Retrospective multi-institutional study of radiotherapy for intracranial non-germinomatous germ cell tumors

Radiotherapy and Oncology 49 (1998) 55–59

Retrospective multi-institutional study of radiotherapy for intracranial non-germinomatous germ cell tumors Hidefumi Aoyama a ,*, Hiroki Shirato a, Hiroshi Yoshida b, Masato Hareyama c, Masamichi Nishio d, Touru Yanagisawa e, Yoshihisa Kakuto f, Jirou Watarai g, Hideki Inakoshi h, Kazuo Miyasaka a a

Department of Radiology, Hokkaido University School of Medicine, North-15, West-7, Kita-Ku, Sapporo 0608638, Japan b Department of Radiology, Asahikawa Medical College, Asahikawa, Japan c Department of Radiology, Sapporo Medical College, Sapporo, Japan d Department of Radiology, Sapporo National Cancer Center, Sapporo, Japan e Department of Radiology, Iwate Medical College, Iwate, Japan f Department of Radiology, Tohoku University School of Medicine, Sendai, Japan g Department of Radiology, Akita University School of Medicine, Akita, Japan h Department of Radiology, Niigata University School of Medicine, Niigata, Japan Received 19 August 1997; revised version received 25 May 1998; accepted 9 July 1998

Abstract The treatment outcome of 24 patients with pathologically-proven non-germinomatous germ cell tumor was retrospectively investigated to determine the effectiveness of radiotherapy. The patients were divided into three groups as follows: group 1, five patients with mature teratoma with or without germinoma; group 2, six patients with immature teratoma with or without germinoma; group 3, 13 patients with other highly malignant tumors. The overall actuarial survival and relapse-free rates at 5 years were 82% and 59%, respectively, with a median follow-up period of 62 months. The actuarial relapse-free rate at 5 years was 100% for group 1, 63% for group 2 and 44% for group 3. There was no difference in the relapse-free rates between total resection and partial resection. Usage of chemotherapy was adversely related to survival probably due to selection bias. No local failure was observed with 10 Gy or more for group 1, 40 Gy or more for group 2 and 54 Gy or more for group 3. In groups 1 and 2, there was no spinal relapses without craniospinal irradiation. In group 3, three of eight patients who did not receive craniospinal irradiation and none of five patients who received craniospinal irradiation experienced spinal relapse. In conclusion, highly malignant GCTs show a high incidence of spinal metastasis and craniospinal irradiation may reduce the risk of spinal metastasis. Radiation dose and volume are to be determined according to the histopathological aggressiveness.  1998 Elsevier Science Ireland Ltd. All rights reserved Keywords: Brain neoplasm; Non-germinomatous germ cell tumor; Radiotherapy

1. Introduction Intracranial non-germinomatous germ cell tumors (NGGCTs) are very rare lesions that are responsible for less than 0.5% of all central nervous system (CNS) tumors [9]. In contrast to the high curability of pure germinoma, the prognosis for intracranial NGGCTs, except for mature teratoma, has been reported to be dismal, even after multimodal treatment. The poor treatment results have been

* Corresponding author.

attributed to the high relapse rate after conventional radiotherapy and surgery. The use of chemotherapy for intracranial NGGCTs has been suggested as it has been effective in the treatment of testicular non-seminomatous germ cell tumors, although there have been contradictory report as to the success of this approach [11]. NGGCTs are often included in one category because of their rarity, but the patterns of relapse are reported to be different among subtypes [3,6,12,16]. We analyzed intracranial NGGCTs retrospectively to identify the prognosis and to investigate the optimal irradiation method for each histopathological subtype of this rare disease.

0167-8140/98/$19.00  1998 Elsevier Science Ireland Ltd. All rights reserved PII S0167-8140 (98 )0 0081-4

56

H. Aoyama et al. / Radiotherapy and Oncology 49 (1998) 55–59

2. Materials and methods A retrospective multi-institutional study was performed among seven university hospitals and one national cancer center in northern Japan. Twenty-four patients were pathologically confirmed at the initial treatment to have had primary intracranial NGGCT between 1971 and 1994. There were 21 males and three females ranging in age from 7 to 24 years (median 14.2 years). Four patients with mature teratoma (MT), five patients with immature teratoma (IMT), two patients with teratoma with malignant transformation, three patients with embryonal carcinoma (EC), one patient with yolk sac tumor (YST) and nine patients with mixed germ cell tumors were included. A solitary tumor was found in 19 patients and multi-focal tumors were found in five patients. The most common location was the pineal region (16/24) followed by the suprasellar region (5/24). Tumors in other sites were seen in three patients at the thalamus, the basal ganglia and the third ventricle. Cerebrospinal fluid cytology was examined in 14 patients before treatment and one (7%) patient with a mixed germ cell tumor (EC/ MT) was found to have tumor cells in the cerebrospinal fluid. One patient was initially diagnosed by ventriculography, 23 patients were diagnosed using computed tomography (CT) and 11 patients were also examined by magnetic resonance imaging. We divided the patients into three groups according to the histopathological component as follows: group 1, MT with or without germinoma (five patients); group 2, IMT with or without germinoma (six patients); group 3, highly malignant tumors including EC, YST, choriocarcinoma and teratoma with malignant transformation (13 patients). There is a classification for NGGCT according to hypothetical embryonic development [13]. The possible neoplasms which may arise at each step of embryo development are choriocarcinoma, yolk sac tumor, embryonal carcinoma and teratoma (in that order). For mixed NGGCTs, patients were included in one of the three groups of more primitive precursor cells according to the hypothetical embryonal development in this study, depending on a hypothesis that prognosis may be affected by the most primitive component of each tumor. There were various types of initial treatment employed. Total resection was used in 15 patients, i.e. four patients in group 1, four patients in group 2 and seven patients in group 3. Partial resection was used in nine patients, i.e. one patient in group 1, two patients in group 2 and six patients in group 3. Radiotherapy was used as part of initial treatment in 22 patients. Local field was used in two patients in group 1, three patients in group 2 and one patient in group 3. Wholeventricle irradiation was used in one patient in groups 2 and 3. Whole-brain irradiation was used in three patients in group 1, one patient in group 2 and four patients in group 3. Craniospinal irradiation was used in one patient in group 2 and five patients in group 3. The median total dose of irradiation to the primary tumor site (including two patients

who did not receive radiotherapy) was 29 Gy (range 10–45 Gy) in group 1, 35 Gy (range 16–55 Gy) in group 2 and 43 Gy (range 0–60 Gy) in group 3. The spinal dose in the craniospinal irradiation ranged from 26 to 40 Gy. Daily doses ranged from 1.8 to 3.0 Gy with a median of 2.0 Gy. Thirteen (54%) of the 24 patients received various types of chemotherapy. Patients in group 3 received chemotherapy more often than patients in groups 1 and 2. One (20%) of five patients in group 1 received chemotherapy which did not consist of platinum compounds after radiation therapy, three (50%) of six patients in group 2 received chemotherapy which consisted of platinum compounds (one patient after and two patients before radiation therapy) and nine (69%) of 13 patients in group 3 received platinum-containing chemotherapy (four patients before, four patients after and one patient during radiation therapy). The follow-up time ranged from 1 to 155 months with a median of 62 months. Relapse rates were compared either by crude or actuarial methods. Actuarial survival and relapse-free rates were calculated using the Kaplan–Meier method and the curves were compared by the log-rank method. The range for the follow-up time was calculated from the first day of initial treatment. A x2-test was used to compare the crude relapse rates. The log-rank test was used for the comparison of variants.

3. Results 3.1. Overall treatment outcome On the whole, the actuarial survival rate at 5 years was 82%. The actuarial relapse-free rate at 5 years was 59%. There was no significant difference in the actuarial relapsefree rate between patients with solitary lesions and patients with either multi-focal tumor or CSF dissemination. There was no significant difference in the actuarial survival rate and the relapse-free rate between the patients who received total resection and the patients who received partial resection. The actuarial relapse-free survival of patients who received chemotherapy was significantly lower than those who did not receive chemotherapy (P = 0.03). The number of patients was too small to analyze the effect of chemotherapy in each histopathological group. The actuarial relapsefree curves were divided into three groups according to the histopathological group and are shown in Fig. 1. The actuarial relapse-free rate at 5 years for group 1 was 100%. One patient in group 1 had experienced an intracranial relapse of germinoma at 8 years and was alive at 5 years after irradiation for the relapse. The actuarial relapse-free rate at 5 years for group 2 was 63%. Two of the six patients with group 2 tumors had died. One patient died within 1 month after resection due to an increasing residual tumor and another patient died 5 months after resection resulting from regrowth of the primary tumor. The actuarial relapse-free survival at 5 years for group 3 was 44%. All three patients

57

H. Aoyama et al. / Radiotherapy and Oncology 49 (1998) 55–59

Fig. 1. Actuarial relapse-free survival curves according to the pathological group. Group 1, mature teratoma with or without germinoma; group 2, immature teratoma with or without germinoma; group 3, teratoma with malignant transformation and tumors including components of embryonal carcinoma, yolk sac tumor and/or choriocarcinoma.

with pure EC experienced a relapse after 16, 22 and 78 months. One patient died 3 months after the relapse and the other two patients were alive 1 and 46 months after the relapse. A patient with pure YST experienced a relapse 5 months after the treatment and died 9 months after the relapse. Two patients with teratoma with malignant transformation remained free from relapse at 14 and 31 months after treatment. Four of the other seven patients with mixed germ cell tumors in group 3 experienced relapse and one of them died. 3.2. Radiotherapy schedule and relapse pattern (Table 1) Overall, relapses occurred in 11 (46%) patients during the study period. Nine relapses occurred within 5 years of the initiation of treatment, but two patients (one with MT and the other with EC) experienced relapses after 5 years. Five relapses were seen at the primary site. Four primary relapses

(IMT in two cases, EC/IMT/G and EC/MT) were detected without any other relapse and one patient with EC experienced both local and spinal canal relapses at the same time. The radiation dose and primary tumor control were investigated. No local failure was seen with 10 Gy or more in patients with group 1 tumor. In patients with group 2 tumor, local failure of primary tumor was seen in two of three patients who received less than 40 Gy and in none of three patients who received 40 Gy or more. In patients with group 3 tumor, local failure of primary tumor was seen in three of 10 patients who received less than 54 Gy and in none of three patients who received 54 Gy or more. A relapse in the spinal canal was observed in three patients. All three patients had group 3 tumors with an element of highly malignant GCT, i.e. EC, YST and EC/IMT, respectively. The incidence of spinal metastasis in group 3 patients was 3/13 (23%). The cerebrospinal fluid cytology was examined in two of the three patients prior to the initial treatment, but tumor cells were not detected. One of the spinal metastases occurred together with a local failure. The three spinal relapses were seen among eight patients who had not received craniospinal irradiation. All five patients who had received craniospinal irradiation did not experience spinal relapse. Intracranial relapse outside the irradiated field without evidence of spinal relapse was observed in one patient in group 1. This was the same patient who experienced a relapse of germinoma along the third ventricular wall 8 years after total resection of MT. No relapse outside the irradiated field was observed in patients with group 2 tumors. Abdominal relapse via a ventriculoperitoneal shunt was observed after 114 months in one patient with EC who had received craniospinal irradiation.

4. Discussion One of the shortcomings of our study is the lack of uniformity in the treatment protocol because of the multi-insti-

Table 1 Characteristics of the initial manifestation of relapse according to the pathological diagnosis Histology

Relapsed sites In fieldd

a

Group 1 Group 2b Group 3c Total a

0 2 2 5

When relapsed? Out fielde h

1 0 2 2

In and outf

Out CNSg

Not irradiated

,5 years

>5 years

0 0 1 1

0 0 1 1

0 0 2 2

0 2 7 9

1 0 1 2

No. of patients (relapse/total) 1/5 2/6 8/13 11/24

Mature teratoma with or without germinoma. Immature teratoma with or without germinoma. c Teratoma with malignant transformation, or tumor consisting of components of embryonal carcinoma, yolk sac tumor and/or choriocarcinoma. d In the irradiated field. e Out of the irradiated field. f Both in and out of the irradiated field. g Abdominal relapse via ventriculoperitoneal shunt. h The histopathology of this relapse was pure germinoma. b

58

H. Aoyama et al. / Radiotherapy and Oncology 49 (1998) 55–59

tutional retrospective nature of the study. Paradoxical results of the negative impact of chemotherapy must reflect the usage of chemotherapy for aggressive tumors. Platinumbased multi-agenic chemotherapy has dramatically improved the outcome for patients with systemic NGGCT [14]. Many case reports and recent prospective studies have suggested that chemotherapy similar to that used for testicular NGGCT may play an important role in treating similar tumors in the CNS [1–5,7,8,14,15]. Balmaceda et al. [3] have shown that four cycles of carboplatin, VP-16 and bleomycin can produce a complete remission rate of 78% and a 2-year survival rate of 62%, with a median follow-up time of 33 months for NGGCT. However, 13 out of 26 NGGCT patients experienced progression or relapse of disease in their series and most of them received radiotherapy as salvage therapy [3]. Allen et al. [2] have shown that neoadjuvant platinum-based chemotherapy and CNS irradiation are effective in achieving objective tumor responses and in producing long-term continuous remission-free survival. In searching for a curable therapeutic mode, the management plan should be directed toward improving the primary response rate for the highly malignant GCTs. The late effects of therapy can be at least partially set aside. The role of radiotherapy in combination with intensive chemotherapy should be investigated more to improve the poor survival rate. Patients with MT had an exceptionally good prognosis after surgical resection with or without radiotherapy. The importance of radiotherapy for MT is questionable because of the lack of a dose–response relationship for local control. Because the mature teratoma has been thought not to be radiosensitive in nature, total removal is recommended in order to eliminate the possibility of a local relapse. If other GCT components are mixed with mature teratoma, radiotherapy and/or chemotherapy should be administered according to the most malignant component of the tumor. A late relapse of the germinoma component in one patient with MT in this study illuminated the difficulty of pathological diagnosis of pure NGGCT. Careful follow-up is mandatory even for patients with mature teratoma. Our study was consistent with previous reports suggesting an intermediate prognosis for those patients with IMT [11,16]. There was no spinal relapse in patients with IMT, irrespective of the fact that five of the six patients did not receive craniospinal irradiation. Our results suggest that immature teratoma rarely disseminates, but rather relapses near the primary tumor in the irradiated area, even after gross total resection. Craniospinal irradiation is not recommended for patients with IMT. The dose to the primary site should be 40 Gy or more based on the present study. A generous local field or partial brain field with 40 Gy in 20 fractions may be a better choice of treatment. A relapse in the spinal canal was observed in 37.5% (3/8) of patients with group 3 tumors who were not treated with craniospinal irradiation and in none of five patients who received craniospinal irradiation in the present study. Mat-

sutani et al. [11] observed a 21.7% (5/23) rate of spinal relapse in patients with highly malignant NGGCTs treated primarily by whole-brain irradiation without spinal irradiation. Schild et al. [16] have shown that patients who had a component of teratoma with malignant transformation had a significantly increased risk of spinal failure as compared with patients with other NCCGT. They showed, however, that none of five patients with teratoma with malignant transformation who received whole-brain or craniospinal irradiation experienced spinal failure. Both patients with teratoma with malignant transformation in our series received craniospinal irradiation and did not experience a relapse, making our results consistent with those of Schild et al. [16], which suggests that a large volume of irradiation for this pathology may prevent spinal relapse. These results suggest that craniospinal irradiation is associated with the prevention of spinal relapse for group 3 tumors, although the number of patients in the present study was too small to find a statistically significant difference. This speculation is in accordance with other reports [4,10,11,16]. Kirkove et al. [10] reported the case of a patient with pineal yolk sac tumor who was treated with surgery, platinum-containing chemotherapy and craniospinal irradiation (25.5 Gy in 17 fractions over 25 days) and was alive without relapse for 2 years. Calaminus et al. [4] used 35–36 Gy craniospinal irradiation followed by a 20 Gy tumor boost after platinum-containing chemotherapy and surgical resection for highly malignant NGGCTs. They reported that 12 of 14 patients were alive and disease-free with a median follow-up of 52 months. Based on these findings, we recommended the use of craniospinal irradiation for patients with group 3 tumors. A reasonable dose of craniospinal irradiation may be between 26 and 40 Gy [4]. The dose to the primary site of group 3 tumors should be 54 Gy or more according to the present study. In conclusion, differences in the prognosis and relapse pattern among histopathologic groups were identified. Highly malignant GCTs show a high incidence of spinal metastasis and craniospinal irradiation may reduce the risk of spinal metastasis. Appropriate radiotherapy in combination with surgery and chemotherapy should be investigated further to improve the outcome of this rare disease in young patients.

References [1] Allen, J.C. Controversies in the management of intracranial germ cell tumors. Neurol. Clin. 9: 441–452, 1991. [2] Allen, J.C., Kim, J.H. and Packer, R.J. Neoadjuvant chemotherapy for newly diagnosed germ cell tumors of the central nervous system. J. Neurosurg 67: 65–70, 1987. [3] Balmaceda, C., Heller, G., Rosenblum, M., et al. Chemotherapy without irradiation – a novel approach for newly diagnosed CNS germ cell tumors: results of an international cooperative trial. J. Clin. Oncol. 14: 2908–2915, 1996. [4] Calaminus, G., Bamberg, M., Baranzelli, M.C., et al. Intracranial

H. Aoyama et al. / Radiotherapy and Oncology 49 (1998) 55–59

[5]

[6]

[7] [8]

[9]

[10]

germ cell tumors: a comprehensive update of the European data. Neuropediatrics 25: 26–32, 1994. Deanaley, D.P., A’Hern, R.P., Whittaker, S. and Bloom, H.J.G. Pineal and CNS germ cell tumors: Royal Marsden Hospital experience 1962–1987. Int. J. Radiat. Oncol. Biol. Phys. 18: 773–781, 1990. Edwards, M.S.B., Hudgins, R.J., Wilson, C.B., Levin, V.A. and Wara, W.M. Pineal region tumors in children. J. Neurosurg. 68: 689–697, 1988. Einhorn, L.H. General Motors Cancer Research Prize-Winners Laureates Lectures. Cancer 71: 3182–3184, 1993. Itoyama, Y., Kochi, M., Yamashiro, S., Yoshizato, K., Kuratsu, J. and Ushio, Y. Combination chemotherapy with cisplatin and etoposide for hematogenous spinal metastasis of intracranial germinoma – case report. Neurol. Medico-Chirurgica 33: 28–31, 1993. Kersh, C.R., Constable, W.C., Eisert, D.R., et al. Primary central nervous system germ cell tumors. Effect of histologic confirmation on radiotherapy. Cancer 61: 2148–2152, 1988. Kirkove, C.S., Brown, A.P. and Syman, L. Successful treatment of a pineal endodermal sinus tumor. Case report. J. Neurosurg. 74: 832– 836, 1991.

59

[11] Matsutani, M., Sano, K., Takakura, K., et al. Primary intracranial germ cell tumors: a clinical analysis of 153 histologically verified cases. J. Neurosurg. 86: 446–455, 1997. [12] Ono, N., Isobe, I., Uki, J., Kurihara, H., Shimizu, T. and Kohno, K. Recurrence of primary intracranial germinomas after complete response with radiotherapy: recurrence patterns and therapy. Neurosurgery 35: 615–620; discussion 620–621, 1994. [13] Sano, K. So-called intracranial germ cell tumours: are they really of germcell origin? Br. J. Neurosurg. 9: 391–401, 1995. [14] Sakai, N., Yamada, H., Andoh, T., Nishimura, Y. and Niikawa, S. Long-term survival in malignant intracranial germ-cell tumors: a report of two cases and a review of the literature. Childs Nervous System 9: 431–436, 1993. [15] Sawamura, Y. Current diagnosis and treatment of central nervous system germ cell tumours. Curr. Opin. Neurol. 9: 419–423, 1996. [16] Schild, S.E., Haddock, M.G., Scheithaauer, B.W., et al. Nongerminomatous germ cell tumors of the brain. Int. J. Radiat. Oncol. Biol. Phys. 36: 557–563, 1996.