Optimal treatment strategy for intracranial germ cell tumors: a single institution analysis

Transcription

1 J Neurosurg Pediatrics 4: , 4: , 2009 Optimal treatment strategy for intracranial germ cell tumors: a single institution analysis Clinical article Ma s ay u k i Ka n a m o r i, M.D., Ph.D., 1 To s h i h i r o Ku m a b e, M.D., Ph.D., 1 Ry u ta Sa i t o, M.D., Ph.D., 1 Yo j i Ya m a s h i t a, M.D., Ph.D., 1 Yu k i h i k o So n o d a, M.D., Ph.D., 1 Hi s a n o r i Ari g a, M.D., Ph.D., 2 Yo s h i h i r o Ta k a i, M.D., Ph.D., 2 a n d Te i j i To m i n a g a, M.D., Ph.D. 1 Departments of 1 Neurosurgery and 2 Radiation Oncology, Tohoku University Graduate School of Medicine, Sendai, Japan Object. This study retrospectively analyzed the long-term outcomes of 108 consecutive patients to establish the classification and optimal treatment strategy for each subgroup of newly diagnosed germ cell tumors (GCTs). Methods. A retrospective review of medical records from the authors department between April 1989 and March 2007 identified 108 patients with newly diagnosed intracranial GCT. The diagnoses were germinoma in 83 patients, and nongerminomatous GCT (NGGCT) in 25 patients. Results. In patients with germinoma, the 10-year overall and progression-free survival (PFS) rates at a median follow-up period of 99 months were 86 and 74%, respectively. Recurrences developed during a range of 6 to 153 months (median 26 months) after starting the initial therapy. Patients treated only with chemotherapy demonstrated a shorter PFS rate, and patients treated with chemotherapy followed by reduced-dose radiation therapy to the whole ventricle, whole brain, or craniospinal axis showed significantly better PFS than patients treated with only radiation or reduced-dose radiation therapy to the focal field. Nongerminomatous GCTs were divided into good, intermediate, and poor prognosis groups as proposed by the Japanese Pediatric Brain Tumor Study Group. In the good and intermediate prognosis groups, the 10-year overall and PFS rates were 100 and 93%, respectively. In the poor prognosis group, the 3-year overall and PFS rates were 56 and 29%, respectively. All patients with NGGCTs, in whom the lesions on MR imaging disappeared after combination therapies consisting of resection, radiation therapy, and chemotherapy, remained alive. Conclusions. Chemotherapy followed by reduced-dose radiation therapy covering the whole ventricle improves the prognosis for patients with germinoma. Combined therapy of radiation therapy, chemotherapy, and radical resection as an initial or salvage treatment achieved excellent tumor control in the intermediate prognosis NGGCT group. The outcomes were still dismal in the poor prognosis NGGCT group, so initial therapy should target complete disappearance of all lesions on MR imaging. (DOI: / PEDS08288) Ke y Wo r d s intracranial germ cell tumor chemotherapy reduced-dose radiation therapy salvage therapy germinoma In t r a c r a n i a l GCTs consist of a heterogeneous group of tumors, including pure germinomas, mature and immature teratomas, teratomas with malignant transformation, yolk sac tumors, choriocarcinomas, embryonal carcinomas, and mixed tumors, and the histological subtype is the single most predictive factor of outcomes. 18 Abbreviations used in this paper: AFP = α-fetoprotein; β-hcg = beta subunit-hcg; CPA = cerebellopontine angle; GCT = germ cell tumor; HCG = human chorionic gonadotropin; JPBTSG = Japanese Pediatric Brain Tumor Study Group; NGGCT = nongerminomatous GCT; PFS = progression-free survival; STGC = syncytiotrophoblastic giant cell. Intracranial GCTs can be divided into germinomas and NGGCTs based on the prognosis and optimal treatment. Germ cell tumors account for 2.8% of all primary brain tumors in Japan. 5 A germinoma is curable using only radiation therapy, with 10-year overall survival rates > 90%. 14,33 However, Gy of radiation therapy to the whole ventricle or larger field is necessary to prevent recurrence, 14,33 so neuroendocrine deficiencies, growth retardation, and cognitive deficits are potential complications. 1,4,28,32,36 Therefore, several trials of chemotherapy regimens without radiation therapy have been performed, but the long-term outcomes indicated unacceptable rates of re- 506 J Neurosurg: Pediatrics / Volume 4 / December 2009

2 Treatment of newly diagnosed intracranial germ cell tumor currence. 2,13,17,32 In contrast, chemotherapy combined with reduced-dose radiation therapy has produced promising results for control of germinoma. 1,4,19 Nongerminomatous GCTs are relatively resistant to radiation therapy and chemotherapy and are associated with a poor outcome. 9,21,26 Nongerminomatous GCTs consist of heterogeneous histological types, with a significant difference in the prognosis between pure malignant GCT and other types of NGGCTs. 21,26 Intracranial GCTs were divided into 3 groups based on treatment outcomes to establish treatment regimens for germinomas and NGGCTs. A mature teratoma was classified into the good prognosis group. Nongerminomatous GCTs, except mature teratomas, were divided as follows: immature teratoma, teratoma with malignant transformation, and mixed tumors mainly consisting of germinomas or teratomas were classified into the intermediate group; and yolk sac tumor, embryonal carcinoma, choriocarcinoma, and mixed tumors consisting mainly of these histological subtypes were classified into the poor prognosis group. 19 Based on this classification system, the JPBTSG conducted a multiinstitutional Phase II study, although using a relatively short follow-up period. 19 The protocol was considered to be effective for patients with germinoma and tumors in the intermediate prognosis group, but not for patients with tumors in the poor prognosis group. 19 The Second International CNS Germ Cell Tumor Study Group found that intensive chemotherapy and salvage surgery/radiation therapy for NGGCTs resulted in survival of 11 of 14 patients with tumors in the intermediate prognosis group and 3 of 6 patients with tumors in the poor prognosis group, at a median follow-up period of 6.3 years. 13 Therefore, the treatment outcomes of the pure malignant and intermediate malignant types of NGGCT should be analyzed independently. The present study retrospectively analyzed the longterm outcomes of 108 consecutive patients to establish the classification and optimal treatment strategy for each subgroup of patients with newly diagnosed GCTs. Methods Patient Characteristics A retrospective review of medical records at our department from April 1989 to March 2007 identified 108 patients with intracranial GCT. Of these 108 patients, 86 were men and 22 were women, with an age range from 2 months to 45 years old (median 14 years). The clinical diagnosis of GCT was based on typical age (GCTs most commonly occur in the second to third decade of life), MR imaging, and serum and/or CSF analysis of the levels of AFP, HCG, and β-hcg. 11 Some of these patients were described previously. 6,11,16,17,22,23,34,35 The histological diagnosis was established from the specimens obtained by endoscopic biopsy procedures or resection via craniotomy, according to the WHO classification of tumors of the nervous system. 18 Informed consent was obtained from all patients involved in this study. Germinoma Treatment Group Patients with pure germinomas and germinomas with J Neurosurg: Pediatrics / Volume 4 / December 2009 STGC and histological verification, as well as clinically diagnosed germinomas (HCG-producing germinomas), were classified in the germinoma treatment group. The clinical diagnosis was based on slightly elevated level of HCG (< 50 IU/L) or β-hcg (< 5 ng/ml) without elevated levels of AFP. The treatment protocol for patients with germinomas between 1989 and 1995 consisted of 20 Gy of radiation therapy to the primary site as a diagnostic therapy without histological verification, followed by additional radiation therapy of Gy to the craniospinal axis, whole brain, ventricles, or the primary tumor site if any response to radiation therapy was detected. 11 The treatment protocol between 1995 and 1997 consisted of 3 cycles of chemotherapy using carboplatin (150 mg/ m 2 on Days 1 3) and etoposide (150 mg/m 2 on Days 1 3; CARE regimen); cisplatin (20 mg/m 2 on Days 1 5) and etoposide (100 mg/m 2 on Days 1 5; PE regimen); ifosfamide (900 mg/m 2 on Days 1 5), cisplatin (20 mg/m 2 on Days 1 5), and etoposide (100 mg/m 2 on Days 1 5; ICE regimen); or bleomycin (20 mg/m 2 on Days 1, 7, and 14), etoposide (100 mg/m 2 on Days 1 5), and cisplatin (20 mg/ m 2 on Days 1 5; BEP regimen). In this period, radiation therapy was omitted. 17 The treatment protocol from 1997 until 2007 consisted of CARE or ICE chemotherapy followed by reduceddose radiation therapy, as proposed by the JPBTSG. 19,20 The radiation field was determined according to the distribution of the tumor, as assessed by brain and spinal MR imaging and CSF cytology. Patients with positive CSF cytology and/or spinal dissemination on MR imaging received 24 Gy of radiation therapy to the craniospinal axis; patients with ventricular dissemination and negative cytology or basal ganglia germinoma received 24 Gy to the whole brain, and patients without ventricular dissemination received 24 Gy to the primary tumor site from , or 24 Gy to the whole ventricle after Patients with residual enhancement at the completion of radiation therapy, or who could not receive chemotherapy due to systemic disorders, received an additional Gy of local radiation therapy. Nongerminomatous GCT Treatment Group Patients with mature teratomas, immature teratomas, teratomas with malignant transformation, and mixed tumors mainly consisting of germinomas or teratomas, as well as yolk sac tumors, embryonal carcinomas, choriocarcinomas, and mixed tumors mainly consisting of these histological subtypes, were included in the NGGTC treatment group. Diagnostic radiation therapy was performed for these patients in the same way as for patients in the germinoma treatment group between 1989 and However, patients with teratomas or highly malignant GCTs diagnosed from the findings of MR imaging and tumor markers underwent resection. 11 Patients with NGGCTs were further classified according to the classification of the JPBTSG after 1995 into the poor prognosis group and the intermediate prognosis group. 19 The diagnoses were based on histological verification or clinical findings. 11 The diagnoses of yolk sac tumor and choriocarcinoma were based on a high AFP (> 2000 ng/ml) or HCG (> 2000 miu/ml) level, respectively. 19 Patients in the good prognosis group 507

3 M. Kanamori et al. underwent resection for mature teratomas or mixed GCTs consisting of mature teratomas and germinomas, or chemotherapy and radiation therapy for mixed tumors consisting of germinomas and mature teratomas. Patients with mature teratomas received no further treatment. Patients in the intermediate and poor prognosis groups received 3 cycles of ICE, BEP, PE, or CARE chemotherapy at 4-week intervals, followed by radiation therapy with various dosages and volumes depending on the histology and distribution of the lesions estimated by MR imaging and cytology. Salvage surgery was performed after or during chemotherapy and radiation therapy if complete remission was not achieved. Thereafter, 1 5 cycles of maintenance chemotherapy were administered. Follow-Up and Statistical Analysis Differences between patients and treatment characteristics were examined using the Fischer exact or Student t-test. The patients were usually followed-up every 4 months until 3 years after the initial treatment, and every 6 12 months thereafter. The correlations between elevation of HCG or β-hcg levels, or the presence of STGCs, and the prognosis remain controversial. 12,21,27,29 The level of HCG or the presence of STGC was not associated with poor prognosis, 19,30 and a slightly elevated level of β-hcg was detected in all patients with germinomas using a highly sensitive method. Therefore, we first analyzed the effect of β-hcg elevation on progression-free and overall survival rates. Progression-free and overall survival rates were defined as months to tumor progression or death from any cause, and months to death from any cause or last follow-up, respectively. The progression-free and overall survival rates were calculated using the Kaplan-Meier method. The prognostic factors were evaluated using the log-rank test. Results Histological Diagnosis The histological diagnosis was established in 52 of the 108 patients before initial treatment (Table 1), and 56 patients received radiation therapy and/or chemotherapy without histological verification as reported previously. 11 Thirteen of these 56 patients underwent salvage resection, and histological diagnoses were established after chemotherapy or radiation therapy (Table 1). Long-Term Outcomes in the Germinoma Group Pretreatment serum and/or CSF β-hcg evaluation was performed in 76 of the 83 patients with histologically verified pure germinomas, germinomas with STGC, and mixed tumors of germinomas and teratomas, and also in clinically diagnosed germinomas, and showed high titer in 36 patients (Table 2). The 2 groups showed no statistically significant differences in age, sex, and treatment modality. However, an elevated β-hcg level was unexpectedly less frequent in basal ganglia germinomas. 27 The 10-year overall survival rates in patients with normal and elevated β-hcg levels were 77 and 97%, respectively (p = 0.16), and the 10-year PFS rates were 81 and TABLE 1: Histological diagnosis of patients with intracranial GCTs Histological Diagnosis No. of Patients histology results before treatment pure germinoma 34 germinoma w/ STGC 2 pure germinoma/mature teratoma 2 mature teratoma 1 immature teratoma 1 mature teratoma w/ malignant component 2 germinoma > embryonal carcinoma w/ malignant component 3 germinoma w/ malignant component 1 embryonal carcinoma 3 yolk sac tumor 1 choriocarcinoma 2 histology results at salvage op pure germinoma 1 mature teratoma 3 mature teratoma w/ malignant component 1 immature teratoma 2 mixed GCT 2 yolk sac tumor 2 choriocarcinoma 1 scar 1 71%, respectively (p = 0.42). The absence of differences in the overall survival and PFS rates confirmed that the outcomes of the 83 patients with histologically verified and clinically diagnosed germinomas with normal and elevated levels of β-hcg could be analyzed together as the germinoma group. Seventy-four of the 83 patients in the germinoma group remained alive during follow-up periods ranging from 4 to 233 months (median 99 months). Three patients died of progressive disseminated disease, 3 of malignant transformation to yolk sac tumors or embryonal carcinomas, 11,16 1 of suicide due to tumor recurrence, 1 of radiation-induced glioblastoma without evidence of recurrent germinoma, and 1 of unknown cause. A male patient with radiation-induced glioblastoma presented with diabetes insipidus at age 16 years. This patient had synchronous lesions in the pineal and neurohypophysial region without elevation of tumor markers, and was treated under a diagnosis of germinoma based on the typical age, sex, and MR imaging findings using 50 Gy of radiation therapy to the extended focal field. The lesion disappeared completely. This patient developed histologically proven glioblastoma in the right cerebellar hemisphere 116 months after diagnosis, and this lesion was diagnosed as radiation-induced glioblastoma. Two patients with malignant transformation underwent endoscopic or stereotactic biopsy procedures, and the diagnoses were pure germinoma. Another patient was diagnosed based only on clinical findings. These 3 patients achieved complete remission of the lesion after chemotherapy and radiation therapy, but then developed 508 J Neurosurg: Pediatrics / Volume 4 / December 2009

4 Treatment of newly diagnosed intracranial germ cell tumor TABLE 2: Summary of patients classified by β-hcg value Variable Normal β-hcg Level (ng/ml) Elevated β-hcg Level (ng/ml) no. of patients age in yrs (median) 8 42 (14) 6 45 (15) sex (M/F) 32:8 30:6 value of β-hcg (ng/ml) < histological diagnosis pure germinoma germinoma w/ STGC 1 1 germinoma/mature teratoma 2 0 unverified location pineal neurohypophysial 10 8 pineal/neurohypophysial basal ganglia 6 1 treatment modality only focal radiation only WV-CS radiation 6 8 focal combined 5 3 WV-CS combined only chemotherapy 6 2 PFS rate (%) 5-yr yr overall survival rate (%) 5-yr yr embryonal carcinoma at 12 and 6 months, and yolk sac tumor at 94 months, respectively 11,16 The treatment modalities used in this group were resection or biopsy in 38 patients, chemotherapy and radiation therapy in 45 patients, and salvage surgery in only 2 patients (Table 3). To determine the optimal treatment for germinoma, the 83 patients were divided into 5 groups: 1) only chemotherapy; 2) only radiation therapy to the focal site (only focal radiation); 3) only radiation therapy to the whole ventricle, whole brain, or craniospinal axis (only WV-CS radiation); 4) reduced-dose focal radiation therapy with chemotherapy (focal combined); and 5) reduceddose whole ventricle, whole brain, or craniospinal axis radiation therapy with chemotherapy (WV-CS combined). The 5-, 10-, and 15-year overall survival rates in the germinoma group were 95, 86, and 83%, respectively (Fig. 1 left). There were no significant differences in overall survival rates between the treatment modalities (Table 4; Fig. 2 left). Tumor recurrence developed in 19 patients during the follow-up period. The 3-, 5-, and 10-year PFS rates were 86, 78, and 74%, respectively (Fig. 1 right), and recurrences developed between 6 and 153 months (median 26 months) after starting the initial therapy (Fig. 1 right). Five of these patients (26%) suffered the first recurrence at 60 months or later. The site of recurrence was the primary J Neurosurg: Pediatrics / Volume 4 / December 2009 TABLE 3: Summary of the treatment of intracranial germ cell tumors Variable Germinoma* Group Intermediate Prognosis NGGCT Poor Prognosis NGGCT no. of patients op total resection subtotal or partial resection biopsy salvage op none chemotherapy CARE PE ICE BEP none radiation therapy craniospinal axis whole brain whole ventricle focal none * Histologically verified germinomas, germinomas with STGC, mixed tumor consisting of germinomas and mature teratomas, and clinically diagnosed germinomas. site in 4 patients, primary site and disseminated disease in 1, ventricular wall in 8, spinal cord in 2, optic nerve in 1, thalamus in 1, CPA and spinal cord in 1, and cerebellar tentorium in 1 (Table 4). There were statistically significant differences in the PFS rates between the treatment modalities. Patients in the chemotherapy-only group had shorter PFS rates compared with the patients in the other groups (Table 4; Fig. 2 right). None of the patients in the WV-CS combined group developed recurrence. The WV-CS combined group demonstrated a significantly better PFS rate compared with the focal combined group (p = 0.003), focal radiation only group (p = 0.035), and WV-CS radiationonly group (p = 0.034; Table 4; Fig. 2 right). Long-Term Outcomes in the NGGCT Group A total of 25 patients with NGGCTs, excluding germinomas with STGC and mixed tumors consisting of germinomas and mature teratomas, were treated in the NGGCT group. Histological diagnoses were established after tumor resection but before radiation therapy and chemotherapy in 14 patients, and chemotherapy and radiation therapy were given without histological diagnosis in 11 patients. Histological diagnoses and treatments are listed in Tables 1 and 3. The patients with NGGCT were divided into 1 pa- 509

5 M. Kanamori et al. Fig. 1. Kaplan-Meier plots comparing overall survival (left) and PFS (right) rates for patients in the germinoma group, the intermediate prognosis NGGCT group, and the poor prognosis NGGCT group. tient in the good prognosis group, 16 in the intermediate prognosis group followed up for months (median 55 months), and 8 in the poor prognosis group followed up for months (median 15 months). The patient in the good prognosis group underwent total resection of a mature teratoma, and did not show any tumor recurrence on follow-up. Seven patients in the intermediate group underwent subtotal or total resection of the tumor as the initial therapy (Table 3). Eight of the 9 patients treated without histological verification had to undergo salvage surgery for residual tumors. The histological diagnoses were mature teratoma in 3, immature teratoma in 2, mature teratoma with malignant transformation in 1, and mixed germ cell tumor in 2 (Table 1). All residual tumors were completely removed at salvage surgery. Fourteen patients received radiation therapy to various fields, and 13 received chemotherapy (Table 3). One of the 16 patients died due to complications from chemotherapy without evidence of recurrence. The 5- and 10-year overall survival rates were both 93% in the intermediate prognosis group (Fig. 1 left). Magnetic resonance imaging showed no detectable lesions after the initial treatment (including initial or salvage resection, radiation therapy, and chemotherapy), and no further recurrences were detected (Fig. 1 right). The 8 patients in the poor prognosis group underwent surgery: subtotal or total resection in 6, and salvage surgery in 2 (Table 3). The histological diagnoses from sal- TABLE 4: Summary of the patients classified by treatment modality* Only Radiation Therapy Combined Therapy Variable Focal Site WV-CS Focal Site WV-CS Only Chemotherapy no. of patients age in yrs (median) 8 45 (14) 9 29 (13.5) 6 26 (18) 8 26 (15) 8 31 (20) sex (M/F) 17:4 13:4 7:2 22:5 7:1 range of follow-up in mos (174) (123) (131) (52) (104) (median) range of radiation dose to (50) (48) (24) (24) 0 tumor site in Gy (median) β-hcg elevation (%) PFS rate (%) 5-yr yr median PFS (mos) NR NR NR NR 32 recurrence site (no. of patients) VW (4), SC (1), cerebellar tentorium (1) PS (1), PS, VW, SC (1), SC (1), CPA, SC (1) PS (1), VW (1), optic nerve (1) none PS (2), VW (3), thalamus (1) overall survival rate (%) 5-yr yr * NR = not reached (median could not be calculated); PS = primary site; SC = spinal cord; VW = ventricular wall; WV-CS = radiation therapy to whole ventricle, whole brain, or craniospinal axis. 510 J Neurosurg: Pediatrics / Volume 4 / December 2009

6 Treatment of newly diagnosed intracranial germ cell tumor Fig. 2. Kaplan-Meier plots comparing overall survival (left) and PFS (right) rates for patients with newly diagnosed germinomas according to treatment modalities, including only chemotherapy, only radiation therapy to the focal site (only focal radiation), only radiation therapy to the whole ventricle, whole brain, or craniospinal axis (only WV-CS radiation), reduced-dose focal radiation therapy with chemotherapy (focal combined), and reduced-dose whole ventricle, whole brain, or craniospinal axis radiation therapy with chemotherapy (WV-CS combined). vage surgery were yolk sac tumor and choriocarcinoma. Seven patients received radiation therapy and chemotherapy, but one patient suffered fatal vasospasm 11 days after tumor resection (Table 3). Three patients had tumor progression during the initial treatment, and one patient at 27 months after the initial treatment. Three patients developed dissemination, and 1 had local tumor progression. Two of the patients with recurrence were alive 13 and 190 months after progression: 1 with progressive spinal dissemination and the other with no recurrent disease after salvage treatment. Therefore, 2 of the 8 patients died of tumor progression, and 1 patient died of postoperative vasospasm. All patients with disappearance of lesions on MR imaging were still alive. The 3-year overall and PFS rates were 56% and 29%, respectively. The poor prognosis group demonstrated a statistically significantly worse overall survival rate compared with that of the germinoma group (p = 0.001), and worse PFS rate compared with that of the germinoma group (p = ) and the intermediate prognosis group (p = ; Fig. 1). Discussion Treatment and Outcomes in the Germinoma Group Most patients in the germinoma group showed complete response to radiation therapy and/or chemotherapy, and only 2 of the 80 patients with detectable lesions on MR imaging before radiation therapy and chemotherapy had to undergo salvage surgery. These good responses to radiation therapy and chemotherapy suggest that the diagnoses of germinoma were appropriate. However, 4 patients developed unusual recurrent diseases during the follow-up period, including radiation-induced glioblastoma in 1 patient and malignant transformation in 2 patients. These 3 cases indicate that the diagnosis at presentation was incorrect, and that the first patient had NGGCT, and the other 2 patients had NGGCT components outside the biopsy specimen. Recently, a high incidence of radiation-induced malignant tumors up to 18% J Neurosurg: Pediatrics / Volume 4 / December 2009 was reported after treatment of germinoma. 8 Similarly, malignant transformation was reported in 2 of 60 patients treated with radiation therapy only, 3 so the probability of malignant transformation can be estimated at 3% from the previous and present results. Therefore, we presume that these patients had unusual courses in the treatment of germinoma not due to misdiagnosis, and that treatment of germinoma can be based on the clinical diagnosis of germinoma, without requiring histological confirmation, as described here and elsewhere. 11,12,25,31,32,37 The present study analyzed the outcomes of reduceddosage radiation therapy to the whole ventricle, whole brain, or craniospinal axis, and platinum-based chemotherapy in the germinoma group compared with historical controls. Before the introduction of platinum-based chemotherapy, patients received a median dosage of 50 Gy of radiation therapy to various fields. After the introduction of chemotherapy, chemotherapy-only regimens or chemotherapy followed by reduced-dosage radiation therapy to focal fields was used to avoid the harmful effects of radiation therapy. Because these strategies result in high rates of recurrence, especially in the ventricular wall outside the radiation field, we treated patients with chemotherapy followed by reduced-dosage radiation therapy covering at least the whole ventricle, and demonstrated the effectiveness of this protocol. The 5- and 10-year overall survival rates in the germinoma group were 95 and 86%, respectively, whereas those of the intermediate prognosis NGGCT group were both 93% (p = 0.91). The 5- and 10-year PFS rates in the germinoma group were 78 and 74%, and those of the intermediate prognosis NGGCT group were both 100% (p = 0.088). The difference in the PFS rates did not reach statistical significance, but it is surprising that the patients with germinoma tended to suffer recurrence during the long follow-up period. We attribute this result to the fact that our germinoma group included 9 patients treated only with chemotherapy, 6 of whom developed recurrent lesions within 5 years. A recent study reported that 36% 511

7 M. Kanamori et al. of patients with germinoma developed recurrence after 5 or more years. 10 Therefore, the latency of recurrence in germinoma could be longer than that in the intermediate prognosis NGGCT group, and result in poorer PFS time in the long-term follow-up period. The present study demonstrated that the WV-CS combined group had a significantly better PFS rate, and the chemotherapy-only group had a significantly worse PFS rate than those of the other groups. Therefore, based on the findings of CSF cytology and MR imaging, reduceddose radiation therapy to the whole ventricle, whole brain, or craniospinal axis, combined with chemotherapy, is effective in patients with newly diagnosed intracranial germinoma. In contrast, 6 of the 9 patients treated only with chemotherapy experienced recurrence within 53 months. Fourteen of the 19 patients with recurrence had new lesions outside the radiation field. These findings are consistent with previous reports, 2,17,32 and suggest that chemotherapy-only treatment cannot replace radiation therapy, and that optimal radiation volume and dosage should be identified for the treatment of germinomas. 17 Induction chemotherapy followed by radiation therapy with a reduced dose and a smaller field has been used to avoid the late adverse effects of radiation therapy. 1,4,12,19,20,24,32 Tumor control and excellent functional outcome rates were achieved in 90% of 16 patients with pure germinoma treated with cisplatin-based chemotherapy, followed by 24 Gy of focal radiation therapy to the tumor site for solitary germinoma, the whole ventricle for multifocal germinoma, and the craniospinal axis for disseminated germinoma. 1 However, the combination of cisplatin-based chemotherapy and Gy of focal radiation therapy to the primary tumor site with 1 2 cm margins resulted in an increased recurrence rate up to 20 44%. 24,32 In addition, we previously reported a case of recurrent intracranial germinoma outside the initial radiation field with progressive malignant transformation. 16 Therefore, neoadjuvant reduced-field radiation therapy not covering the whole ventricle may be associated with increased risk of recurrence, and radiation therapy to cover the whole ventricle is essential for longterm tumor control. The adequate radiation volume for tumor control is important to establish. In our study, 10 of 27 patients in the WV-CS combined group showed no dissemination or positive CSF cytology before starting initial treatment, and therefore could be successfully treated with radiation therapy to the whole ventricle, but not to a larger volume. Recurrence is rare in the spinal cord of patients without tumor dissemination who did not receive radiation therapy to the spinal cord; 7,12,21,39 therefore radiation therapy to the craniospinal axis should be limited to patients with evidence of CSF dissemination or spinal metastasis. In addition, radiation therapy to the whole brain combined with chemotherapy was limited to only those patients with dissemination to the ventricular wall and negative cytology after Nine of 13 patients had recurrence with intracranial dissemination limited to the ventricular system. Of the other 4 patients with extraventricular recurrence, 2 had germinomas in the basal ganglia, 1 had recurrence in the optic nerve, and the other had recurrence in the optic nerve, thalamus, and white matter in the left temporal lobe. 22 Our recent study of patients with germinoma in the basal ganglia found that the target volume of radiation therapy in treating germinoma in the basal ganglia is difficult to define, so recurrence could develop in the extraventricular region, including the optic nerve. 34 Therefore, whole brain radiation therapy may be the optimal treatment for patients with basal ganglia lesions. Excluding the patients with germinomas in the basal ganglia, only 2 of the present patients treated with 24 Gy of radiation to the whole ventricle or 48 Gy of focal radiation developed intracranial extraventricular dissemination in the CPA and cerebellar tentorium. This finding suggests that the incidence of intracranial extraventricular dissemination is low, even after focal or whole-ventricle radiation therapy with or without chemotherapy. Therefore, we presume that routine radiation therapy to the whole brain is unnecessary except in patients with germinoma in the basal ganglia and with intracranial dissemination without positive CSF cytology, and that patients without positive cytology or disseminated disease can be successfully treated with radiation therapy to the whole ventricle. To determine the optimal radiation dosage, we analyzed the relationship between radiation dosage and recurrence. The radiation dose was reduced to 24 Gy in 19 of the 27 patients in the WV-CS combined group without increasing the recurrence rate. Additional radiation therapy was administered for residual disease in 5 patients, with contraindication for chemotherapy due to renal disorder in 2 patients. Previously, patients with germinoma with elevated β-hcg, treated with reduced-dose combination therapy, had poor outcomes with a 5-year PFS rate of only 44%, so Gy of radiation therapy combined with effective chemotherapy was necessary to prevent tumor recurrence. 1 In our study, half the patients in the WV-CS combined group had elevated β-hcg, but none developed recurrence. The overall and PFS rates of all patients were also similar for patients with normal and elevated β-hcg levels. These issues should be further investigated by prospective study. Treatment of the Intermediate Prognosis and Poor Prognosis NGGCT Groups The present study demonstrated the usefulness of the JPBTSG classification to reflect the treatment outcomes for NGGCTs, which include heterogeneous histological types. Eight of 12 patients in the intermediate prognosis group who did not undergo total resection required salvage surgery. The residual lesions were totally resected, and the histological diagnoses were mature or immature teratomas, or scars, as described previously. 11,38 As we previously reported, neoadjuvant chemotherapy and radiation therapy contribute to the reduction of tumor volume and vascularity, and malignant components. These effects facilitate complete resection at salvage surgery. Therefore, we could achieve total resection of the tumor in all patients in the intermediate prognosis group after neoadjuvant therapy. 11 Magnetic resonance imaging showed no residual lesions, and these patients developed no recurrent disease at the end of the initial treatment in the intermediate prognosis group. No recurrence developed during the long follow-up period, in contrast to the germinoma group. Previous stud- 512 J Neurosurg: Pediatrics / Volume 4 / December 2009

8 Treatment of newly diagnosed intracranial germ cell tumor ies have also demonstrated excellent results with salvage surgery and subsequent chemotherapy in the treatment of tumors in the intermediate prognosis group. 15,38 The poor prognosis group showed 3 characteristic patterns of recurrence and response to the treatment. First, tumor progression occurred very early, even during the initial treatment, which reflects the aggressiveness of these tumors. Second, 3 of 4 patients who completed initial treatment without detectable lesions on MR imaging remained alive without recurrence, as also found in the intermediate prognosis group, but not in the germinoma group. Excellent results have been reported in patients in the poor prognosis group treated with neoadjuvant therapy, consisting of combined chemotherapy and radiation therapy, followed by complete resection of residual tumors, and all 7 patients survived for months. 15 In our series, 2 patients, in whom gross-total resection was not achieved because of the tumor s anatomical location, experienced rapid growth and dissemination of the tumor, suggesting that more intensive neoadjuvant therapy is necessary to allow complete resection in the treatment of tumors in the poor prognosis group. Third, histological examination of the specimens from salvage surgery demonstrated that the treatment-resistant components consisted of fibrous tissue and yolk sac tumor or choriocarcinoma in the poor prognosis group, in contrast to teratoma or scar in the intermediate prognosis group. Similarly, specimens from salvage surgery consisted of viable cells in 2 and necrosis or mesenchymal tissue in 3 in the poor prognosis group, and mature teratoma in 2 and viable cells in 1 in the intermediate prognosis group. 15 These previous data and our present findings suggest that the histological findings of tumors in the poor prognosis group at salvage surgery are different from those of the intermediate group, and adjuvant therapy is indispensable even after salvage surgery. 15 Surgery and postoperative radiation therapy and chemotherapy are important, as all 8 patients who underwent resection and received radiation therapy, or who underwent incomplete resection, died of progressive disease, whereas 2 of 3 patients who underwent macroscopic resection and received both radiation therapy and chemotherapy survived without tumor recurrence. 26 Conclusions Chemotherapy followed by reduced-dose radiation therapy to the whole ventricle, whole brain, or craniospinal axis improves the prognosis for patients with germinomas. Excellent tumor control was achieved using a combination of radiation therapy, chemotherapy, and resection as an initial or salvage treatment in the intermediate prognosis NGGCT group. Outcomes continued to be dismal in the poor prognosis NGGCT group, so initial treatment should target complete disappearance of the lesion on MR imaging. Disclaimer The authors do not report any conflict of interest concerning the materials or methods used in this study or the findings specified in this paper. J Neurosurg: Pediatrics / Volume 4 / December 2009 References 1. Aoyama H, Shirato H, Ikeda J, Fujieda K, Miyasaka K, Sawamura Y: Induction chemotherapy followed by low-dose involved-field radiotherapy for intracranial germ cell tumors. J Clin Oncol 20: , Balmaceda C, Heller G, Rosenblum M, Diez B, Villablanca JG, Kellie S, et al: Chemotherapy without irradiation A novel approach for newly diagnosed CNS germ cell tumors: results of an international cooperative trial. The First International Central Nervous System Germ Cell Tumor Study. J Clin Oncol 14: , Bamberg M, Kortmann RD, Calaminus G, Becker G, Meisner C, Harms D, et al: Radiation therapy for intracranial germinoma: results of the German cooperative prospective trials MAKEI 83/86/89. J Clin Oncol 17: , Baranzelli MC, Patte C, Bouffet E, Couanet D, Habrand JL, Portas M, et al: Nonmetastatic intracranial germinoma: the experience of the French Society of Pediatric Oncology. Cancer 80: , Committee of Brain Tumor Registry of Japan: Report of Brain Tumor Registry of Japan ( ). Neurol Med Chir (Tokyo) 43 Suppl:1 111, Endo T, Kumabe T, Ikeda H, Shirane R, Yoshimoto T: Neurohypophyseal germinoma histologically misidentified as granulomatous hypophysitis. Acta Neurochir (Wien) 144: , Haas-Kogan DA, Missett BT, Wara WM, Donaldson SS, Lamborn KR, Prados MD, et al: Radiation therapy for intracranial germ cell tumors. Int J Radiat Oncol Biol Phys 56: , Jabbour SK, Zhang Z, Arnold D, Wharam MD: Risk of second tumor in intracranial germinoma patients treated with radiation therapy: the Johns Hopkins experience. J Neurooncol 91: , Jennings MT, Gelman R, Hochberg F: Intracranial germ-cell tumors: natural history and pathogenesis. J Neurosurg 63: , Kamoshima Y, Sawamura Y, Ikeda J, Shirato H, Aoyama H: Late recurrence and salvage therapy of CNS germinomas. J Neurooncol 90: , Kanamori M, Kumabe T, Tominaga T: Is histological diagnosis necessary to start treatment for germ cell tumours in the pineal region? J Clin Neurosci 15: , Kawabata Y, Takahashi JA, Arakawa Y, Shirahata M, Hashimoto N: Long term outcomes in patients with intracranial germinomas: a single institution experience of irradiation with or without chemotherapy. J Neurooncol 88: , Kellie SJ, Boyce H, Dunkel IJ, Diez B, Rosenblum M, Brualdi L, et al: Intensive cisplatin and cyclophosphamide-based chemotherapy without radiotherapy for intracranial germinomas: failure of a primary chemotherapy approach. Pediatr Blood Cancer 43: , Kersh CR, Constable WC, Eisert DR, Spaulding CA, Hahn SS, Jenrette JM III, et al: Primary central nervous system germ cell tumors: effect of histologic confirmation on radiotherapy. Cancer 61: , Kochi M, Itoyama Y, Shiraishi S, Kitamura I, Marubayashi T, Ushio Y: Successful treatment of intracranial nongerminomatous germ cell tumors by administering neoadjuvant chemotherapy and radiation before excision of residual tumors. J Neurosurg 99: , Kon H, Kumabe T, Jokura H, Shirane R: Recurrent intracranial germinoma outside the initial radiation field with progressive malignant transformation. Acta Neurochir (Wien) 144: , Kumabe T, Kusaka Y, Jokura H, Ikeda H, Shirane R, Yoshimoto T: [Recurrence of intracranial germinoma initially treated with chemotherapy only.] No Shinkei Geka 30: , 2002 (Jpn) 513

9 M. Kanamori et al. 18. Louis DN, Ohgaki H, Wiestler OD, et al (eds): WHO Classification of Tumours of the Central Nervous System, ed 4. Lyon: IARC Press, Matsutani M: Combined chemotherapy and radiation therapy for CNS germ cell tumors. The Japanese experience. J Neurooncol 54: , Matsutani M: Clinical management of primary central nervous system germ cell tumors. Semin Oncol 31: , Matsutani M, Sano K, Takakura K, Fujimaki T, Nakamura O, Funata N, et al: Primary intracranial germ cell tumors: a clinical analysis of 153 histologically verified cases. J Neurosurg 86: , Nakajima T, Kumabe T, Jokura H, Yoshimoto T: Recurrent germinoma in the optic nerve: report of two cases. Neurosurgery 48: , Nakamura T, Kanamori M, Sonoda Y, Watanabe M, Kumabe T, Tominaga T: [Primary yolk sac tumor in the cerebellar vermis. Case report.] No Shinkei Geka 37: , 2009 (Jpn) 24. Nguyen QN, Chang EL, Allen PK, Maor MH, Ater JL, Mahajan A, et al: Focal and craniospinal irradiation for patients with germinoma and pattern of failure. Cancer 107: , Ogawa K, Shikama N, Toita T, Nakamura K, Uno T, Onishi H, et al: Long-term results of radiotherapy for intracranial germinoma: a multi-institutional retrospective review of 126 patients. Int J Radiat Oncol Biol Phys 58: , Ogawa K, Toita T, Nakamura K, Uno T, Onishi H, Itami J, et al: Treatment and prognosis of patients with intracranial nongerminomatous malignant germ cell tumors: a multiinstitutional retrospective analysis of 41 patients. Cancer 98: , Ogino H, Shibamoto Y, Takanaka T, Suzuki K, Ishihara S, Yamada T, et al: CNS germinoma with elevated serum human chorionic gonadotropin level: clinical characteristics and treatment outcome. Int J Radiat Oncol Biol Phys 62: , Oka H, Kawano N, Tanaka T, Utsuki S, Kobayashi I, Maezawa H, et al: Long-term functional outcome of suprasellar germinomas: usefulness and limitations of radiotherapy. J Neurooncol 40: , Sawamura Y, Shirato H, Ikeda J, Tada M, Ishii N, Kato T, et al: Induction chemotherapy followed by reduced-volume radiation therapy for newly diagnosed central nervous system germinoma. J Neurosurg 88:66 72, Shibamoto Y, Takahashi M, Sasai K: Prognosis of intracranial germinoma with syncytiotrophoblastic giant cells treated by radiation therapy. Int J Radiat Oncol Biol Phys 37: , Shikama N, Ogawa K, Tanaka S, Toita T, Nakamura K, Uno T, et al: Lack of benefit of spinal irradiation in the primary treatment of intracranial germinoma. A multiinstitutional retrospective review of 180 patients. Cancer 104: , Shim KW, Kim TG, Suh CO, Cho JH, Yoo CJ, Choi JU, et al: Treatment failure in intracranial primary germinomas. Childs Nerv Syst 23: , Shirato H, Nishio M, Sawamura Y, Myojin M, Kitahara T, Nishioka T, et al: Analysis of long-term treatment of intracranial germinoma. Int J Radiat Oncol Biol Phys 37: , Sonoda Y, Kumabe T, Sugiyama S, Kanamori M, Yamashita Y, Saito R, et al: Germ cell tumors in the basal ganglia: problems of early diagnosis and treatment. J Neurosurg Pediatr 2: , Sugiyama S, Kumabe T, Mino M, Fujimura M, Fujiwara I, Tominaga T: [Precocious puberty caused by hcg-producing germinoma involving the bilateral basal ganglia and cerebral white matter without typical radiologic findings: case report.] No Shinkei Geka 34: , 2006 (Jpn) 36. Syndikus I, Tait D, Ashley S, Jannoun L: Long-term follow-up of young children with brain tumors after irradiation. Int J Radiat Oncol Biol Phys 30: , Tian C, Pu C, Wu W, Huang X, Lang S, Yu S, et al: Is biopsy needed to guide management for all patients with presumed intracranial ectopic germinomas? J Neurooncol 92:37 44, Weiner HL, Lichtenbaum RA, Wisoff JH, Snow RB, Souweidane MM, Bruce JN, et al: Delayed surgical resection of central nervous system germ cell tumors. Neurosurgery 50: , Wolden SL, Wara WM, Larson DA, Prados MD, Edwards MS, Sneed PK: Radiation therapy for primary intracranial germcell tumors. Int J Radiat Oncol Biol Phys 32: , 1995 Manuscript submitted September 16, Accepted July 14, Address correspondence to: Toshihiro Kumabe, M.D., Department of Neurosurgery, Tohoku University Graduate School of Medicine, 1-1 Seiryo-machi, Aoba-ku, Sendai, Japan kuma@ nsg.med.tohoku.ac.jp. 514 J Neurosurg: Pediatrics / Volume 4 / December 2009