Change in treatment strategy for intracranial germinoma: Long-term follow-up experience at a single institute

Authors

  • Yi-Wei Chen MD,

    1. Institute of Clinical Medicine, National Yang-Ming University, Taipei, Taiwan
    2. Faculty of Medicine, National Yang-Ming University, Taipei, Taiwan
    3. Division of Radiation Oncology, Taipei Veterans General Hospital, Taipei, Taiwan; Cancer Center, Taipei Veterans General Hospital, Taipei, Taiwan
    Search for more papers by this author
  • Pin-I Huang MD,

    1. Institute of Clinical Medicine, National Yang-Ming University, Taipei, Taiwan
    2. Faculty of Medicine, National Yang-Ming University, Taipei, Taiwan
    3. Division of Radiation Oncology, Taipei Veterans General Hospital, Taipei, Taiwan; Cancer Center, Taipei Veterans General Hospital, Taipei, Taiwan
    Search for more papers by this author
  • Donald Ming-Tak Ho MD,

    1. Faculty of Medicine, National Yang-Ming University, Taipei, Taiwan
    2. Department of Pathology, Taipei Veterans General Hospital, Taipei, Taiwan
    Search for more papers by this author
  • Yu-Wen Hu MD,

    1. Division of Radiation Oncology, Taipei Veterans General Hospital, Taipei, Taiwan; Cancer Center, Taipei Veterans General Hospital, Taipei, Taiwan
    Search for more papers by this author
  • Kai-Ping Chang MD,

    1. Department of Pediatrics, Taipei Veterans General Hospital, Taipei, Taiwan
    Search for more papers by this author
  • Shih-Hwa Chiou MD, PhD,

    1. Institute of Clinical Medicine, National Yang-Ming University, Taipei, Taiwan
    2. Department of Medical Research and Education, Taipei Veterans General Hospital, Taipei, Taiwan
    Search for more papers by this author
  • Wan-Yuo Guo MD, PhD,

    1. Faculty of Medicine, National Yang-Ming University, Taipei, Taiwan
    2. Department of Radiology, Taipei Veterans General Hospital, Taipei, Taiwan
    Search for more papers by this author
  • Feng-Chi Chang MD,

    1. Faculty of Medicine, National Yang-Ming University, Taipei, Taiwan
    2. Department of Radiology, Taipei Veterans General Hospital, Taipei, Taiwan
    Search for more papers by this author
  • Muh-Lii Liang MD,

    1. Institute of Clinical Medicine, National Yang-Ming University, Taipei, Taiwan
    2. Faculty of Medicine, National Yang-Ming University, Taipei, Taiwan
    3. Division of Pediatric Neurosurgery, Taipei Veterans General Hospital, Taipei, Taiwan; Neurological Institute, Taipei Veterans General Hospital, Taipei, Taiwan
    Search for more papers by this author
  • Yi-Yen Lee MD,

    1. Institute of Clinical Medicine, National Yang-Ming University, Taipei, Taiwan
    2. Faculty of Medicine, National Yang-Ming University, Taipei, Taiwan
    3. Division of Pediatric Neurosurgery, Taipei Veterans General Hospital, Taipei, Taiwan; Neurological Institute, Taipei Veterans General Hospital, Taipei, Taiwan
    Search for more papers by this author
  • Hsin-Hung Chen MD,

    1. Faculty of Medicine, National Yang-Ming University, Taipei, Taiwan
    2. Division of Pediatric Neurosurgery, Taipei Veterans General Hospital, Taipei, Taiwan; Neurological Institute, Taipei Veterans General Hospital, Taipei, Taiwan
    Search for more papers by this author
  • Ting-Rong Hsu MD,

    1. Institute of Clinical Medicine, National Yang-Ming University, Taipei, Taiwan
    2. Faculty of Medicine, National Yang-Ming University, Taipei, Taiwan
    3. Department of Pediatrics, Taipei Veterans General Hospital, Taipei, Taiwan
    Search for more papers by this author
  • Shih-Chieh Lin MD,

    1. Faculty of Medicine, National Yang-Ming University, Taipei, Taiwan
    2. Department of Pathology, Taipei Veterans General Hospital, Taipei, Taiwan
    Search for more papers by this author
  • Tai-Tong Wong MD,

    1. Faculty of Medicine, National Yang-Ming University, Taipei, Taiwan
    2. Division of Pediatric Neurosurgery, Taipei Veterans General Hospital, Taipei, Taiwan; Neurological Institute, Taipei Veterans General Hospital, Taipei, Taiwan
    Search for more papers by this author
  • Sang-Hue Yen MD

    Corresponding author
    1. Division of Radiation Oncology, Taipei Veterans General Hospital, Taipei, Taiwan; Cancer Center, Taipei Veterans General Hospital, Taipei, Taiwan
    2. Department of Biomedical Imaging and Radiological Sciences, National Yang-Ming University, Taipei, Taiwan
    • Cancer Center, Taipei Veterans General Hospital, 201, Sec.2, Shih-Pai Road, Taipei 112, Taiwan; Fax (886) 2-28749425; shyen@vghtpe.gov.tw; OR Tai-Tong Wong, MD, Division of Pediatric Neurosurgery, Neurological Institute, Taipei Veterans General Hospital, 201, Sec.2, Shih-Pai Road, Taipei 112, Taiwan

    Search for more papers by this author

Abstract

BACKGROUND:

Previous intracranial germinoma (IG) studies have investigated the effect of different radiotherapy (RT) volumes and the necessity for adjunctive chemotherapy, but there is currently no consensus on the best treatment for this tumor.

METHODS:

From January 1989 to December 2009, 80 IG patients (≤20 years old) were treated with various RT regimens. Of them, 14 patients had craniospinal irradiation (CSI) + primary boost (PB); 8 patients had whole-brain irradiation (WBI) + PB; 31 patients had whole ventricular irradiation (WVI) + PB; and 27 patients had focal RT only. Twenty-nine patients (36.2%) also received systemic chemotherapy (CHT). Survival was estimated by the Kaplan-Meier method and variables affecting survival were analyzed by the Cox proportional hazard model.

RESULTS:

Eleven patients (13.8%) developed local recurrence or dissemination after treatment, and 10 of these patients were in the focal RT group. The 5-year relapse-free survival (RFS) for the CSI, WBI, WVI, and focal RT patients were 100%, 85.7%, 100%, and 84.6%, respectively (P = .001). The 5-year overall survival (OS) for CSI, WBI, WVI, and focal RT patients was 100%, 83.3%, 100%, and 87.9%, respectively (P = .125). Focal irradiation (P = .02) and initial use of CHT (P = .021) were negatively associated with RFS.

CONCLUSIONS:

Focal RT plus CHT were associated with inferior control of IG and a higher incidence of CHT-related toxicities. Adjustment of the radiation volume to the whole ventricular system without CHT is sufficient for treatment of nondisseminated IGs, even with lower primary RT doses (<36 Gy). Cancer 2011. © 2011 American Cancer Society.

INTRODUCTION

Intracranial germ cell tumors (GCTs) are brain tumors that occur in diverse histological types, are diagnosed primarily in preadults (<20 years old), and are the third most common type of brain tumor, behind gliomas and medulloblastomas.1-3 In Western populations, GCTs account for only 0.5% to 3% of all brain tumors,4, 5 but in Taiwan GCTs account for 10% of preadult brain tumors.2 Intracranial germinoma (IG) is the most common histologic type of GCT, and IGs account for about 60% to 70% of central nervous system GCTs.2, 6 Thus, clinical diagnosis of IG and the development of appropriate treatment regimens are critical.

In the past, when there was little understanding of IG, craniotomy with gross tumor excision was the primary treatment. However, neurosurgery is associated with significant risks, especially in children and adolescents. These tumors are now known to be highly sensitive to radiotherapy (RT) and chemotherapy (CHT), so surgery is currently utilized mainly for diagnosis or decompression of IGs.7 In particular, treatment by external beam RT or chemotherapeutic agents results in excellent disease control and increased overall survival (OS).8, 9 The current international guidelines for treatment of IG mainly recommend use of RT, possibly with adjunctive staggered CHT to reduce the side effects of RT. However, the optimal RT volume and dosage, and the role of systemic CHT remain controversial. Our multidisciplinary pediatric neuro-oncology team at Taipei Veterans General Hospital (VGHTPE), which treats more than half of all pediatric brain tumor patients in Taiwan, made 3 major changes in RT strategies for treatment of IG in the past 30 years. Before the early 1990s, we used large radiation volumes, including the whole craniospinal axis or whole brain, and high primary RT dosage (>50 Gy), based on the potential for neuraxis spreading of IGs. In the early 1990s, we attempted to prevent the side effects from large-field RT, and shifted to a platinum-based CHT alone based on the report that IGs were chemosensitive.10 However, more than half of our patients who received CHT alone experienced relapse within 1 to 2 years, even after initially complete remission.11 In the mid- to late 1990s, we mainly used combined focal low-dose radiotherapy (30 Gy) and CHT in an attempt to maintain disease control and reduce RT-related neural toxicities. During that period, some IG patients exhibited long-term disease control, despite the use of lower primary radiation doses (<40 Gy) and no CHT. In addition, most failure sites were within the ventricular system and outside the radiation field.

Based on these observations, we further adjusted the radiation volume and dose. Our current consensus treatment entails irradiation of the whole ventricular system (including the bilateral and lateral ventricles and the third and fourth ventricles) with low-dose radiation (30 Gy) using a 3-dimensional conformal technique (3D-CRT) without CHT.12 Young pediatric or adolescent patients who were given this treatment experienced no local recurrence or seeding in the neuraxis and no severe post-RT sequelae. Here, we report a comparison of the different RT and CHT strategies used to treat IG patients over the past 21 years at a single center in Taiwan.

MATERIALS AND METHODS

Patient Characteristics

A retrospective review of pediatric neuro-oncology patients treated at the VGHTPE between January 1989 and December 2009 was performed by review of the electronic RT registry. The institutional review board of the VGHTPE approved this study and all patients or their parents provided written informed consent. A total of 80 IG patients (<20 years of age) who had pathological or clinical diagnosis of IG and received various forms of RT were included. Six of these patients with initial seeding in the spinal column received CSI plus focal boost to the primary tumors and seeding sites followed by CHT. Before treatment, the spinal status was confirmed by cerebrospinal fluid (CSF) cytology or whole-spine magnetic resonance imaging (MRI) (after the mid-1990s). Table 1 summarizes the characteristics of the other 74 patients who had nondisseminated IG.

Table 1. Characteristics of Patients Who Received Treatment for Nondisseminated Intracranial Germinoma During Phase I, Phase II, and Phase III
Treatment Phase CharacteristicPhase I (N = 16)Phase II (N = 27)Phase III (N = 31)All Patients (N = 74)
  1. Abbreviations: DI, diabetes insipidus; hCG, human choriogonadotropin; KPS, Karnofsky performance score.

  2. Data in parenthesis are percentages in each group.

Age (y)    
 Range5-177-197-205-20
 Median13131112
Gender    
 Male14 (88)20 (74)23 (74)57 (77)
 Female2 (22)7 (23)8 (26)17 (23)
Performance status (KPS)    
 80-1009 (56)21 (78)25 (81)55 (74)
 60-702 (13)3 (11)5 (16)10 (14)
 <604 (25)1 (4)1 (3)6 (8)
 Not available1 (6)2 (7)0 (0)3 (4)
Tumor location    
 Sellar and suprasellar4 (25)11 (41)8 (26)23 (31)
 Pineal6 (38)5 (19)3 (10)14 (19)
 Thalamus/basal ganglion (unilateral/bilateral)4/1 (25/6)9/1 (33/4)9/1 (29/3)22/3 (30/4)
 Multifocal1 (6)1 (4)10 (32)12 (16)
Initial symptoms/signs    
 Headache4 (25)7 (23)3 (10)14 (19)
 Visual disturbance1 (6)5 (19)9 (29)15 (20)
 Movement disorder5 (31)11 (41)5 (16)21 (28)
 Pituitary dysfunction    
  Growth retardation1 (6)1 (4)3 (10)5 (7)
  DI1 (6)5 (19)8 (26)14 (19)
  Sex hormone disorder0 (0)2 (7)7 (23)9 (12)
Maximal tumor diameter (cm)    
 <33 (19)15 (56)15 (48)33 (45)
 3-58 (50)9 (33)14 (45)31 (42)
 >50 (0)2 (7)2 (7)4 (5)
 Not available5 (31)1 (4)0 (0)6 (8)
β-hCG level    
 >10 mIU/mL2 (13)3 (11)4 (13)9 (12)
 ≤10 mIU/mL14 (87)24 (89)27 (87)65 (88)
Diagnosis    
 Pathological14 (87)15 (56)15 (48)44 (59)
 Clinical2 (13)12 (44)16 (52)30 (41)

Radiotherapy and Treatment Phases

Before 1990, patients received gamma irradiation by a therapeutic cobalt unit (n = 6). After 1990, all patients were treated by linear accelerators with mega-voltage X-rays (4, 6, and 10 MV). After 2000, 3D-CRT was used. All patients received RT for 5 consecutive days each week.

Before the early 1990s (phase I), craniospinal irradiation (CSI) plus primary boost (PB; n = 8) or whole-brain irradiation (WBI) plus PB (n = 8) were the principal RT regimens, even without evidence of seeding in the neuraxis. PB was administered to the main tumor after CSI or WBI. The median total dose for CSI was 27 Gy (range, 25.5-28.5), the daily fraction was 1.5 to 1.8 Gy, and the median dose to the primary tumor was 50 Gy (range, 40.5-54). The median dose for WBI was 29.4 Gy (range, 8-30.9), the daily fraction was 1.8 to 2 Gy, and the median dose to the primary tumor was 50 Gy (range, 32.9-50.4).

From the mid- to late 1990s (phase II), 27 patients were treated with focal RT, which included the image-defined gross tumor and a 1.5-2 cm margin. In this group, the median dose to the primary tumor was 30 Gy (range, 28-50), and the daily fraction was 1.8 to 2 Gy. The median primary dose was lower in this group because approximately 50% of these patients also received systemic CHT. Since the early 2000s (phase III), the RT volume used to treat IGs gradually changed to whole ventricular irradiation (WVI) with a PB (n = 31). WVI encompassed the gross tumor and a 1.5- to 2-cm margin as well as the whole ventricular system (including the bilateral lateral ventricles and the third and fourth ventricles to the foramen of Magendie) for lesions located in the pineal, suprasellar, basal ganglia, and thalamus. An expansion of at least 0.5 cm was delineated from the rim of the ventricle lining, and unnecessary high-dose irradiation at the cortex was avoided as much as possible. In this group, the median dose for WVI was 24.5 Gy (range, 23.3-30), the daily fraction was 1.8 to 2 Gy, and the median dose to the primary tumor was 30 Gy (range, 29.3-50).

Chemotherapy

Twenty-nine (36.2%) of 80 patients also received systemic CHT before RT (n = 15), after RT (n = 17), or before and after RT (n = 3). This CHT consisted of 6 to 8 courses of the VBEP regimen (vincristine, bleomycin, etoposide, and cisplatin), with 5 consecutive days per course. No patient received CHT and RT concurrently. In phase I (extensive RT field), 9 of 16 patients received VBEP-CHT (neoadjuvant, 6 patients; adjuvant, 3 patients). In phase II (focal RT-field), 14 of 27 patients received CHT (neoadjuvant, 7 patients; adjuvant, 7 patients). Three of the patients in this phase II group received RT because of recurrence after initial CHT alone, although initial complete regression was observed. For the recurrent IG patients who had already received VBEP, the CHT regimen was changed to ICE (ifosfamide, cisplatin, and etoposide) for another 6 to 8 courses. In phase III (WVI), only 2 of 31 patients received VBEP after RT. However, patients with initial spinal seeding (n = 6) received the VBEP regimen after the initial extensive CSI.

Assessment of Treatment Results

Treatment results were retrospectively assessed by use of post-treatment computed tomography (CT) (before the 1990s) or MRI (after the 1990s) every 3 months in the first 2 years, every 6 months in the first 5 years, and then at 1-year intervals. All tumor lesions were measured according to Response Evaluation Criteria in Solid Tumors (RECIST)13 criteria. Relapse was defined as the new appearance of a tumor lesion in the primary or other site of the neuraxis after a complete response (CR). Adverse treatment effects were reviewed from available records and reassessed according to the latest National Cancer Institute Common Toxicity Criteria (version 4.0).14

Statistical Analysis

Descriptive statistics (medians and proportions) were used to characterize the patients and treatment results. Relapse-free survival (RFS) was the time from the last day of RT to the day of any evidence of relapse or the last follow-up. OS was the time from the last day of RT to the day of death or the last follow-up. Survival rates of the CSI, WBI, WVI, and focal RT groups were analyzed by the Kaplan-Meier method and compared by the log-rank test. Variables affecting survival were analyzed by the Cox proportional hazard model. A P value <.05 was defined as statistically significant.

RESULTS

Response to Therapy and Failure Patterns

Table 2 shows the outcomes of patients in the 5 different RT groups. The median follow-up time was 82.3 months (range, 5.1-236.8). Seventy-two patients (90%) experienced CR immediately after treatment and 8 patients experienced partial response (PR), with some residual tumors or small cyst-like lesions. For the 8 PR patients, the IGs shrank gradually, and complete regression was evident within 3 months of completion of RT or CHT. Only 1 patient in the late-response group experienced a relapse.

Table 2. Disease Control and Survival Rates of Intracranial Germinoma Patients in the Five Different Treatment Groups
Group OutcomeCSI+ PB (n = 8)CSI+ Boost for Initial Seeding (n = 6)WBI + PB (n = 8)Focal plus Margins (n = 27)WVI + PB (n = 31)Total (n = 80)
  1. Abbreviations: CSI, craniospinal irradiation; FU, follow-up; PB, primary boost; WBI, whole-brain irradiation; WVI, whole-ventricle irradiation.

Median FU (mo)/range145.8 (19.9-236.8)78.3 (19.3-142.3)106.5 (30.4-217.7)98.2 (5.1-177.8)54.7 (6.1-158.9)82.3 (5.1-236.8)
Relapse cases00110011
Mortality cases001405
Relapse-free survival (%)      
5 y100.0100.085.784.6100.091.4
10 y100.0100.085.779.9100.078.5
Overall survival (%)      
5 y100.0100.083.387.9100.094.0
10 y100.0100.083.379.1100.090.1

Eleven of the 80 patients experienced relapses after treatment (Table 3). Five of these patients had pathologic diagnoses, and 6 had clinical diagnoses. Ten of the 11 patients were in the focal RT group, 5 of whom were given CHT, and 6 of whom were not given CHT. Twelve relapse events were recorded, and 9 of these (75%) were outside of the RT field (primary site [n = 2], neuraxis [n = 4], ventricular system [n = 5], another cranial site [n = 1]). Relapse patient No. 11 underwent gamma knife radiosurgery (GKRS) as an initial treatment at another hospital, but relapsed 1 year later. He received GKRS again as a salvage treatment. Unfortunately, multiple dissemination in the neuraxis occurred 1 year after the second GKRS. At this point, the patient transferred to VGHTPE for salvage treatment. After CSI and systemic CHT, CR was achieved and has been maintained for approximately 20 months (at the time of this report). Among the 11 relapsed patients, CSI plus systemic CHT was the most common type of therapy. After salvage treatment, 7 patients (64%) had CR and remain disease-free. Only 5 IG patients expired after RT. The causes of death were RT-induced secondary angiosarcoma (n = 1), sepsis during scheduled CHT (n = 2), acute lymphoblastic leukemia (n = 1), and severe diarrhea unrelated to the disease (n = 1) (Table 3).

Table 3. Characteristics of Intracranial Germinoma Patients Who Experienced Treatment Failure
Patient o.AgeGenderPrimary Tumor LocationDiagnostic MethodsMaximal Tumor Diameter (cm)Pretreatment β-hCG LevelRT FieldPrimary RT Dose (Gy) and FractionsInitial Use of CHT as Adjunctive TreatmentFailure PatternDisease-Free Period (mo)Salvage TreatmentCurrent Status
  1. Abbreviations: ALL, acute lymphocytic leukemia; C, clinical; CHT, chemotherapy; CSI, craniospinal irradiation; F, female; FU, follow-up; GKRS, gamma knife radiosurgery; hCG, human choriogonadotropin; M, male; N, no; N/A, not available; NED, no evidence of disease; OP, operation; P, pathological; PB, primary boost; RT, radiotherapy; WB, whole brain; WNL, within normal limit; Y, yes.

15MPinealP< 3 ill- definedWNLWB + PB50.4/28NPrimary (In-field)23.0GKRSDied after RT induced secondary angiosarcoma
215MSuprasellarP2.5N/AFocal30/15NSpinal seeding (out-field)6.9CSI + CHTAlive and NED
315MRight basal ganglion, suprasellar and optic tractC>3; <5 ill- definedWNLFocal30/15YSpinal seeding (out-field)3.1CSI + CHTDied after ALL
412MPinealP2.0ElevatedFocal30/15NPrimary (In-field)136.2No TxLoss of FU
57MSuprasellarP1.5WNLFocal30/15N1. Base of the 4th ventricle (out-field)1.69.21. OP+ focal RTAlive and NED
          2. pituitary stalk (out-field)2. 83.12. CSI + CHT 
616MSuprasellarP5.0WNLFocal30/15YIntracranial and spinal seeding (out-field)7.6CSI + CHTDied after CHT induced sepsis
713FRight basal ganglion and suprasellarC5.2WNLFocal50/25YMultiple within ventricles (out-field)104.4CSI + CHTAlive and NED
815MLeft basal ganglionC1.7WNLFocal with margins encompassing partial ventricular system30/15YContralateral basal ganglion (out-field)88.9CSI+CHTAlive and NED
911MSuprasellarC< 3, ill- definedWNLFocal30/15YWithin ventricles (out-field)101.9CSI + CHTAlive and NED
1014MRight basal ganglionC2.5WNLFocal30/15NWithin ventricle (out-field)15.7OP + CS I +CHTAlive and NED
117MPinealC<3WNLGKRS × 2N/AN1. Primary (In-field)1. 12.0CSI + CHTAlive and NED
          2. Multiple neuraxis (out-field)2. 11.5  

The 5- and 10-year RFS for all patients who underwent RT were 91.4% and 78.5%, respectively, and the 5- and 10-year OS were 94% and 90.1%, respectively (Table 2). When stratified by treatment volume, the 5- and 10-year PFS and OS were 100% for the CSI (nondisseminated or disseminated) and the WVI groups but the PFS and OS were worse for the WBI and focal RT groups (Table 2). There was also a significant difference between the focal RT group and all other treatment groups in the 5- and 10-year RFS rates. Kaplan-Meier analysis and log-rank testing of the RFS curves indicated that this difference was significant (Fig. 1A; P = .001). However, there was no difference in OS curves because of the high IG salvage rates (Fig. 1B; P = .125).

Figure 1.

Shown are Kaplan-Meier curves and log-rank tests of (A) relapse-free survival (P = .001) and (B) overall survival (P= .125) in patients who received different radiation therapies.

Prognostic Factors

Table 4 summarizes the effect of various prognostic factors on RFS and OS. Univariate analyses indicated that focal RT (P = .02) and the initial use of CHT (P = .021) were negatively associated with RFS, but not OS. Other factors, such as age, gender, diagnostic method (pathological vs clinical), number of lesions (single, multifocal, dissemination), primary RT dose (≥36 Gy or <36 Gy), tumor size (≥3 cm or <3 cm), and elevated β-hCG (>10 mIU/mL), had no significant impact on OS or RFS in any of the treatment groups.

Table 4. Univariate Analysis of Risk Factors for Relapse-Free Survival and Overall Survival in Intracranial Germinoma Patients (n = 80)
Survival VariableRelapse-free survivalOverall Survival
HR (95% CI)PHR (95% CI)P
  1. Abbreviations: CI, confidence interval; CSI, craniospinal irradiation; HR, hazard ratio; RT, radiotherapy; WBI, whole-brain irradiation; WVI, whole-ventricular irradiation.

  2. β-hCG status: Within normal limit (≤10 mIU/mL); Elevation (>10 mIU/mL).

  3. aRepresents the baseline group for comparison.

Age (continuous, n = 80)0.909 (0.753-1.098).3210.890 (0.676-1.172).407
Gender    
 Male (n = 62)a    
 Female (n = 18)0.410 (0.052-3.212).3962.949 (0.487-17.848).239
Diagnostic methods    
 Pathological (n = 49)a    
 Clinical (n = 31)1.296 (0.394-4.269).6690.355 (0.040-3.182).355
Number of lesions    
 Single (n = 60)a    
 Multifocal or dissemination (n = 20)0.344 (0.044-2.708).311No significance1.000
Primary RT dose    
 ≥ 36 Gy (n = 27)a    
 < 36 Gy (n = 53)4.451 (0.912-21.718).0650.988 (0.160-6.115).990
Tumor size    
 ≥3 cm (n = 39)a    
 <3 cm (n = 35)2.392 (0.693-8.256).1680.423 (0.044-4.083).457
Initial chemotherapy as adjunctive treatment    
 Yes (n = 28)a    
 No (n = 52)0.256 (0.053-0.784).021No significance1.000
Irradiation volume    
 CSI/WBI (n = 22)a    
 WVI (n = 31)No significance1.000No significance1.000
 Focal (n = 27)11.682 (1.472-92.666).0203.520 (0.393-31.526).260
Beta-hCG status    
 Within normal limit (n = 71)a    
 Elevation (n = 9)1.941 (0.248-15.212).5281.525 (0.170-13.686).706

Toxicities

Hematologic complications were the most common adverse effects observed during the all 3 treatment phases. Pancytopenia was much more common in patients who received CHT plus non-CSI RT (n = 20), CSI alone (n = 5), or CHT + CSI (n = 9, 6 for initial seeding) than in the patients whose RT volume did not encompass the vertebral bone (non-CHT plus non-CSI RT, n = 46). Grades III, IV, and V febrile neutropenia were observed in the CHT plus non-CSI RT group (3 patients, 3 patients, 2 patients, respectively), in the CSI alone group (1 patient, 0 patients, 0 patients, respectively), and in the CHT + CSI group (4 patients, 2 patients, 0 patients, respectively). There were no cases of grade III, IV, or V febrile neutropenia in the non-CHT plus non-CSI RT group. Grades I, II, III, and IV anemia were observed in the CHT plus non-CSI RT group (9 patients, 6 patients, 2 patients, 2 patients, respectively), in the CSI alone group (3 patients, 2 patients, 0 patients, 0 patients, respectively), in the CHT + CSI group (3 patients, 2 patients, 2 patients, 1 patient, respectively), and in the non-CHT plus non-CSI RT group (8 patients, 4 patients, 1 patient, 0 patients, respectively).

Twenty-eight patients (35%) with various hormone deficiencies at diagnosis had tumors located in the sellar and suprasellar (neurohypophyseal) regions. During follow-up, no newly diagnosed hormonal deficiencies were evident after control of the IGs in all groups. Six of the 28 patients (21%) were able to reduce hormone supplementation after completion of all treatments. The other 22 patients required long-term administration of specific hormones (cortisone, antidiuretic hormone, growth hormone, or thyroid hormone). More than 50% of the patients (17 of 25) with tumors located initially in the basal ganglia or thalamus continued to experience movement disorders, such as hand dystonia or hemiparesis, even after completion of all treatments.

There was 1 radiation-induced secondary malignancy (relapse patient No. 1). This patient received initial WBI and PB for a pineal IG at age 5. He suffered from an angiosarcoma with rapid progression 2 years after the initial extensive RT and expired soon after failure of salvage treatment. Nine patients in all groups (CSI, 2 patients; WB, 1 patient; Focal, 5 patients; WVI, 1 patient) experienced difficulty in returning to school or resuming social life during the follow-up period. One patient in the CSI group with a high primary dose (>50 Gy) had serious mental impairment, although his IG was controlled over the long term. Regular examinations were not given to assess neurocognitive impairment, but 4 of the 80 patients demonstrated excellent school performance and were among the top 5 students in their classes (Focal RT, 1 patient; WVI, 3 patients).

DISCUSSION

IG is a highly curable tumor that typically occurs in young patients, and disease control and patient OS are generally very good.1, 8, 9, 12 Thus, treatment regimens that cause less treatment-related toxicity are crucial for these young patients, who are likely to experience long-term survival. In this study, most of the relapses (Table 3) occurred in the focal RT group (with or without systemic CHT), similar to other international case reports.15-17 Nine of our 11 failure cases were outside the RT fields, with no apparent relation to RT dosage, and 5 of these patients had intended to use adjunctive CHT to lower the primary RT dosage. This implies that an insufficient irradiation volume may be more dangerous than a low radiation dosage without CHT. An analysis of the relationship of the dose-volume and relapse (Table 5) indicated that the WVI group had a relatively low failure rate, similar to the CSI group (even with a primary RT dose <36 Gy). This suggests that our current RT regimen for treatment of patients with nondisseminated IG is an effective treatment.

Table 5. Dose and Volume Relationship for Relapse in Patients with Nondisseminated Intracranial Germinoma (n = 74)
Primary Radiotherapy Dose (Gy)Radiotherapy Volume
CSI + PBWBI + PBFocalWVI + PBTotal
  1. Abbreviations as in Table 2.

  2. Numbers in parenthesis are failure cases.

<400 (0)1 (0)24 (9)26 (0)51 (9)
40–504 (0)5 (0)0 (0)4 (0)13 (0)
≥ 504 (0)2 (1)3 (1)1 (0)10 (2)
Total8 (0)8 (1)27 (10)31 (0)74 (11)

Previous studies reported that extra-gonadal germinoma was as radiosensitive as germinomas that arise within the reproductive organs, and that 10-20 Gy delivered with a daily fraction size of 1.8-2 Gy can lead to significant tumor regression.12, 18, 19 We found that primary RT dose was not a definitive predictive factor for RFS or OS (Table 4). In particular, low-dose RT (<36 Gy) provided good long-term control of IG. Several Asian study reports of treatment of IGs with RT but not CHT have similar results.12, 20, 21 One previous study suggested low-dose RT (<50 Gy) might be a reason for treatment failure,22 but this study was performed more than 20 years ago and utilized radiation techniques that are now obsolete. Therefore, the optimal RT dose for treatment of IG remains somewhat controversial.

Early studies of IG treatment examined the effect of platinum-based CHT alone. However, more than 50% of these patients experienced recurrence within 1 to 2 years, even those who received an initial CR.10, 11 Therefore, CHT alone is clearly not an effective treatment for IG. Thus, more recent clinical trials of IG treatment examined the effect of induction CHT plus low-dose focal RT.5, 23 Although CHT can lower the RT dosage-related toxicities, CHT itself has some unacceptable adverse effects for young patients, such as bone marrow suppression, gastrointestinal disruption, and blood malignancies. Additionally, patients given this treatment experience a high incidence of tumor relapse outside the RT field, suggesting that CHT cannot independently control microscopic IGs and that RT volume is the major issue.5, 24 Although extensive RT (including CSI and WBI) leads to better control of IGs,20, 25 careful tailoring of the feasible RT volume is required, because it may also lead to toxicities such as neurocognitive impairment, hormonal dysfunction, RT-induced secondary tumors, and RT-induced vasculopathy.26-28

In the present study, we found that WVI plus PB is an effective regimen for treatment of IG. Most failures resulted from use of focal RT and occurred within the ventricular system (Table 3). These results are comparable with the current worldwide trend for the management of IG.21, 29 The most experienced GCT study group in Asia, the Japanese GCT study group (JGCTSG), also examined the use of WVI rather than focal RT after induction CHT.23, 30 Our results (median follow-up time, 54.7 months; Table 2) indicated that the WVI group had no recurrences and had a higher post-RT quality of life than the extensive RT group. Based on our own experience and on international trends to abandon CSI,31, 32 we suggest use of 20 to 24 Gy WVI plus PB (total dose, 30-36 Gy) without CHT for treatment of nondisseminated, low-risk IG (ie, no initial spinal seeding, no poor pathology components, no elevated beta-hCG levels at diagnosis [>200 mIU/mL], and rapid tumor regression after low-dose radiation).

In present study, the incidence of IGs in the basal ganglion or thalamus region was higher (25 of 80) than that in the suprasellar (23 of 80) or pineal regions (14 of 80), which were reported as the most common locations of IGs in Western populations.6 Some researchers recommended extensive whole brain RT for IGs arising from the basal ganglion or thalamus region.30 However, based on our large patient cohort series (n = 25; 5 for extensive RT), no WVI or focal RT IG patients experienced relapse in the cortical region (Fig. 2), and most early recurrent events occurred in the ventricular system (Table 3). Thus, we recommend against the use of large field irradiation because pediatric and adolescent patients are vulnerable to the extensive RT-related adverse effects and appear to experience no additional benefit. We further suggest that large field irradiation encompassing the normal brain cortex should be avoided in order to protect the neurocognitive function of patients who typically experience long-term survival.

Figure 2.

A 12-year-old boy with an intracranial germinoma that originated from the right basal ganglion and the thalamus region. After partial tumor removal and 30 Gy-whole ventricle irradiation only (no chemotherapy), a complete response was noted with 6 years of disease control (MRI, T1-weighted image with contrast enhancement). Coronal views before (A) and 6 years after treatment (B); axial views before (C) and 6 years after treatment (D).

CSI is mandatory for patients who have IGs with initial multiple disseminations so that the full extent of the tumor is covered. In our analysis, none of the 6 patients with disseminated IG experienced relapse after receiving CSI plus primary and seeding boost and 6 to 8 courses of adjuvant CHT. These results suggest that initial spreading in the whole neuraxis or multifocal IGs were not associated with poor prognosis. However, this finding should be viewed with caution due to the small number of patients and the retrospective nature of this study. Unlike other central nervous system malignancies, initial disseminated IGs can be controlled long-term by use of a suitable treatment protocol (CSI plus systemic CHT). None of our patients who had multifocal IGs (including double midline bifocal germinomas in the pineal and suprasellar regions) experienced poor outcome. Although the tumorigenesis of multifocal lesions may be synchronous or metastatic, our previous report showed that WVI plus PB without CHT was still associated with better prognosis in this subgroup of IG patients.33

We must acknowledge several limitations of our study. First, our study was retrospective, so our results must be considered tentative. Second, the patient group was somewhat heterogeneous, especially with regard to different treatment periods and tumor statuses. Third, the total number of enrolled patients was small. These limitations could all be overcome by use of a large prospective trial of patients with IGs; however, such a study is clearly impractical for such a rare tumor. Moreover, our results indicate that the WVI + PB regimen resulted in 100% 5-year and 10-year OS and RFS.

In conclusion, our results indicate that IGs are extremely radiosensitive tumors, can be effectively treated by RT alone, and that appropriate adjustment of the extent and dosage of irradiation can result in excellent treatment efficacy. Adjustment of the RT volume to include the whole ventricular system can reduce treatment-related side effects in pediatric and adolescent patients and provide excellent long-term survival and good quality of life.

FUNDING SUPPORT

This work was supported by grants from Taipei Veterans General Hospital (VGH98B1-006 and V100B-022); Department of Health, R.O.C. (DOH100-TD-C-111-007), and the Charity Foundation of JUT Land Development Group, Taiwan.

CONFLICT OF INTEREST DISCLOSURES

The authors made no disclosures.

Ancillary