 (37.7%), and BX was performed for 25 (15.7%). Estimates of overall survival, progression-free survival, and time to malignant transformation were based on 70, 113, and 66 events, respectively. The calculated 5-year overall survival rate was 77.5%, the 5-year progression-free survival rate was 43.2%, and the risk of malignant transformation at 5 years postsurgery was 32.7%; for the 129 patients with de novo tumors, the corresponding rates were 76.0%, 44.7%, and 35.5%, respectively (Fig. 1). In addition, for the study cohort as a whole, the median postrecurrence survival duration was 32 months.
The frequencies and patterns of occurrence associated with TP53 mutations have been described previously and are summarized in Table 2.14 In brief, TP53 mutations were found in 49.0% of the overall study cohort (73 of 149) and in 43.4% of patients with de novo tumors (53 of 122). In seven cases, two mutations were found in a single tumor, and in two cases, three mutations were found in a single tumor. Thirty-six mutations were found in hotspot codons (codon 175 [n = 5], codon 248 [n = 4], or codon 273 [n = 27]). Among the 84 mutations detected in the current study cohort, there were 3 deletions with ensuing frameshift and 81 missense mutations. Relative to other histologic subtypes, the gemistocytic subtype was significantly more frequently associated with the presence of TP53 mutations (P < 0.05). Development of new TP53 mutations at the time of progression or recurrence (n = 113) or at the time of malignant transformation (n = 66) was not noted in any of the investigated cases; in other words, the TP53 status of the primary tumor was predictive of the TP53 status of the recurrent, progressive, or newly malignant tumor in all cases.
|Event/variable||No. of occurrences (%)|
|TP53 mutation||73 (49.0)|
|In de novo tumors||53 (43.4a)|
|In recurrent tumors||20 (66.6b)|
|No. of mutations per tumor|
|Total no. of mutations||84|
|P53 overexpression/accumulation||70 (46.8)|
|In tumors with TP53 mutations||68 (93.1c)|
|In tumors without TP53 mutations||2 (2.7d)|
P53 overexpression/accumulation was noted in 46.8% of all tumors and in 91.8% of tumors that were positive for TP53 mutations (Table 2).
Prognostic Factors Influencing Overall Survival
Unfavorable predictors of overall survival on univariate analysis included age > 50 years (P < 0.003), positive TP53 mutation status (P < 0.015), and gemistocytic subtype (P < 0.003); the prognostic significance of P53 overexpression/accumulation (P = 0.064) was less noteworthy. Nonsignificant variables included KPS, gender, occurrence of seizures at the time of presentation, tumor location, oligoastrocytic subtype, and extent of resection. In the optimized multivariate model, age > 50 years (risk ratio [RR], 2.4; 95% confidence interval [CI], 1.4–4.2; P < 0.002), gemistocytic subtype (RR, 3.3; 95% CI, 1.4–7.8; P < 0.01), and positive TP53 mutation status (RR, 1.7; 95% CI, 1.0–2.8; P < 0.05) continued to be unfavorable predictors of overall survival. The 5-year survival rate was 82.1% for patients without TP53 mutations and 72.1% for patients with TP53 mutations (P < 0.013) (Fig. 2), and the same survival pattern held true for the subgroup of patients with de novo tumors (data not shown). In contrast, postrecurrence survival rates were comparable for patients with TP53 mutations and patients without TP53 mutations (P > 0.3) (data not shown).
Prognostic Factors Influencing Progression-Free Survival
On univariate analysis, the following variables were significantly associated with poorer progression-free survival: increasing age (continuously scaled; P < 0.04), positive TP53 mutation status (P < 0.001), P53 overexpression/accumulation (P < 0.002), and gemistocytic subtype (P < 0.001). Nonsignificant variables included KPS, gender, occurrence of seizures at the time of presentation, tumor location, oligoastrocytic subtype, and extent of resection. In the optimized multivariate model, increasing age (continuously scaled; RR, 1.01; 95% CI, 1.00–1.04; P < 0.05), gemistocytic subtype (RR, 4.90; 95% CI, 2.3–10.6; P < 0.001), and positive TP53 mutation status (RR, 1.8; 95% CI, 1.2–2.7; P < 0.003) continued to be significant risk factors. In an alternative multivariate model that included P53 expression/accumulation in place of TP53 mutation status, P53 expression/accumulation did not exhibit prognostic significance; in addition, the fit of this alternative model was significantly poorer than the fit of the optimized model. The unadjusted 5-year progression-free survival rates for patients without TP53 mutations and patients with TP53 mutations were 54.3% and 28.7%, respectively (Fig. 2). Similar findings were made in a separate analysis of the 129 patients with de novo tumors (data not shown).
Prognostic Factors Influencing Time to Malignant Transformation
Positive TP53 mutation status was the lone significant risk factor with respect to time to malignant transformation. Although the significance of this risk factor was evident in the overall study cohort (RR, 1.7; 95% CI, 1.1–2.8; P < 0.03), it was even more apparent among patients with de novo tumors (RR, 2.1; 95% CI, 1.2–3.5; P = 0.008). The prognostic influence of P53 protein overexpression/accumulation (P = 0.06) was less noteworthy. Age (dichotomized or continuously scaled), tumor histology, KPS, gender, occurrence of seizures at the time of presentation, tumor location, and extent of resection did not have significant prognostic value. The unadjusted 5-year malignant transformation rate was 23.9% for patients without TP53 mutations and 43.9% for patients with TP53 mutations (Fig. 2). Prognostic factors for each study endpoint are summarized in Table 3.
|Variable||Overall survival||Progression-free survival||Malignant transformation|
|RR (95% CI)||P value||RR (95% CI)||P value||RR (95% CI)||P value|
|Age||2.4 (1.4–4.2)b||< 0.002b||1.01 (1.00–1.04)c||< 0.05c||—||—|
|Histologic subtype (gemistocytic vs. other)||3.3 (1.4–7.8)||< 0.01||4.9 (2.3–10.6)||< 0.001||—||—|
|TP53 mutation status||1.7 (1.0–2.8)||< 0.05||1.8 (1.2–2.7)||< 0.003||1.7 (1.1–2.8)||< 0.03|
Prognostic Influence of Multiple Mutations and Hotspot Codon Mutations
Patients with multiple mutations (n = 9) fared worse than did those with single mutations; however, this difference was not statistically significant for any of the study endpoints (5-year survival rate, 63.5% vs. 73.2% [P > 0.5]; 5-year progression/recurrence-free survival rate, 12.5% vs. 28.5% [P > 0.3]; 5-year malignant transformation rate, 66.6% vs. 42.1% [P > 0.19]). Patients who had hotspot mutations within codon 175 (n = 5) had a significantly higher rate of malignant transformation compared with patients who had mutations that were not located within a hotspot codon (5-year malignant transformation rate, 100% vs. 47%; P < 0.03). In contrast, patients who had other types of hotspot mutations (i.e., mutations within codon 248 or codon 273) were not at an elevated risk relative to other patients with TP53 mutations.
It has been confirmed that TP53 mutations are relatively common early events in the pathogenesis of WHO Grade II astrocytomas and oligoastrocytomas and that these mutations occur preferentially in gemistocytic lesions.2, 11, 12, 14 Nonetheless, the prognostic relevance of these findings has remained unclear, and the current long-term analysis of a large, homogeneously treated population of adult patients with supratentorial WHO Grade II astrocytoma or oligoastrocytoma was undertaken to shed light on this issue. To our knowledge, the current study was the first to assess the prognostic impact of TP53 mutation status on all critical endpoints (i.e., death, tumor progression, and time to malignant transformation) after adjustment for the effects of relevant patient- and tumor-related covariates. For each study endpoint, such an analysis became possible due to the occurrence of a sufficient number of events during the extended follow-up period. Still, the results of the current study should be regarded with caution. For example, the prognostic influence of pretreatment tumor volume could not be analyzed, and objective (i.e., MRI-based) data regarding extent of resection were not available; as a result, the constructed prognostic models should not be considered fully descriptive of the risk factors affecting adult patients with supratentorial WHO Grade II astrocytoma or oligoastrocytoma. Furthermore, investigation of loss of heterozygosity on 1p/19q was not performed in the current study, and consequently, the possibility that the incidence of TP53 mutations was underreported, particularly for patients with oligoastrocytomas, could not be ruled out.
Survival, Progression-Free Survival, and Time to Malignant Transformation
The survival rates obtained in the current study (5-year overall survival rate, 77.5%) were near the top of the range of literature-reported survival rates following microsurgical treatment. Whether this finding was related to the use of aggressive surgical techniques or to favorable patterns of treatment-independent prognostic factors could not be determined within the framework of the current retrospective series.
Long-term data regarding progression-free survival and time to malignant transformation previously were not available in the literature.2, 5–8, 10–13 Kreth et al.17 did report a 5-year malignant transformation rate of 34% after interstitial radiosurgery for patients with supratentorial WHO Grade II astrocytoma or oligoastrocytoma, but the median follow-up period in that study was only ∼40 months. For the current series of homogeneously treated patients (who underwent microsurgery without external-beam radiotherapy), none of the presented Kaplan–Meier curves leveled off beyond the 5-year time point; in addition, only 46 of 159 patients in the current series presented with stable disease at the time of last follow-up. These findings underscore the primarily palliative nature of the surgical/radiotherapeutic treatment strategy that was used.
In accordance with previous studies, the current investigation demonstrated the unfavorable influence of both increased age and gemistocytic subtype.1, 17 Extent of resection was not found to have any prognostic relevance; however, this finding should be regarded with caution, because early postoperative MRI was not stringently performed and subjective intraoperative evaluation of extent of resection (which was performed) has been shown to be unreliable. Although KPS typically is considered to be a powerful prognostic indicator,1, 17 this was not the case in the current series; this lack of observed prognostic significance may be attributable to the fact that most patients had pretreatment KPS scores of ∼80.
Positive TP53 mutation status continued to be an independent unfavorable predictor of overall and progression-free survival after adjustment for the effects of age and gemistocytic subtype in the optimized multivariate model. This finding highlights the importance of long-term follow-up, which increases statistical power (through the occurrence of additional events) and allows detection of the ‘late effects’ of TP53 mutations. For example, in the current series, significant divergence in the survival plots for patients with TP53 mutations and patients without TP53 mutations became evident only beyond the 5-year time point (Fig. 2). Consequently, the significant influence of TP53 mutation status went undetected in our previous report, in which the median follow-up period was 49 months.14
Postrecurrence survival was not associated with TP53 mutation status. In addition, in terms of response to delayed external-beam radiotherapy for progressive or malignantly transformed lesions, patients with TP53 mutations did not differ significantly from patients without TP53 mutations.
Patients with oligoastrocytoma were similar to other patients in terms of both the prevalence of TP53 mutations and overall prognosis. This finding should be regarded with caution, given the limited number of patients with oligoastrocytoma in the current study.
P53 overexpression/accumulation possessed unfavorable prognostic influence only with respect to progression-free survival, and only then on univariate analysis. Multivariate prognostic models that included P53 status did not yield as good a fit as did models that included TP53 mutation status instead. The limitations and inaccuracy associated with immunohistochemical determination of P53 status have been reported previously and were confirmed in the current investigation.
Positive TP53 mutation status (and not P53 overexpression/accumulation) was the lone risk factor with respect to malignant transformation in the current series. This finding, which confirms preliminary data reported by Ishii et al.,2 underscores the significant problem of grade instability in low-grade astrocytoma. Other reports have emphasized the unfavorable prognostic impact of older age (> 40 years) and increased tumor volume1, 17; however, the effects of tumor volume could not be analyzed in the current study, and patient age did not exhibit prognostic relevance with respect to malignant transformation (P = 0.2). Thus, uncertainties remain regarding the prognostic influence of these two covariates.
Prognostic Relevance of Multiple Mutations and Hotspot Codon Mutations
Compared with all other patients who had TP53 mutations, patients who had hotspot mutations located within codon 175 (n = 5) fared significantly worse in terms of time to malignant transformation. Progression-free survival, however, was not significantly poorer among patients who had mutations located within codon 175. (This finding was in disagreement with the results of our previous report.) Given the limited size of this patient subgroup, all of these findings should be regarded with caution.
As in our previous analysis, patients with multiple mutations had poorer outcomes than did patients with single mutations, although this difference was not statistically significant. More data on multiple TP53 mutations must be obtained before a reliable assessment of the prognostic impact of this relatively rare phenomenon can be made.
TP53 mutations are common early events in the pathogenesis of supratentorial WHO Grade II astrocytoma or oligoastrocytoma in adults. The overall prognosis for patients with such malignancies depends strongly on their pretreatment prognostic profiles. In the current study, positive TP53 mutation status was found to be an independent unfavorable predictor of overall survival, progression-free survival, and time to malignant transformation. Other factors associated with reduced overall and progression-free survival include older age and gemistocytic subtype, while the prognostic relevance of hotspot mutations and multiple mutations remains subject to debate. As a final note, we believe that the current analysis also underscores the importance of adequate techniques for the detection of TP53 mutations.
The authors thank Otmar D. Wiestler for performing histologic review and Paul Kleihues for providing helpful advice regarding the technical aspects of TP53 mutation analysis and P53 immunostaining. The authors also thank Julia Winkler for her technical assistance and Mélanie F. Daffner, Konstantinos Sarvanakis, and Magnus Meschede for their assistance with the acquisition of patient data. Finally, the authors thank Dr. Hans-Jürgen Reulen, who has developed a state-of-the-art method for the microsurgical treatment of low-grade glioma, for lending his support to the current study.
- 16Regression models and life tables. J R Stat Soc. 1972; 34: 187–220..