Tumor recurrence and malignant progression of gangliogliomas
Transcrição
Tumor recurrence and malignant progression of gangliogliomas
3355 Tumor Recurrence and Malignant Progression of Gangliogliomas Michael Majores, MD1,3 Marec von Lehe, MD2 Jana Fassunke, PhD1 Johannes Schramm, MD2 Albert J. Becker, MD1 Matthias Simon, MD2 1 Department of Neuropathology, University of Bonn Medical Center, Bonn, Germany. 2 Department of Neurosurgery, University of Bonn Medical Center, Bonn, Germany. 3 Department of Pathology, University of Bonn Medical Center, Bonn, Germany. BACKGROUND. Most gangliogliomas (GGs) are benign tumors, but tumor recurrence and malignant progression are observed in some patients. METHODS. The authors analyzed their experience with 4 recurrent/progressive GGs (World Health Organization [WHO] grade I), 21 tumors with atypical features (WHO grade II), and 5 tumors with anaplastic histologic features (WHO grade III). Histopathologic findings (23 patients) were reviewed. The mean follow-up was 5.9 years (median, 4.5 years; range, 0.5-14.7 years). RESULTS. The 5-year survival rates were only 79% for patients who had tumors with atypical features and 53% for patients who had WHO grade III tumors. Secondary glioblastomas were diagnosed in 5 of 11 patients (45%) who underwent surgery for tumor recurrence. Age at surgery <40 years (P 5 .007) was associated significantly with better overall survival (OS), but it was not associated with better progression-free survival (PFS). Clinical presentation (drug-resistant epilepsy vs all other patients with seizures vs no seizures) was associated significantly with better OS (P 5 .005) and PFS (P < .001). Patients who had extratemporal tumors had a significantly shorter PFS (P 5 .01) but not OS. A complete resection was correlated strongly with both OS (P 5 .002) and PFS (P 5 .001). Neuropathologic examination revealed the presence of a gemistocytic cell component (PFS, P 5 .025), a lack of protein droplets (OS, P 5 .04; PFS, P 5 .05), and focal tumor cell-associated CD34 immunolabeling (OS, P 5 .03) as significant predictors of an adverse clinical course. CONCLUSIONS. The current data supported a 3-tiered GG histopathologic grading system that included an intermediate diagnostic category (atypical GG, WHO grade II). Careful attention to histopathologic findings and clinical parameters usually will identify patients who are at risk for an adverse clinical course. Cancer 2008;113:3355–63. 2008 American Cancer Society. KEYWORDS: ganglioglioma, malignancy, recurrence, neuropathology. Supported by the Deutsche Krebshilfe, the German Glioma Network, the Deutsche Forschungsgemeinschaft (SFB TR3), the German–Israeli collaborative research program of the Bundesministerium fur Bildung und Forschung, the Ministry of Science, and Bonn Forschung. The first and second authors contributed equally to this article. Address for reprints: Michael Majores, MD, Department of Neuropathology, University of Bonn Medical Center, Sigmund-Freud Strasse 25, D-53105 Bonn, Germany; Fax: (011) 49-228287-14331; E-mail: [email protected] Received April 16, 2008; revision received July 6, 2008; accepted July 25, 2008. ª 2008 American Cancer Society G angliogliomas (GGs) are glioneuronal tumors composed of neoplastic glial and dysplastic neuronal elements. Both cell populations show marked heterogeneity, ranging from a predominantly neuronal phenotype to variants with a prominent glial population. GGs are characterized by a high epileptogenic potency. GGs are the most common tumors encountered in young patients who have chronic temporal lobe epilepsy.1-3 Histopathologic differential diagnoses comprise both high-grade and low-grade neoplasms, such as diffuse astrocytomas, oligodendrogliomas, dysembryoplastic neuroepithelial tumors, pilocytic astrocytomas (PAs) and pleomorphic xanthoastrocytomas (PXAs).4-8 The pathogenesis of GG is not very well understood. Individual cases have been described in association with a family history of neurofibromatosis,9 in a patient with Peutz-Jeghers syndrome,10 and DOI 10.1002/cncr.23965 Published online 5 November 2008 in Wiley InterScience (www.interscience.wiley.com). 3356 CANCER December 15, 2008 / Volume 113 / Number 12 in a patient with Turcot syndrome.11 An anaplastic GG was detected in the Eker rat, a model for human tuberous sclerosis carrying a mutation of the tuberous-sclerosis complex 2 (TSC2) gene.12 However, patients with epilepsy who have GG usually lack the typical clinical stigmata of tuberous sclerosis.13 TSC gene mutations have been identified in a few GGs.14 It has been proposed that malignant GGs have a monoclonal origin, suggestive of an initial transformation of a single neuroglial precursor cell with subsequent malignant progression.15 Loss of p19 expression and p53 mutations also may be associated with malignant progression.15,16 Several studies have suggested a very benign clinical course in most patients with GG. However, tumor recurrence, malignant progression, and secondary glioblastoma multiforme (GBM) are observed in some patients.17-20 Malignant transformation has been associated with incomplete tumor resections18 and radiation therapy.21 Overall, only sparse information is available with respect to patients who had tumors with an adverse clinical course. Therefore, for the current study, we analyzed our experience with recurrent and/or histologically atypical or anaplastic GG. Our results support a 3-tiered GG histopathologic grading system that includes an intermediate diagnostic category (atypical GG; World Health Organization [WHO] grade II). Our data also indicate that an adverse clinical course in a patient with GG often can be predicted based on histopathologic findings and clinical data, such as age, clinical presentation, tumor location, and degree of resection. MATERIALS AND METHODS Patients and Clinical Data In a series of 203 patients with GG who underwent surgery in the Department of Neurosurgery at the University of Bonn from 1992 to 2006 (excluding brainstem and spinal cord tumors), we identified 5 patients with anaplastic GG, WHO grade III (2.5%); 4 patients with recurrent/progressive GG, WHO grade I (2%); and 21 patients (10.3%) who had tumors with atypical histologic characteristics, such as increased cellularity and increased mitotic activity, but that lacked unequivocal anaplastic features.22 These latter tumors were diagnosed as ‘atypical GG, WHO grade II.’ Ten patients from the current series were included in the large collection of mostly benign GGs reported by Luyken and coworkers.19 Our series did not include any patients with newly diagnosed GG who had WHO grade IV (glioblastomatous) changes in the glial component.23 We cannot exclude the pos- sibility that some patients with recurrent WHO grade I GG may have presented at another institution. However, this does not seem very likely. A previous report from our institution that covered the years 1988 through 2001 with a median follow-up of 8 years included only 1 recurrent tumor in a total of 171 WHO grade I GGs.19 Informed, written consent was obtained from all patients. The respective patients’ charts, operative notes, preoperative and postoperative radiology reports, and/or magnetic resonance images (MRIs) were reviewed. Follow-up data were obtained through telephone interviews if necessary. The mean follow-up was 5.9 years (median, 4.5 years; range, 0.5-14.7 years). Routine follow-up consisted of clinical and MRI examinations every year for patients with WHO grade I GG, twice a year for patients with WHO grade II GG, and every 4 months for patients with WHO grade III GG. Tumor recurrence/progression was defined as radiographic evidence of tumor regrowth with or without clinical signs. Tumors were resected by using standard supratentorial approaches. In patients who had drugresistant epilepsy, the resection always included a 0.5 to 1 cm rim of adjacent cortical tissue; frequently included an adjoining ictogenic cortical area; and, in some patients who underwent temporal resection, it included the hippocampus and the amygdala, as indicated by the preoperative workup or intraoperative corticography. Multiple subpial transections were used in 1 patient with a left temporo-occipital GG.24 Adjuvant therapy consisted of fractionated radiotherapy in 2 patients and radiotherapy and chemotherapy with combined procarbazine, lomustine, and vincristine in 1 of 5 patients with WHO grade III tumors. No adjuvant therapies were prescribed for patients with WHO grade I or II GG. All but 3 of 13 patients with recurrent tumors underwent surgery. Treatment at first tumor recurrence included radiotherapy in all patients who had histologically proven, secondary GBM and in 1 patient who had a recurrent WHO grade II tumor. Tumor recurrence after radiotherapy was treated with chemotherapy (temozolomide in 5 patients, nitrosourea-based protocols in 3 patients). Neuropathology Surgical specimens were fixed in formaldehyde overnight and embedded in paraffin. Specimens were reviewed at the Institute for Neuropathology/German Brain Tumor Reference Center at the University of Bonn and were classified according to the WHO 2000 classification system for central nervous system tumors.22 In total, 23 specimens from primary Recurrence and Malignancy in GG/Majores et al 3357 FIGURE 1. Histopathologic characteristics of ganglioglioma (GG). (a) Typical GG specimens are characterized by a biphasic (ie, glial and neuronal) phenotype (hematoxylin and eosin stain; original magnification, 3200). (b) The presence of dysplastic neuronal elements is highlighted by immunohistochemistry (IHC) using antibodies against synaptophysin, frequently revealing a perisomatic accumulation (original magnification, 3200). (c) The lack of a dense reticulin fiber network is helpful in the differential diagnosis between GG and pleomorphic xanthoastrocytomas (original magnification, 3200). (d) Perifocal occurrence of ramified CD34 immunoreactivity is highly suggestive of a GG (original magnification, 3200). (e) In some GG examples, CD34 immunoreactivity also was detected in association with tumor cell processes (original magnification, 3400). (f) Malignant progression to a glioblastoma multiforme usually reveals a loss of GG features, including loss of CD34 immunolabeling (original magnification, 3200). Increased proliferative activity (g) (Ki67 [MIB-1]; original magnification, 3400) and (h) a prominent gemistocytic tumor cell component (h) are atypical findings in GG (glial fibrillary acidic protein; original magnification, 3200). (i) Malignant progression to a glioblastoma is indicated by a brisk mitotic and proliferative activity as well as palisading necrosis in concert with loss of GG features (IHC; original magnification, 3100). tumors and 12 recurrent samples (from 9 patients) were available for a retrospective neuropathologic workup. In addition to conventional hematoxylin eosin stains, immunohistochemical analyses using antibodies against glial and neuronal epitopes were performed to reveal the biphasic nature of the tumors. Typical histopathologic and immunohistochemical characteristics are summarized in Figure 1. Statistical Analysis Commercially available software was used for standard statistical analysis (SPSS version 14.0; SPSS Inc., Chicago, Ill). For survival analysis, Kaplan-Meier estimates and log-rank tests were used. The level of significance was set at P < .05 (2-sided). RESULTS Demographics and Clinical Presentation The current study was comprised of 21 atypical WHO grade II tumors, 5 anaplastic WHO grade III tumors with or without recurrence, and 4 recurrent/ progressive GGs that initially were assigned WHO grade I. Pertinent clinical data are summarized in Table 1. The mean age at initial surgery was 19.8 3358 CANCER December 15, 2008 / Volume 113 / Number 12 TABLE 1 Clinical Characteristics of Gangliogliomas (World Health Organization [WHO] Grade I), Atypical Gangliogliomas (WHO Grade II), and Anaplastic Gangliogliomas (WHO Grade III) Gangliogliomas Characteristic No. of patients Age, y Mean Median Range Aged 16 y, % Men, % Drug-resistant epilepsy, % Tumor location, % Temporal Frontal Other Recurrence rate, % (no./total) Malignant progression, % (no./total) 5-Y survival, % 5-Y PFS, % WHO Grade I WHO Grade II WHO Grade III 177 21 5 25 23 1-64 28 51 91 33.5 32 8-45 19 67 62 23.4 22 3-60 40 80 40 78 8 14 2 (4/177) 0.6 (1/177) — 99 97 62 24 14 33 (7/21) 14 (3/21) 24 (5/21)* 79 70 80 20 — 60 (3/5) 20 (1/5) 40 (2/5)* 53 30 PFS indicates progression-free survival. * Includes patients who had malignant progression diagnosed by magnetic resonance imaging only (ie, without histologic confirmation). years for patients with recurrent/progressive grade I tumors (median age, 17.5 years; range, 10–34 years.). Refractory epilepsy was observed in 3 patients (75%) with recurrent/progressive WHO grade I tumors. Clinical Prognostic Factors Age at surgery <40 years (P 5 .007) was associated significantly with better overall survival (OS), but not with better progression-free survival (PFS). Clinical presentation was associated significantly with OS (P 5 .005) and PFS (P < .001). Patients with drug-resistant epilepsy fared best, the prognosis for all other patients who presented with seizures was intermediate, and survival was worst in patients without seizures. Tumor Location and Radiologic Findings Temporal lobe tumors were diagnosed in 17 patients, including 10 temporomesial GGs. In addition, there were 3 insular GGs with temporal lobe extensions, and there was 1 temporo-occipital GG. Five frontal lobe GGs were observed, and there were 4 tumors of the parietal and/or occipital lobe. In 1 patient with a WHO grade II GG, we observed an additional contrast-enhancing lesion in the fourth ventricle that remained stable in size over a follow-up of 176 months. Patients with extratemporal tumors had a significantly shorter PFS (P 5 .01), but OS was not affected. All available neuroimaging data from our patients were reviewed. The preoperative differential diagnosis included GG in 47% of patients. No specific neuroimaging patterns that were suggestive of the presence of WHO grade II or III GG or that pointed to an adverse clinical course in general could be delineated. However, perifocal edema was observed in 26% of patients. Typical imaging findings (42% of patients) included strongly contrast-enhancing tumor nodules (or at least some contrast-enhancing areas) often accompanied by intratumoral cysts within a temporal (or frontal) tumor. The nonenhancing tumor component involved the cerebral cortex as well as white matter and was delineated best on fluid-attenuated inversion recovery images. Only 19% of tumors showed no contrast enhancement. Intratumoral cysts were observed in 50% of patients (Fig. 2). Computed tomography scans were available from 11 patients, and 72% of those scans showed calcifications. Surgical Therapy and Epileptologic Outcomes Tumor growth in the basal ganglia and/or insular region prevented radical surgery in some patients. Complete resection was achieved at 80% of primary surgeries. A complete resection proved to be a strong predictor of both OS (P 5 .002) and recurrence-free survival (P 5 .001). Intraoperative findings that suggested a diagnosis of GG, ie, the presence of a relatively firm, well delineated, and sometimes cystic and calcified mass of yellow-brown or gray color, were recorded in only 11 of 21 patients (52%) with WHO grade II tumors and in 1 of 5 patients (20%) with WHO grade III tumors. Postoperatively, there was 1 new minor dysphasia, 1 patient with worsened hemiparesis, and 1 patient with worsened hemiparesis and dysphasia. Visual field defects (quadrantanopias) were observed after temporomesial and/or occipital resection in 9 patients. There was no mortality. Epileptologic outcomes were available for 15 of 18 patients who presented with drug-resistant epilepsy. Thirteen patients (87%) were seizure-free or suffered only from isolated, nondebilitating seizures at the 1-year follow-up (Engels Class I).25 Tumor Recurrence and Malignant Progression Kaplan-Meier estimates of OS and PFS are shown in Figure 3. The 5-year survival rates for patients with atypical GG (WHO grade II) and grade III tumors were 79% and 53%, respectively (see also Table 1). The recurrence rates were 33% (7 of 21 patients) for Recurrence and Malignancy in GG/Majores et al atypical GG (median follow-up, 55 months; range, 6176 months) and 60% (3 of 5 patients) for WHO grade III tumors (median follow-up, 38 months; range, 15-118 months). The mean time interval between the initial surgery and tumor recurrence was 23 19 months (median, 18 months; range 6-72 months). Malignant progression to glioblastoma was observed in 5 of 11 patients (45%; 4 men and 1 woman) who underwent surgery for recurrent tumors (Table 1). Secondary glioblastoma was observed in 1 of 4 patients (25%) with recurrent WHO grade I GG, in 3 of 21 patients (14%) with atypical GG (Fig. 2C,D), and in 1 of 5 patients (20%) with WHO grade III GG. The mean age at initial surgery of patients who later suffered from secondary GBM was 45.2 years (me- 3359 dian, 60 years; range, 22-60 years) as compared with 27 years (median, 29 years; range, 3–55 years) for patients without malignant progression (P 5 .03). The primary tumor was located in the temporal lobe in 2 patients, in the frontal lobe in 2 other patients, and in the parietal lobe in 1 patient. Only 2 patients FIGURE 2. Magnetic resonance imaging (MRI) features of anaplastic (World Health Organization [WHO] grade III) gangliogliomas (GG) (a) and atypical GG (WHO grade II) (b-e). (a) A man aged 45 years with a left temporomesial, anaplastic GG (WHO grade III) who presented with a single, generalized seizure. From left to right: T1-weighted axial scan with a typical modular contrast enhancing lesion. T2-weighted coronal image depicting the tumor’s (mild) mass effect. A sagittal fluid-attenuated inversion recovery (FLAIR) scan shows the true extent of the lesion. The MRI was read as typical GG. The tumor was completely resected, and the patient underwent adjuvant radiotherapy. Tumor recurrence consistent with glioblastoma multiforme (GBM) (WHO grade IV) according to MRI criteria was diagnosed after 11 months. Despite chemotherapy with temozolomide, the patient died from tumor progression 4 months later. (b) A young woman aged 16 years with a left temporopolar, atypical GG WHO grade II. The patient suffered from focal seizures. From left to right: T1-weighted axial scan after administration of contrast medium. Axial FLAIR image. T2-weighted axial scan. This MRI was also read as typical GG. Note similarities to the patient shown in a. A complete resection was performed. The patient was followed for 6 months without evidence of tumor recurrence. (c) A man aged 60 years with a right frontocentral, atypical GG (WHO grade II) and refractory epilepsy. From left to right: T1-weighted axial scans before and after application of contrast medium. Axial FLAIR scan. The tumor could be completely resected. Tumor recurrence was diagnosed 20 months later. The patient died 5 months after reoperation (histology: GBM, WHO grade IV) and radiotherapy from tumor progression. (d) A man aged 60 years with a left temporoinsular, atypical GG (WHO grade II) involving the basal ganglia and that presented with hemiparesis. From left to right: T1-weighted coronal scan before and sagittal scan after contrast application. Axial FLAIR image. The MRI depicts a large mass that includes many small cysts and strong contrast enhancement. There is not much perifocal edema. Tumor progression was noted only 6 months after a subtotal (>70%) resection (residual tumor in the basal ganglia). The patient died 5 months after reoperation (histology: GBM, WHO grade IV), radiotherapy, and chemotherapy (temozolomide). (e) A man aged 37 years with a left frontocentroinsular tumor involving the basal ganglia. The patient presented with a single, generalized seizure and minimal dysphasia. From left to right: MR scans obtained after an initial biopsy (histology: atypical GG, WHO grade II). A T1-weighted coronal scan reveals a large tumor with a considerable mass effect and focal contrast enhancement. Axial FLAIR images depict tumor in the frontocentral region, the insula, and the basal ganglia. There is surprisingly little perifocal edema. A subtotal resection was performed (90% of the tumor). Symptomatic tumor progression 10 months later (repeat biopsy: atypical GG, WHO grade II) was treated successfully with radiotherapy and temozolomide chemotherapy. At the time of the current report (39 months after the first biopsy), the patient was evaluated for salvage chemotherapy for further clinical progression. 3360 CANCER December 15, 2008 / Volume 113 / Number 12 FIGURE 3. Kaplan-Meier estimates of overall survival (OS) and progression/recurrence-free survival (PFS). The World Health Organization (WHO) tumor grade was correlated with OS (P < .001) (a) and PFS (P < .001) (b). For comparison, survival data are provided from the patients with WHO grade I tumors who were treated during the study period. A complete resection resulted in longer OS (P 5 .002) (c) and longer PFS (P 5 .001) (d). Clinical presentation proved to be an important prognostic parameter: OS (P 5 .005) (e) and PFS (d) were best in patients with refractory epilepsy and worst in patients without seizures. initially presented with drug-resistant epilepsy. One patient with a secondary glioblastoma who initially was diagnosed with a GG WHO grade III of the right temporal lobe developed leptomeningeal seeding shortly before his death. Follow-up MRI scanning diagnosed recurrent tumors consistent with glioblas- toma in 2 additional patients with atypical GG and in 1 patient with WHO grade III GG who did not undergo repeat surgery. All patients with histologically proven or presumed secondary glioblastomas died from their disease, accounting for all 8 tumorrelated deaths in this series. Recurrence and Malignancy in GG/Majores et al Histologic Analyses and Prognosis Specific histopathologic features were tested as possible predictors of an adverse clinical course. Twenty-four percent of specimens from the current series contained a substantial gemistocytic component, which was correlated significantly with shorter PFS (P 5 .025). The presence of protein droplets predicted longer PFS (P 5 .05) and OS (P 5 .04). Expression of the stem cell marker CD34, which is expressed transiently during early neurulation, was assessed by immunohistochemistry.26 Seventy-seven percent of the initial tumor specimens, but only 18% of recurrent samples, demonstrated satellitosis-like, perilesional CD34 expression. Intralesional CD34 labeling, ie, (focal) staining of the tumor cell processes, was observed in 52% of tumor samples and was correlated with shorter OS in the current series (P 5 .03). Somatic accumulation of synaptophysin or neurofilament protein27 was observed in 39% and 54% of specimens, respectively. No correlations with OS or PFS were observed. DISCUSSION In most patients, the prognosis after surgery for intracranial GG is good. Many patients can expect to be cured from their tumors and also often from their epilepsy.17-21,28 However, some patients will suffer from recurrent tumor and/or malignant progression. In 1 series of 58 GGs, 40 tumors were assigned to histologic grade I, but 16 tumors were grade II, and 2 tumors were grade III. The event-free 5-year survival rate in that cohort was 95% for GG of the cerebral hemispheres.28 Rumana et al observed a surprisingly high number of tumor-related deaths (n 5 10) in a series of 42 patients with supratentorial GG.20 Blumcke and Wiestler reported 30 WHO grade II tumors (9%) and 17 WHO grade III tumors (5%) in a cohort of 326 patients with GG. Tumor recurrence was observed in 25% of patients with grade II tumors and in 38% of patients with grade III tumors.17 Similarly, we identified 21 WHO grade II tumors (10.3%) and 5 WHO grade III tumors (2.5%) in a series of 203 patients with GG. The overall recurrence rate was 7% (14 of 203 patients). Recurrence rates for patients with WHO grade II and III tumors were 33% and 60%, respectively. Im and coworkers studied 34 patients with GG and observed tumor progression in 3 patients: Two of those patients had secondary glioblastomas identified at their second surgery.18 Secondary glioblastomas were diagnosed in 5 of 11 patients (45%) in our series who underwent repeat surgery. GG differs vastly from diffuse low-grade glioma (LGG) with respect to the rate of recurrence. 3361 Nevertheless, a similar 50% rate of malignant progression was reported in patients who underwent surgery for LGG recurrence.29 Similar to LGG regrowth, GG recurrence is not a benign condition. Histologic grading for GG is controversial. The new WHO 2007 classification distinguishes between benign GG WHO grade I, and anaplastic WHO grade III tumors. An intermediate category (formerly labeled atypical GG WHO grade II) is no longer included in the WHO grading system.6,22 The diagnosis of an atypical GG, WHO grade II was applied to tumors that had increased cellularity, nuclear pleomorphy, and/or increased proliferative activity in the glial cell component but that lacked definitive criteria for anaplasia, such as palisading necrosis or a brisk mitotic count. Seventy percent of the tumors analyzed in the current study belonged to this category. Our data support a 3-tiered rather than a 2-tiered histopathologic grading system for GG. The 5-year survival rate for atypical (WHO grade II) GG was 79% in this series. The recurrence rate of these growths was 33% after a median follow-up of 55 months compared with only 2% after surgery for WHO grade I tumors. Conversely, outcomes after surgery for WHO grade III GG were considerably worse (recurrence rate, 60%; 5-year survival rate, 53%) than outcomes for atypical (WHO grade II) tumors despite postoperative radiotherapy in 4 of 5 patients. On the basis of these data, a distinction of WHO grade II GG from both GG WHO grades I and III seems warranted, because it provides valuable prognostic information. To arrive at a better histologic characterization of atypical and anaplastic GG, we investigated specific histopathologic features for possible correlations with the patient outcomes in our series. A gemistocytic differentiation pattern in GG has been suggested as a possible marker for atypia.19 Indeed, the presence of a gemistocytic cell component was correlated significantly with shorter PFS in the current series. Conversely, patients who had tumors that contained protein droplets had better OS and longer PFS. This is not surprising, because protein droplets are a frequent finding in typical benign GG and other low-grade neoplasms, such as PA, in young patients. Finally, tumor cell process-associated CD34 immunoreactivity, albeit only focal in most instances, was correlated with shorter OS. In such patients, a PXA has to be ruled out, especially for extratemporal tumor locations. It is noteworthy that composite PXA-GG lesions have been reported.8,30,31 Can an adverse clinical outcome also be predicted by clinical parameters? Most GG series include 3362 CANCER December 15, 2008 / Volume 113 / Number 12 slightly more patients who are men.19 It has been suggested that being a man rather than a woman may be a negative prognostic parameter.20 Indeed, men clearly were over-represented among the recurrent and WHO grade II and III tumors in the current study. Age 40 years was correlated strongly with shorter OS (but not PFS, although patients with tumor recurrence after resection for GG WHO grade II and III, on average, were older than those without recurrence). There was a significant correlation between age at initial surgery and the formation of a secondary GBM. Older age also has been associated with an adverse prognosis by other authors.20 It was suggested that anaplastic GGs occur more often in older patients compared with their nonmalignant counterparts.5,17,19,32 Clinical presentation was correlated strongly with both OS and PFS in this series. The best outcomes were observed in patients with drug-resistant epilepsy. The prognosis was worst in patients without seizures. Tumor site was correlated with PFS but not with OS in the current analysis. All 5 frontal tumors from this series recurred. The analysis of GG series that included mostly benign tumors has provided similar data, but some studies failed to identify these associations.18-20 Finally, the degree of resection proved to be an important prognostic parameter in our patients. A strong correlation was observed with OS and PFS. Similar results were obtained by Luyken and coworkers in their series, which included mostly benign tumors.19 Im and coworkers reported tumor recurrences after 3 of 9 incomplete resections, and none of their completely resected tumors relapsed.18 These data support the role of aggressive surgery in the management of higher grade GG. In contrast to all other prognostic parameters discussed here, the degree of resection is the only variable that can be influenced, at least in part, by the treating physician. We conclude from our data that the results of a careful histopathologic analysis, together with certain clinical parameters allow an individualized assessment of the risk of tumor recurrence and tumorrelated death in patients with GG. The typical patient who is at risk for an adverse clinical course is 1) a man, 2) aged >40 years, 3) who presents with signs and symptoms other than (drug- resistant) epilepsy, and 4) has undergone an incomplete resection for 5) an extratemporal tumor. Histopathologic analysis will reveal 6) some atypical features (such as the presence of a gemistocytic cell component) or some overtly anaplastic features. Early repeat surgery for resectable residual tumor should be contemplated in such patients, and patients should undergo regular surveillance imaging for more than 5 years. Our data do not allow us to make firm conclusions with respect to adjuvant therapies. Radiotherapy and chemotherapy may have some efficacy (Fig. 2). REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. Wolf HK, Muller MB, Spanle M, Zentner J, Schramm J, Wiestler OD. Ganglioglioma: a detailed histopathological and immunohistochemical analysis of 61 cases. Acta Neuropathol (Berl). 1994;88:166-173. Wolf HK, Wiestler OD. Neuropathologic findings in chronic epilepsy. Laryngorhinootologie. 1995;74:127-129. Zentner J, Wolf HK, Ostertun B, et al. Gangliogliomas: clinical, radiological, and histopathological findings in 51 patients. J Neurol Neurosurg Psychiatry. 1994;57:1497-1502. Evans AJ, Fayaz I, Cusimano MD, Laperriere N, Bilbao JM. Combined pleomorphic xanthoastrocytoma-ganglioglioma of the cerebellum. Arch Pathol Lab Med. 2000;124:17071709. Hirose T, Scheithauer BW, Lopes MB, Gerber HA, Altermatt HJ, VandenBerg SR. Ganglioglioma: an ultrastructural and immunohistochemical study. Cancer. 1997;79:989-1003. Louis DN, Ohgaki H, Wiestler OD, et al. The 2007 WHO classification of tumours of the central nervous system. Acta Neuropathol. 2007;114:97-109. Prayson RA. Composite ganglioglioma and dysembryoplastic neuroepithelial tumor. Arch Pathol Lab Med. 1999;123: 247-250. Yeh DJ, Hessler RB, Stevens EA, Lee MR. Composite pleomorphic xanthoastrocytoma-ganglioglioma presenting as a suprasellar mass: case report. Neurosurgery. 2003;52:14651469. Mekni A, Chelly I, Haouet S, Zitouna M, Kchir N. [Malignant cerebellar ganglioglioma. A case report and review of the literature]. Neurochirurgie. 2006;52(2-3 pt 1):119-122. Resta N, Lauriola L, Puca A, et al. Ganglioglioma arising in a Peutz-Jeghers patient: a case report with molecular implications. Acta Neuropathol. 2006;112:106-111. Tamiya T, Hamazaki S, Ono Y, et al. Ganglioglioma in a patient with Turcot syndrome. Case report. J Neurosurg. 2000;92:170-175. Mizuguchi M, Takashima S, Yamanouchi H, Nakazato Y, Mitani H, Hino O. Novel cerebral lesions in the Eker rat model of tuberous sclerosis: cortical tuber and anaplastic ganglioglioma. J Neuropathol Exp Neurol. 2000;59:188-196. Roach ES, Gomez MR, Northrup H. Tuberous sclerosis complex consensus conference: revised clinical diagnostic criteria. J Child Neurol. 1998;13:624-628. Becker AJ, Lobach M, Klein H, et al. Mutational analysis of TSC1 and TSC2 genes in gangliogliomas. Neuropathol Appl Neurobiol. 2001;27:105-114. Pandita A, Balasubramaniam A, Perrin R, Shannon P, Guha A. Malignant and benign ganglioglioma: a pathological and molecular study. Neuro Oncol. 2007;9:124-134. Fukushima T, Katayama Y, Watanabe T, Yoshino A, Komine C, Yokoyama T. Aberrant TP53 protein accumulation in the neuronal component of ganglioglioma. J Neurooncol. 2005; 72:103-106. Blumcke I, Wiestler OD. Gangliogliomas: an intriguing tumor entity associated with focal epilepsies. J Neuropathol Exp Neurol. 2002;61:575-584. Recurrence and Malignancy in GG/Majores et al 18. Im SH, Chung CK, Cho BK, et al. Intracranial ganglioglioma: preoperative characteristics and oncologic outcome after surgery. J Neurooncol. 2002;59:173-183. 19. Luyken C, Blumcke I, Fimmers R, Urbach H, Wiestler OD, Schramm J. Supratentorial gangliogliomas: histopathologic grading and tumor recurrence in 184 patients with a median follow-up of 8 years. Cancer. 2004;101:146-155. 20. Rumana CS, Valadka AB, Contant CF. Prognostic factors in supratentorial ganglioglioma. Acta Neurochir (Wien). 1999;141:63-69. 21. Rumana CS, Valadka AB. Radiation therapy and malignant degeneration of benign supratentorial gangliogliomas. Neurosurgery. 1998;42:1038-1043. 22. Kleihues P, Cavenee WK. Pathology and Genetics of Tumours of the Nervous System. Lyon, France: IARC; 2000. 23. Dash RC, Provenzale JM, McComb RD, Perry DA, Longee DC, McLendon RE. Malignant supratentorial ganglioglioma (ganglion cell-giant cell glioblastoma): a case report and review of the literature. Arch Pathol Lab Med. 1999;123: 342-345. 24. Schramm J, Clusmann H. The surgery of epilepsy. Neurosurgery. 2008;62(suppl 2)463-481; discussion 481. 25. Engel J Jr. Outcome with respect to epileptic seizures. In: Engel J Jr, ed. Surgical Treatment of the Epilepsies. New York, NY: Raven Press; 1993:609-621. 3363 26. Blumcke I, Giencke K, Wardelmann E, et al. The CD34 epitope is expressed in neoplastic and malformative lesions associated with chronic, focal epilepsies. Acta Neuropathol. 1999;97:481-490. 27. Wierzba-Bobrowicz T, Schmidt-Sidor B, Gwiazda E, Bertrand E. The significance of immunocytochemical markers, synaptophysin and neurofilaments in diagnosis of ganglioglioma. Folia Neuropathol. 1999;37:157-161. 28. Lang FF, Epstein FJ, Ransohoff J, et al. Central nervous system gangliogliomas. Part 2: clinical outcome. J Neurosurg. 1993;79:867-873. 29. Schmidt MH, Berger MS, Lamborn KR, et al. Repeated operations for infiltrative low-grade gliomas without intervening therapy. J Neurosurg. 2003;98:1165-1169. 30. Furuta A, Takahashi H, Ikuta F, Onda K, Takeda N, Tanaka R. Temporal lobe tumor demonstrating ganglioglioma and pleomorphic xanthoastrocytoma components. Case report. J Neurosurg. 1992;77:143-147. 31. Perry A, Giannini C, Scheithauer BW, et al. Composite pleomorphic xanthoastrocytoma and ganglioglioma: report of 4 cases and review of the literature. Am J Surg Pathol. 1997;21:763-771. 32. Brainer-Lima PT, Brainer-Lima AM, Azevedo-Filho HR. Ganglioglioma: comparison with other low-grade brain tumors. Arq Neuropsiquiatr. 2006;64(3A):613-618.