Malignant extracranial germ cell tumors account for approximately. Brain Metastases of Malignant Germ Cell Tumors in Children and Adolescents

Transcription

1 620 Brain Metastases of Malignant Germ Cell Tumors in Children and Adolescents Sheri L. Spunt, M.D. 1,2 Michael F. Walsh, B.A. 3 Matthew J. Krasin, M.D. 4 Kathleen J. Helton, M.D. 4 Catherine A. Billups, M.S. 5 Alvida M. Cain, C.C.R.P. 1 Alberto S. Pappo, M.D. 1,2 1 Department of Hematology/Oncology, St. Jude Children s Research Hospital, Memphis, Tennessee. 2 Department of Pediatrics, The University of Tennessee College of Medicine, Memphis, Tennessee. 3 School of Medicine, University College Dublin, Dublin, Ireland. 4 Department of Radiological Sciences, St. Jude Children s Research Hospital, Memphis, Tennessee. 5 Department of Biostatistics, St. Jude Children s Research Hospital, Memphis, Tennessee. Supported in part by Cancer Center Grant CA and Cancer Center Support (CORE) Grant P30 CA from the National Cancer Institute and by the American Lebanese Syrian Associated Charities. The authors thank Sharon Naron, E.L.S., for her editorial assistance. Alberto Pappo s current address: Department of Pediatrics, The Hospital for Sick Children, Toronto, Ontario, Canada. Address for reprints: Sheri L. Spunt, M.D., Department of Hematology/Oncology, St. Jude Children s Research Hospital, 332 N. Lauderdale Street, Memphis, TN ; Fax: (901) ; sheri.spunt@stjude.org Received December 1, 2003; revision received April 29, 2004; accepted May 6, BACKGROUND. Brain metastases of pediatric germ cell tumors are uncommon, and there is limited information regarding their incidence, clinical presentation, response to treatment, and influence on survival. METHODS. The authors reviewed the experience with brain metastases from pediatric germ cell tumors at St. Jude Children s Research Hospital (Memphis, TN) over a 40-year period. RESULTS. Between March 1962 and February 2002, 16 of 206 patients with germ cell tumors (7.8%) had brain metastases at the time of initial presentation (n 2), later in the course of the illness (n 12), or at autopsy (n 2). Twelve of 16 patients (75%) had symptoms referable to the brain (nausea/emesis, headaches, or seizures), and 14 (88%) had pulmonary metastases at the time brain metastases were identified. Patients with brain metastases were more likely to have an extragonadal primary tumor (P 0.013), advanced-stage disease at initial presentation (P 0.016), and choriocarcinoma within the primary tumor (P 0.001). The incidence of brain metastases was significantly lower in the second 2 decades of the study period (5 of 135 patients [3.7%]) than in the first 2 decades (11 of 71 patients [15.5%]; P 0.005). Two of the 16 patients in the current study are long-term survivors. CONCLUSIONS. Brain metastases are uncommon in childhood germ cell tumors, and their incidence appears to be decreasing. In the current study, most patients with such metastases were symptomatic and had pulmonary metastases at the time brain metastases were identified. Patients with the highest risk of developing brain metastases include those with extragonadal tumors, those with high disease stage at initial presentation, and those with choriocarcinoma as a component of the primary tumor. The probability of survival is poor, although a small proportion of patients may become long-term survivors. Cancer 2004;101: American Cancer Society. KEYWORDS: neoplasms (germ cell and embryonal), germinoma, choriocarcinoma, endodermal sinus tumor, teratoma, neoplasm metastasis, infant, child, adolescent, brain. Malignant extracranial germ cell tumors account for approximately 5% of all malignancies in patients less than 20 years of age. 1 Most affected children become long-term survivors. 2 However, 15 30% of patients have distant metastases at the time of initial diagnosis, 2,3 most commonly involving the regional lymph nodes, lungs, liver, or bones. Brain metastases are uncommon, and the literature on pediatric patients offers little information about the incidence, clinical presentation, response to treatment, and effect on survival associated with these metastases. In adults, most of the information on brain metastases in patients with germ cell tumors is derived from studies of patients with testic American Cancer Society DOI /cncr Published online 22 June 2004 in Wiley InterScience (

2 Brain Metastases in Childhood Germ Cell Tumors/Spunt et al. 621 ular and extragonadal primary tumors. At the time of initial presentation, brain metastases are identified in approximately 1.3% of this population. 4 Among adult patients with testicular germ cell tumors, those with brain metastases usually had advanced-stage disease at initial presentation and have concomitant pulmonary metastases at the time of detection of brain involvement. 5 The presence of choriocarcinoma or yolk sac histology also appears to predispose patients to the development of brain metastases. 6 Although outcome historically has been poor, several recent reports suggest that long-term survival can be achieved with aggressive local and systemic treatment. 5,7,8 We reviewed the experience at St. Jude Children s Research Hospital (Memphis, TN) over a 40-year period. Our goals were to more completely characterize brain involvement in childhood malignant germ cell tumors, to identify risk factors for the development of this complication, and to assess factors associated with disease control at this site. MATERIALS AND METHODS Patients We reviewed the medical records of all children and adolescents with germ cell tumors metastatic to the brain who were treated between March 1962 and February 2002 at St. Jude Children s Research Hospital. Patients were identified from a database containing information collected prospectively on all patients with solid tumors. Each patient s database record is periodically updated throughout the duration of treatment and follow-up. For each patient with a germ cell tumor and brain metastases, we evaluated demographic data (gender, race, age at the time of diagnosis, and duration of follow-up), primary tumor site (obtained from the initial tumor biopsy or resection report), and tumor histology (obtained from the associated pathology report). The surgical findings, results of diagnostic imaging studies, and measurement of serum tumor markers were used to assign a stage of disease according to the Children s Oncology Group system. 9 Evaluation of disease extent at the time of initial diagnosis varied according to the site of the primary tumor and the year of diagnosis. Depending on the available technology, the primary tumor and regional lymph nodes were imaged using X-ray, ultrasonography, computed tomography (CT), or magnetic resonance imaging (MRI). All patients received an X- ray or CT scan of the chest to identify pulmonary metastases. For some patients, a skeletal survey or 99m Tc bone scan was used to detect bone metastases, and some patients underwent CT or MRI scanning of the head to screen for brain metastases. Therapy The initial surgical approach depended on the site of the primary tumor. Patients with testicular tumors underwent inguinal orchiectomy with or without retroperitoneal lymph node sampling or dissection. Patients with ovarian tumors underwent unilateral salpingo-oophorectomy, biopsy of the contralateral ovary, collection of pelvic fluid for cytologic evaluation, and abdominal exploration with biopsy of suspicious areas. Patients with extragonadal primary tumors underwent tumor biopsy and/or resection with or without regional lymph node sampling or dissection. We recorded the types of chemotherapy and radiotherapy administered to each patient and reviewed surgical notes and pathology reports to identify any surgical procedures undertaken after the initial diagnosis. Where applicable, we noted the timing and pattern of treatment failure and any subsequent therapy administered. Brain Metastases For each patient, we determined whether brain metastases were identified at the time of initial diagnosis, during treatment or follow-up, or at the time of autopsy. We also made note of which patients brain metastases were detected incidentally by screening imaging studies. The presenting signs and symptoms of brain involvement in symptomatic patients were noted, as were the tests used to establish the diagnosis of metastatic brain disease. The type of treatment administered after identification of brain metastases was recorded, as was the extent of control of brain involvement. Among surviving patients, we sought to identify any long-term effects associated with these brain metastases or their treatment. The autopsy reports of patients who had died were reviewed, if available. Statistical Considerations Correlations involving categoric variables were examined using the Fisher exact test. Age at diagnosis was compared between patients with and without brain metastases using the exact Wilcoxon rank sum test. Differences between these groups in disease stage were examined using the exact Kruskal Wallis test (stage was considered an ordinal variable). The duration of survival was defined as the interval between the date of diagnosis and the date of death due to any cause or the date of last follow-up, and the probability of survival was estimated using the method of Kaplan and Meier. Survival distributions were compared using the exact log-rank test.

3 622 CANCER August 1, 2004 / Volume 101 / Number 3 TABLE 1 Treatment of Brain Metastases and Outcomes in 16 Patients with Malignant Germ Cell Tumors Patient no. Time to development of brain metastases (mos) Chemotherapy Radiotherapy Outcome (mos after diagnosis brain metastases) Cause of death 1 25 VAC 20 Gy partial-brain DOD (3) Brainstem herniation 2 4 None None DOD (0) Brainstem herniation 3 30 None 35 Gy partial-brain DOD (5) Brainstem herniation 4 At autopsy None None DOD (0) Respiratory failure 5 7 Doxorubicin None DOD (2) Brainstem herniation 6 17 None None DOD (1) Tumor progression, but exact cause unknown 7 22 Amsacrine None DOD (1) Brainstem herniation 8 7 Cisplatin/doxorubicin 35 Gy whole-brain 3.5 Gy boost DOD (4) Respiratory failure 9 At autopsy None None DOD (0) Respiratory failure 10 3 None None DOD (0) Brainstem herniation 11 7 Etoposide 30 Gy whole-brain 6 Gy boost DOD (4) Respiratory failure Etoposide 34 Gy whole-brain 14.8 Gy boost NED (223) N/A 13 2 PVB 33 Gy whole-brain 12 Gy boost DOD (19) Tumor progression, but exact cause unknown 14 At diagnosis FLMP None DOD (0) Respiratory failure 15 At diagnosis a PEB None NED (37) N/A a None Dose unknown DOD (5) Respiratory failure DOD: died of disease; FLMP: 5-fluorouracil/leucovorin/methotrexate/cisplatin; Gy: grays, N/A: not applicable; NED: no evidence of disease; PEB: cisplatin/etoposide/bleomycin; PVB: cisplatin/vinblastine/ bleomycin; VAC: vincristine/actinomycin D/cyclophosphamide. a Patient was asymptomatic. Brain metastases were detected on surveillance neuroimaging. RESULTS Clinicopathologic Features and Treatment Of 12,308 patients evaluated for malignancy between March 1962 and February 2002 at St. Jude Children s Research Hospital, 206 (1.7%) had a diagnosis of germ cell tumor. Of these 206 patients, 16 (7.8%) experienced metastatic involvement of the brain at some point. Brain metastases were identified at the time of initial presentation (n 2), during the course of the illness (n 12), or at autopsy (n 2). The median age at the time of initial presentation was 6.5 years (range, years). Ten patients (63%) were male, and 15 patients (94%) were white. The primary tumor arose in the ovary or testis in 5 patients and at extragonadal sites in 11 patients. More than one malignant component of the tumor was identified in most cases. The most common malignant components were yolk sac tumor (n 9) and choriocarcinoma (n 7). Eight of the 16 patients had distant metastases at the time of initial diagnosis. After initial surgery, 14 patients received chemotherapy (n 9), radiotherapy (n 2), or both (n 3). Two patients with Stage I testicular disease received no adjuvant treatment. The most common chemotherapeutic regimens used were vincristine, dactinomycin, and cyclophosphamide (n 5) and cisplatincontaining regimens (n 4). Brain Metastases Data on the timing and treatment of brain metastases are shown in Table 1. Brain metastases were identified a median of 7 months (range, 0 37 months) after initial diagnosis. In Patients 4 and 9, brain metastases were identified at autopsy. Two additional patients (Patients 15 and 16) had no symptoms referable to the central nervous system (CNS) and brain involvement was identified by screening imaging studies at the time of initial diagnosis (n 1) or tumor recurrence (n 1; Fig. 1). Brain metastases were diagnosed clinically in 12 patients. These 12 patients had some combination of nausea/emesis, headaches, and seizures. Three patients had neck or back pain, one patient reported changes in vision, and one infant had increasing head circumference. Neurologic abnormalities were detectable on physical examination in 6 of these 12 patients (50%). The imaging studies used to identify brain metastases varied with the era of presentation and with physician preference. CT scans were used most frequently (in 8 of 14 patients diagnosed using diagnostic imaging methods [57%]). Most brain metastases were supratentorial (n 9) or involved both supratentorial and infratentorial regions (n 4). Pulmonary metastases were present at the time brain metastases were documented in 14 of 16 patients (88%). Of the two

4 Brain Metastases in Childhood Germ Cell Tumors/Spunt et al. 623 patients who had no pulmonary metastases, one had a large mediastinal tumor, and one had tumor involvement of the spinal meninges. Five patients received no treatment, either because brain metastases were not detected until autopsy (n 2) or because of advanced metastatic disease (n 3). The remaining 11 patients were treated with chemotherapy (n 4), radiotherapy (n 2), or both (n 5). None of the patients underwent surgical resection of brain metastases. Comparison of Patients Who Had Germ Cell Tumors with Brain Metastases and Patients Who Had Germ Cell Tumors without Brain Metastases We compared the clinical features of the 16 patients who had brain metastases with those of the other 190 patients with germ cell tumors who were treated at St. Jude Children s Research Hospital. Patients with brain metastases were significantly more likely to have an extragonadal tumor (P 0.013), to have advancedstage disease at initial presentation (P 0.016), and to have a component of choriocarcinoma within the tumor (P 0.001). Patients with brain metastases were also slightly more likely to be male (P 0.109). We found no difference between the two groups in terms of age or race. Of note was the finding that the proportions of males (34 of 84 [40%]) and females (43 of 122 [35%]) with extragonadal tumors were similar (P 0.47). The incidence of brain metastases in patients treated for germ cell tumors was significantly lower in the more recent 2 decades of the study period (5 of 135 [3.7%]) than in the earlier 2 decades (11 of 71 [15.5%]; P 0.005). Because the 4 patients with brain metastases detected by screening imaging studies or at autopsy were clinically indistinguishable from the other 190 patients with germ cell tumors, we repeated the comparison of patients with and without brain metastases after excluding these 4 patients. The results were virtually identical to the findings reported earlier in the current report. FIGURE 1. Imaging studies performed to screen asymptomatic patients for brain metastases. (A) Magnetic resonance imaging scan. Axial gradient-echo image (Patient 15) shows two right frontal hemorrhagic foci at the gray-white junction. Additional punctate lesions that enhanced minimally were detected in the left frontal, temporal, and parietal lobes and in the right occipital lobe (not shown). The study was performed at the time of initial presentation in a patient with widespread lung, liver, and retroperitoneal lymph node metastases. The patient s primary testicular tumor was composed exclusively of choriocarcinoma, a histologic type known to be associated with hemorrhagic brain metastases. 23 (B) Computed tomographic scan. Enhanced axial image (Patient 16) shows a2cm 2 cm, slightly hyperdense, vividly enhancing mass in the right frontal lobe extending into the genu of the corpus callosum. Image was obtained at the time of tumor progression in a patient with extensive pulmonary metastases. Outcome Figure 2 shows the survival estimates of patients with germ cell tumors who did and did not have brain metastases. The median period of follow-up for all survivors (n 140) was 13.5 years. Patients with brain metastases had a poor outcome. Only two remained alive at the time of the current study, both without evidence of disease. One (Patient 15), who had asymptomatic brain metastases at initial presentation, received cisplatin-based chemotherapy alone and remained free of disease 4 years after initial diagnosis. The other (Patient 12) was free of disease 18 years after initial diagnosis, having received etoposide chemo-

5 624 CANCER August 1, 2004 / Volume 101 / Number 3 FIGURE 2. Survival of 206 patients with germ cell tumors who did (n 16) or did not (n 190) have brain metastases. therapy and whole-brain radiotherapy with a boost to the tumor bed for brain metastases that developed 17 months after the initial diagnosis of germ cell tumor. Neither of these patients has any discernible long-term complications referable to brain metastases or their treatment. The 9 patients who died after receiving treatment for brain metastases survived for a median of 4 months (range, 0 19 months) after detection of brain metastases. Four of these nine patients died as a direct result of tumor involvement of the brain, whereas the remaining five died of other complications or of unknown causes. DISCUSSION This single-institution review of 206 children and adolescents with malignant germ cell tumors showed that brain metastases are present at the time of initial diagnosis in approximately 1% of patients and develop at some point in approximately 8% of patients. In a previous report from our institution, only patients with melanoma and rhabdoid tumors of the kidney had a greater risk of developing brain metastases. 10 The 1% incidence of brain metastases among newly diagnosed pediatric patients with germ cell tumors in the current series is quite similar to the 1.3% incidence reported in a large adult series of patients with testicular and extragonadal germ cell tumors. 4 Although there are published studies on children with brain metastases of neuroblastoma, 11 rhabdomyosarcoma, 12,13 osteosarcoma, 14,15 Ewing sarcoma family of tumors, 12,16 Wilms tumor, 17 and melanoma, 18 to our knowledge, no report has specifically addressed brain metastases in children with germ cell tumors. In 2 recent series comprising 640 children with solid tumors treated between 1982 and 1995, none of the 19 patients with brain metastases had germ cell tumors. 19,20 We found that most patients with germ cell tumors and brain metastases had symptoms and signs referable to the CNS, such as nausea/emesis, headaches, seizures, and focal neurologic findings. However, some patients in the current series had asymptomatic brain metastases that were identified by screening imaging studies or were detected at the time of autopsy. Because many patients did not undergo brain imaging in the absence of symptoms and may have been cured with chemotherapy alone, it is possible that the current study underestimated the incidence of brain metastases. Whether asymptomatic brain metastases are common in children treated with modern chemotherapy regimens is unknown. However, our observation that the incidence of brain metastases was lower in the more recent 2 decades of the study period suggests that the advent of more effective platinum-based chemotherapy may have decreased the incidence of metastatic spread to the brain. The finding that two asymptomatic patients in the current series had brain metastases detected at autopsy also suggests that brain involvement may be more common than is clinically suspected, particularly among children with refractory disease. A relatively small proportion of patients in the current series who died of germ cell tumors underwent an autopsy, and an even smaller subset underwent an autopsy that included an evaluation of the brain. Thus, it is difficult to accurately assess the incidence of brain metastases in the current population of children who died of germ cell tumors. However, a 1983 autopsy series showed that four of eight children who died of germ cell tumors had brain involvement. 21 We identified several tumor characteristics that were associated with an increased risk of brain metastases. Patients whose primary germ cell tumors arose in extragonadal sites, who had advanced-stage disease at initial presentation, or whose primary tumors had choriocarcinoma as a component had the greatest risk. Almost all patients with brain metastases (14 of 16) had pulmonary metastases at the time brain involvement was detected. Exceptions included one child with a large mediastinal primary tumor and one with leptomeningeal involvement of a sacrococcygeal tumor. Isolated brain involvement as the first site of tumor recurrence has been reported previously in a child with a mediastinal germ cell tumor, although this patient also had pulmonary involvement at initial presentation. 22 Risk factors for the development of brain metastases in adults with germ cell tumors are

6 Brain Metastases in Childhood Germ Cell Tumors/Spunt et al. 625 similar. In a series of 44 adults with testicular germ cell tumors who had brain metastases, most had advanced-stage disease (84%) and pulmonary tumor involvement (89%). 5 The association between choriocarcinoma and an elevated risk of brain metastases also has been observed previously. 6 On the basis of our findings, we recommend that children with advanced-stage germ cell tumors that contain a choriocarcinoma component and children with pulmonary metastases undergo screening imaging for brain metastases. Brain imaging should also be considered for pediatric patients with advanced-stage mediastinal primary tumors and those with spinal leptomeningeal tumor involvement. Only two patients in this study were long-term survivors. One received chemotherapy alone for brain metastases present at initial diagnosis, and the other received chemotherapy and whole-brain radiotherapy with a boost to the tumor bed for brain metastases that developed 17 months after initial presentation. Our limited data preclude any definitive recommendations regarding the optimal treatment of brain metastases for children with germ cell tumors. The approach to such treatment is controversial in adults. In a large study of brain metastases in adults with testicular germ cell tumors, univariate analysis showed combined chemotherapy and radiotherapy to be more effective than either modality alone. 5 However, in a second study, radiotherapy was not independently predictive of favorable outcome in patients treated with cisplatin-based chemotherapy for brain metastases that were evident at initial presentation. 7 Irradiation of the brain must be used cautiously in children, given its potential morbidity and the lack of definitive evidence that it is necessary in all cases. Because brain metastases may be cured by chemotherapy alone in some patients, a trial of chemotherapy to assess response may be indicated before radiotherapy is administered. High-dose chemotherapy with autologous stem cell transplantation has not been studied in childhood germ cell tumors associated with brain metastases. However, a recent study in adults with germ cell tumors and brain metastases at initial diagnosis found high progression-free and overall survival rates after frontline high-dose chemotherapy with autologous stem cell transplantation. 8 Whether this approach is superior to standard treatment strategies remains to be proven. Although the incidence of brain metastases of childhood germ cell tumors appears to be decreasing, the outlook for patients with this complication remains poor. Further studies are needed to define the treatment approach that provides the best chance of cure along with an acceptable toxicity profile. REFERENCES 1. Ries LA, Smith MA, Gurney JG, et al. Cancer incidence and survival among children and adolescents: United States SEER program Bethesda, MD: National Cancer Institute, Mann JR, Raafat F, Robinson K, et al. The United Kingdom Children s Cancer Study Group s second germ cell tumor study: carboplatin, etoposide, and bleomycin are an effective treatment for children with malignant extracranial germ cell tumors, with acceptable toxicity. J Clin Oncol. 2000;18: Marina N, Fontanesi J, Kun L, et al. Treatment of childhood germ cell tumors. Review of the St. Jude experience from 1979 to Cancer. 1992;70: International Germ Cell Cancer Collaborative Group. International Germ Cell Consensus Classification: a prognostic factor-based staging system for metastatic germ cell cancers. J Clin Oncol. 1997;15: Bokemeyer C, Nowak P, Haupt A, et al. Treatment of brain metastases in patients with testicular cancer. J Clin Oncol. 1997;15: Williams SD, Einhorn LH. Brain metastases in disseminated germinal neoplasms: incidence and clinical course. Cancer. 1979;44: Fossa SD, Bokemeyer C, Gerl A, et al. Treatment outcome of patients with brain metastases from malignant germ cell tumors. Cancer. 1999;85: Kollmannsberger C, Nichols C, Bamberg M, et al. First-line high-dose chemotherapy radiation therapy in patients with metastatic germ-cell cancer and brain metastases. Ann Oncol. 2000;11: Cushing B, Perlman EJ, Marina NM, Castleberry RP. Germ cell tumors. In: Pizzo PA, Poplack DG, editors. Principles and practice of pediatric oncology (4th edition). Philadelphia: Lippincott Williams & Wilkins, 2002: Spunt SL, Thompson SJ, Pappo AS. Brain metastases in paediatric solid tumours. Forum (Genova). 2001;11: Kramer K, Kushner B, Heller G, Cheung NK. Neuroblastoma metastatic to the central nervous system. The Memorial Sloan-Kettering Cancer Center experience and a literature review. Cancer. 2001;91: Parasuraman S, Langston J, Rao BN, et al. Brain metastases in pediatric Ewing sarcoma and rhabdomyosarcoma: the St. Jude Children s Research Hospital experience. J Pediatr Hematol Oncol. 1999;21: Spunt SL, Anderson JR, Teot LA, Breneman JC, Meyer WH, Pappo AS. Routine brain imaging is unwarranted in asymptomatic patients with rhabdomyosarcoma arising outside of the head and neck region that is metastatic at diagnosis. A report from the Intergroup Rhabdomyosarcoma Study Group. Cancer. 2001;92: Baram TZ, van Tassel P, Jaffe NA. Brain metastases in osteosarcoma: incidence, clinical and neuroradiological findings and management options. J Neurooncol. 1988;6:47 52.

7 626 CANCER August 1, 2004 / Volume 101 / Number Marina NM, Pratt CB, Shema SJ, Brooks T, Rao B, Meyer WH. Brain metastases in osteosarcoma. Report of a longterm survivor and review of the St. Jude Children s Research Hospital experience. Cancer. 1993;71: Shuper A, Cohen IJ, Mor C, Ash S, Kornreich L, Zaizov R. Metastatic brain involvement in Ewing family of tumors in children. Neurology. 1998;51: Lowis SP, Foot A, Gerrard MP, et al. Central nervous system metastasis in Wilms tumor. A review of three consecutive United Kingdom trials. Cancer. 1998;83: Rodriguez-Galindo C, Pappo AS, Kaste SC, et al. Brain metastases in children with melanoma. Cancer. 1997;79: Bouffet E, Doumi N, Thiesse P, et al. Brain metastases in children with solid tumors. Cancer. 1997;79: Tasdemiroglu E, Patchell RA. Cerebral metastases in childhood malignancies. Acta Neurochir (Wien). 1997;139: Graus F, Walker RW, Allen JC. Brain metastases in children. J Pediatr. 1983;103: Berkow RL, Kelly DR. Isolated CNS metastasis as the first site of recurrence in a child with germ cell tumor of the mediastinum. Med Pediatr Oncol. 1995;24: Kobayashi T, Kida Y, Yoshida J, Shibuya N, Kageyama N. Brain metastasis of choriocarcinoma. Surg Neurol. 1982;17: