Is High-Dose Chemotherapy After Primary Chemotherapy a Therapeutic Option for Patients With Primary Mediastinal Nonseminomatous Germ Cell Tumor?

Transcription

1 Biology of Blood and Marrow Transplantation 12: (2006) 2006 American Society for Blood and Marrow Transplantation /06/ $32.00/0 doi: /j.bbmt Is High-Dose Chemotherapy After Primary Chemotherapy a Therapeutic Option for Patients With Primary Mediastinal Nonseminomatous Germ Cell Tumor? Giuseppe Luigi Banna, 1 Ugo De Giorgi, 3 Benvenuto Ferrari, 1 Luca Castagna, 1 Marco Alloisio, 2 Maurizio Marangolo, 3 Giovanni Rosti, 3 Armando Santoro 1 1 Department of Medical Oncology and Hematology and 2 Unit of Thoracic Surgery, Istituto Clinico Humanitas, Rozzano, Milan, Italy; 3 Department of Oncology and Hematology, Istituto Oncologico Romagnolo, Ravenna, Italy Correspondence and reprint requests: Giuseppe L. Banna, MD, Department of Medical Oncology and Hematology, Istituto Clinico Humanitas, Via Manzoni, 56, 20089, Rozzano, Milan, Italy ( gbanna@yahoo.com). Received April 19, 2006; accepted June 8, 2006 ABSTRACT Patients with primary mediastinal nonseminomatous germ cell tumors have a poor prognosis, with a 5-year overall survival of nearly 50%. We investigated the feasibility and activity of early high-dose chemotherapy (HDCT) in these patients. After conventional induction chemotherapy, patients underwent a single shot of HDCT consisting of carboplatin, etoposide, and cyclophosphamide, followed by peripheral blood progenitor cell support. Twenty-one patients were considered for treatment with HDCT. Median age was 29 years (range, years). Eight (38%) patients had lung metastases. After primary chemotherapy, 7 patients achieved complete remission, 4 achieved partial remission with negative marker, 1 achieved partial remission with positive marker, 2 had stable disease, and 7 progressive disease. Twelve patients were not treated with HDCT due to progressive disease and poor physical conditions. No HDCT-related deaths or irreversible organ toxicities were observed. Residual surgery after HDCT was performed in 4 patients and resulted in 3 pathologic complete remissions. With a median follow-up of 52 months (range, months) in 9 patients treated with HDCT, 8 have been continuously free of disease. Of 12 patients who did not receive HDCT, 0 was alive at 2 years from diagnosis. A single course of HDCT after induction chemotherapy appeared to be inapplicable in most of our patients, mainly due to early progressive disease. These data should be considered in the analysis of retrospective series and in the design of new prospective trials with HDCT in these patients. Earlier HDCT administration followed by residual surgery should be considered for further investigation American Society for Blood and Marrow Transplantation KEY WORDS Primary mediastinal nonseminomatous germ cell tumor High-dose chemotherapy Poor prognosis Extragonadal germ cell tumor Residual surgery INTRODUCTION Although patients with advanced testicular germ cell malignancies present a high cure rate, patients with primary mediastinal nonseminomatous germ cell tumor (PMNSGCT) have a poor prognosis [1,2]. The 5-year overall survival (OS) rate is nearly 50%, and the 5-year disease-free survival rate approximately 40% for these patients after conventional-dose cisplatin and etoposide-based chemotherapy and surgery, if necessary [1,2]. The presence of lung metastases represents another unfavorable prognostic factor [3,4]. High-dose chemotherapy (HDCT) with peripheral blood progenitor cell (PBPC) support has been investigated, with controversial results in these patients [5-9]. Recently, multicycle upfront HDCT has shown interesting results in these patients [10]. We investigated the feasibility and activity of an early single HDCT course with carboplatin, etoposide, and cyclophosphamide (the CARBOPEC regi- 1085

2 1086 G. L. Banna et al. men) with PBPC support in patients who did not progress to standard-dose cisplatin and etoposidebased chemotherapy. METHODS Patients All patients with PMNSGCT who were followed at the Istituto Clinico Humanitas (Milan, Italy) and the Istituto Oncologico Romagnolo (Ravenna, Italy) between 1993 and 2004 were prospectively enrolled. Written informed consent was obtained. No prior treatments were allowed. In the absence of histologic diagnosis (due to the need for urgent chemotherapy treatment or difficulty in obtaining an adequate biopsy), high serum levels of -human chorionic gonadotropin ( -hcg) and/or -fetoprotein (AFP) and/or lactate dehydrogenase (LDH) with a clinical picture and staging compatible with PMNSGCT were accepted. Evaluation before chemotherapy included a detailed history, full physical examination, chest radiograph, computed tomography of the thorax, abdomen, and pelvis, standard hematologic and biochemical parameters, and measurement of serum tumor markers (AFP, -hcg, and LDH). Computed tomography or magnetic resonance imaging of the brain and radionuclide bone scan were performed in symptomatic patients. Treatment Induction chemotherapy consisted of cisplatin, etoposide, and bleomycin (PEB) or ifosfamide (VIP). Chemotherapy was given if neutrophil and platelet counts were 500/ L and / L, respectively, at day 1 of each course. Prophylactic administration of granulocyte colony-stimulating factor (G- CSF) was used for all subsequent chemotherapy courses in case of neutropenic fever during 1 preceding course. Responding patients underwent PBPC mobilization preferentially with cyclophosphamide 4.5 g/m 2 plus G-CSF 5 g/kg per day or G-CSF 10 g/kg per day alone. After chemotherapy, G-CSF was started 1-2 days after cessation of chemotherapy and was continued until the day of PBPC harvesting. PBPCs were harvested with a Cobe Spectra (Denver, CO, USA) 3000 device, and the targeted dose was fixed at CD34 cells/kg. All patients undergoing HDCT were required to have an Eastern Cooperative Oncology Group performance status of 0-1, no prohibitive medical comorbidities, and no progressive disease after induction chemotherapy. HDCT consisted of carboplatin at a dose of 250, 400, or 550 mg/m 2 intravenously on days 6 to 3, if creatinine clearance was 30-59, , or 100 ml/min, respectively; etoposide 450 mg/m 2 intravenously on days 6 to 3; and cyclophosphamide 1600 mg/m 2 intravenously on days 6 to 3, with mesna 1600 mg/m 2 intravenously over 24 hours of continuous infusion on days 7 to 1 (CARB- OPEC regimen). On day 0, PBPCs were reinfused followed by G-CSF 5 g/kg per day from day 1 until hematologic recovery. All patients with residual radiographic disease after completion of induction chemotherapy were evaluated for complete residual surgery before undergoing HDCT. Patients who were deemed not operable by the surgeons underwent HDCT and were subsequently reassessed for surgery in case of disease persistence. Response Evaluation After induction chemotherapy and after HDCT, tumor response was evaluated by computed tomography of the chest, abdomen, and pelvis and by measurement of serum tumor markers. After treatment was completed, patients were followed up every 3 months in the first year, every 4 months in the second and third years, every 6 months up to 5 years, and annually thereafter. Tumor response was classified as follows. Clinical complete remission (ccr) was defined as complete disappearance of all clinical, radiologic, and biochemical evidence of disease, with normalization of levels of tumor markers ( -hcg and/or AFP and/or LDH) for 1 month. A partial response (PR) was defined as a decrease 50% of the sum of the products of perpendicular diameters of measurable disease that lasted 1 month. In addition, patients with PR and normalization of previously elevated tumor marker levels were classified as tumor-markernegative partial responders (PRm ) and those with 90% decrease in markers but without normalization as tumor-marker-positive partial responders (PRm ). Stable disease (SD) was defined as a decrease 50% or an increase 25% in bidimensional tumor measurements or stable tumor marker levels. After complete residual surgery, patients with normal tumor marker levels were defined as having pathologic complete remission (pcr) if necrosis/fibrosis and/or mature teratoma was histologically proved, or surgical complete remission (scr) when viable malignant tumor cells were documented. Disease progression (PD) was defined as residual lesions increasing in size or as the occurrence of new tumor lesions and/or serial elevation of tumor markers at repeated controls. Treatment-related toxicity was assessed with NCI- CTC 2.0 (National Cancer Institute, Bethesda, Md). Statistical Analysis The study was planned as a 2-stage design according to the design of Simon [11]. Alpha and beta errors were set at 10%. A response rate 15% was consid-

3 High-Dose Chemotherapy for Mediastinal Germ Cell Tumors 1087 ered to be of interest, and 19 patients were entered in the first stage. If no responses were noted in the first stage, that arm would be terminated. Otherwise, if 1 response was noted, an additional 18 patients needed to be accrued for a total of 37 patients. Data were collected retrospectively into a database and are presented as number and proportion or as median and range, where appropriate. Progression-free survival (PFS) and OS curves were generated with Kaplan- Meier statistical analysis [12]. PFS time was calculated from enrollment to progression. OS time was calculated from enrollment to death or last contact. All analyses were made with Statistica 6.0 ( RESULTS Patients and Treatment Between March 1993 and September 2004, 21 consecutive patients with PMNSGCT were enrolled. The trial was closed and the analyses performed after the first step because of slow recruitment. Median age was 29 years (range, years). Eight patients (38%) had lung metastases. Clinical features are summarized in Table 1. Surgery at diagnosis was performed in 6 patients but was radical in only 2 cases. All but 2 patients received 4 courses of PEB chemotherapy; bleomycin was substituted for ifosfamide in 2 patients according to the VIP regimen. Two patients received an additional PEB course and 1 patient an additional VIP course, instead of the standard 4 courses, with the purpose of further decreasing residual disease; 1 patient received only 1 PEB course because of PD. Residual surgery after induction chemotherapy was performed in 6 patients. PBPCs were mobilized with cyclophosphamide plus G-CSF (n 4), G-CSF alone (n 2), etoposide plus G-CSF (n 1), Table 1. Patient Characteristics at Diagnosis Characteristic n % Patients 21 Age, years Median (range) 29 (19-55) Metastatic sites Lung 8 38 Central nervous system 1 5 Other (pleura, pericardium) 2 10 Serum tumor markers Raised AFP (>7 IU/mL) 14* 70 Median (range) 2618 ( ) Raised -hcg (>5 IU/L) Median (range) 94 ( ) Raised LDH (>243 IU/L) 9 60 Median (range) 1264 ( ) *One patient not assessable. Three patients not assessable. Six patients not assessable. cisplatin and etoposide plus G-CSF (n 1), and paclitaxel, ifosfamide, and cisplatin plus G-CSF (n 1). Mobilization with G-CSF alone was repeated in 2 patients because of infectious contamination of the first collection in 1 patient and poor mobilization after cyclophosphamide in the other. One patient received HDCT with carboplatin alone, at the dose corresponding to Area Under Plasma Concentration Curve on days 6 to 3, because he had received an additional fifth PEB course as induction chemotherapy and etoposide 2.5 g/m 2 as mobilization. A median dose of CD34 cells/kg (range, ) was administered. Treatment regimens are listed in Table 2. Response and Survival After induction and residual surgery, 4 patients achieved ccr (19%), 1 achieved pcr (5%), 2 achieved scr (10%), 4 were PRm (19%), 1 was PRm (5%); 2 had SD (10%) and 7 (33%) showed PD. Twelve patients (57%) did not undergo HDCT: 7 patients due to PD after induction regimen, 2 due to tumultuous PD with poor physical condition during mobilization after initial CR to induction chemotherapy, 1 with SD after induction regimen but with poor performance status, 1 due to poor mobilization, and 1 because of an unwillingness to undergo HDCT. Nine patients (43%) underwent HDCT. At the time of HDCT, 2 patients were in ccr, 1 was in pcr, 1 was in scr, 3 were PRm, 1 was PRm, and 1 had SD. After HDCT, 7 of 9 (78%) assessable patients were in CR (including 3 in pcr after subsequent residual thoracic surgery performed in 4 patients); 1 patient was in PRm (but a viable tumor was disclosed by subsequent residual thoracic surgery); and 1 patient had PD. Among the 7 complete responders to HDCT, 4 underwent transplantation in CR, 2 in PRm, and 1 in SD; 3 of them (43%) had lung metastases. Different salvage chemotherapy regimens were administered after induction or HDCT failure; these included etoposide, ifosfamide, vinblastine, paclitaxel, gemcitabine, and doxorubicin but in all cases resulted in PD and subsequent progression-related deaths. OS and PFS curves are shown in Figures 1 and 2, respectively. OS and PFS at 3 years were 41% (SE, 11.1%) and 43% (SE, 10.8%), respectively. Median OS time was 19 months and median PFS was 10 months. With a median follow-up of 52 months (range, months), 8 patients are alive (38%), 7 with durable CR and 1 with PRm. The remaining 13 patients (62%) died from PD. All but 1 of 9 patients (89%) who underwent HDCT are alive and in disease remission, with a median follow-up of 52 months (range, months). Of 5 patients who did not

4 1088 G. L. Banna et al. Table 2. Results of Induction Treatment and High-Dose Chemotherapy Characteristic n % Surgery at diagnosis 6 29 Chemotherapy PEB 19 VIP 2 Response after induction/residual surgery* ccr 4 19 pcr 1 5 scr 2 10 PRm 4 19 PRm 1 5 SD 2 10 PD 7 33 CD34 mobilization Cyclophosphamide 4.5 g/m 2 4 G-CSF 10 g/kg per day alone 2 Etoposide/PE/TIP 1/1/1 HDCT CARBOPEC 8 43 CarboAUC24 1 CD34 cell dose, median (range) kg ( ) Disease status before HDCT ccr 2 22 pcr 1 11 scr 1 11 PRm 3 33 PRm 1 11 SD 1 11 Disease status after HDCT/residual surgery ccr 4 44 pcr 3 33 PRm 1 11 PD 1 11 Follow-up (mo) Median (range) 52 (15-71) Alive and disease-free 8 38 Died PEB indicates cisplatin, etoposide, bleomycin; VIP etoposide, ifosfamide, cisplatin; ccr clinical complete remission; pcr pathologic CR; scr surgical CR; PRm partial remission with negative tumor markers; PRm partial remission with positive tumor markers; SD stable disease; PD progressive disease; G-CSF granulocyte colony-stimulating factor; PE cisplatin, etoposide; TIP paclitaxel, ifosfamide, cisplatin; CARBOPEC high doses of carboplatin, etoposide, cyclophosphamide; CarboAUC24 carboplatin at the dose corresponding to AUC 24; HDCT high-dose chemotherapy. *Residual surgery after induction chemotherapy was performed in 6 patients. Includes 1 poor mobilizer and 1 patient with contaminated yield after cyclophosphamide. Responses after HDCT and residual surgery performed in 4 patients. Patient underwent residual thoracic surgery with evidence of viable tumor. receive HDCT despite no evidence of PD after induction chemotherapy (including 2 with ccr after induction chemotherapy but with PD during mobilization), 4 died of disease. HDCT Toxicity No toxic deaths occurred. Median durations of G4 neutropenia and thrombocytopenia were 9 days (range, 2-12 days) and 6 days (range, 4-9 days), respectively. All but 1 patient developed neutropenic fever that resolved in all cases with antibiotics. Seven patients developed grade 3-4 mucositis, which required opiate analgesics and parenteral nutrition. Most patients had nausea, vomiting, and diarrhea, which were controlled with medication. No other grade 3-4 toxicities and/or major complications were observed. No cases of secondary acute leukemia and/or other neoplasms have been observed. DISCUSSION HDCT based on carboplatin and etoposide is actively investigated in patients with germ cell tumor and poor prognostic features [4,5]. In this study, we evaluated the feasibility and activity of a single course of early HDCT with CARBOPEC after induction chemotherapy for patients with PMNSGCT. Of 21 patients who started induction chemotherapy, only 9 (43%) received HDCT. Twelve patients were not treated with HDCT mainly due to PD and poor physical condition. A harvesting failure rate of 25%-30% was previously reported in patients with heavily pretreated germ cell tumor [13]. In our study, patients with PMNSGCT were not heavily pretreated, but patients did not receive HDCT due to early PD during/after induction chemotherapy. The role of HDCT in patients with PMNSGCT was very difficult interpret in our study, mainly due to patient selection. Moreover, HDCT in patients who do not achieve a CR or PRm status after 3-4 courses of an induction chemotherapy could be considered as a second-line treatment, and very poor results have been reported with HDCT in this setting for patients with PMNSGCT [9]. We reported 3-year OS and PFS of 41% and 43%, respectively, which did not appear to differ from those reported by other investigators who used conventional treatments [1,2,4]. However, all but 1 of 9 patients (89%) who underwent HDCT are alive and in disease remission with a median follow-up of 52 months (range, months), whereas 4 of 5 patients (80%) who had disease response after induction treatment but did not undergo HDCT relapsed and died from PD. Thirteen patients (62%) progressed and died despite different salvage chemotherapy regimens administered. In our series, 6 patients underwent early residual surgery, performed after standard-dose chemotherapy, thus possibly delaying the administration of HDCT. Instead, 4 patients underwent residual surgery after HDCT. Of note, 5 of 6 patients who un-

5 High-Dose Chemotherapy for Mediastinal Germ Cell Tumors 1089 Figure 1. Probability of overall survival. derwent early residual surgery had histologic evidence of residual disease, whereas 3 of 4 patients who underwent residual surgery after HDCT were in pcr. Residual surgery has an important role in the treatment of PMNSGCT, although no optimal timing has been established in HDCT programs [14]. Our findings may support the role of dose intensity in the treatment of PMNSGCT for the maintenance and induction (3 of 9 patients who underwent HDCT were not in CR or PRm ) of disease response; earlier HDCT administration and performing residual surgery after earlier HDCT may be suggested and considered for further evaluation in prospective studies. Comparison of our study with other similar trials is limited by few reports on HDCT for these patients. Moreover, most investigators included few cases of PMNSGCT in larger groups of patients with poor prognosis [1]. Decatris et al [7] reported 4-year PFS and OS of 42%, and 25%, respectively, for patients with PMNSGCT who were treated with HDCT after 3-4 courses of PEB. A recent analysis of the Euro- Figure 2. Probability of progression-free survival.

6 1090 G. L. Banna et al. pean Bone Marrow Transplantation Solid Tumors Working Party [8] showed a disease-free survival of 64% in the 11 patients with PMNSGCT who received early HDCT. However, in these trials, data referred only to patients with responsive disease who had undergone HDCT, but no data were available about patients who started the induction regimen but did not receive HDCT [7-9]. The only available phase III trial in patients with poor prognosis, which was presented only as an extended abstract, showed no advantage for the HDCT arm, but lacked data on patients with PMNSGCT [15]. Bokemeyer et al [6] used a matched-pair analysis among 147 patients with poor prognosis who were treated with a single course of mobilizing chemotherapy and 3-4 courses of high-dose cisplatin, etoposide, and ifosfamide and 309 patients who were treated with standard-dose PEB or VIP chemotherapy and reported a significant improvement in PFS (75% versus 59%, P ) and OS (82% versus 71%, P ). However, no data on patients with PMNSGCT, which represented 10% of all cases, were available. Recently, multicycle upfront HDCT has shown interesting results in patients with poor prognosis and 2-year PFS and disease-specific survival rates of 69% and 79% [16]. In a subgroup with PMNSGCT, these results were substantially similar, with 2-year PFS and OS rates of 64 and 68%, respectively [10]. This strategy was explored in a phase I-II study in 221 patients with Indiana advanced disease or International Germ Cell Cancer Collaborative Group poor prognosis criteria and consisted of the administration of 3-4 sequential cycles of high-dose VIP (cisplatin, etoposide, and ifosfamide) at 6 consecutive dose levels after 1 cycle of standard induction/mobilization VIP [16]. The results observed in the subgroup with PMNS- GCT might indicate an approximately 15% survival improvement compared with data of an international database analysis including 253 patients with PMNS- GCT treated with conventional chemotherapy [10]. Other possible strategies to improve outcome of patients with PMNSGCT consist of intensive chemotherapeutic regimens such as intensive induction and sequential, alternating, or dose-dense chemotherapy, and the addition of new active drugs (eg, paclitaxel) [17-19]. Recently, Christian et al [20] reported the results of an intensive induction chemotherapy with bleomycin, vincristine, and cisplatin alternating with PEB (CBOP/PEB regimen) with a 3-year OS of 77% for patients with PMNSGCT. However, nearly 50% of patients with PMNSGCT did not receive a complete course of CBOP/BEP and toxicities were considerable. In conclusion, a single course of HDCT after primary chemotherapy appeared to be inapplicable in most of our patients with PMNSGCT, mainly due to early PD. Patients with PMNSGCT have a risk of 30%-35% of early PD during or after induction chemotherapy [3-5]. This issue should be considered in the analysis of retrospective series and in the design of clinical trials with HDCT in PMNSGCT. Earlier HDCT administration followed by residual surgery should be considered for further evaluation in prospective studies. Recently, multicycle upfront HDCT has shown interesting results in these patients. This strategy is currently being investigated in 2 ongoing, phase III, randomized studies in Europe and the United States. Results of these trials may better define the role of HDCT in these patients. ACKNOWLEDGMENTS The authors thank Dr Emanuela Morenghi for help in the statistical review of data. REFERENCES 1. The International Germ-Cell Cancer Collaborative Group. International Germ Cell Consensus Classification: a prognostic factor-based staging system for metastatic germ cell cancers. International Germ Cell Cancer Collaborative Group. J Clin Oncol. 1997;15: Fizazi K, Culine S, Droz JP, et al. Primary mediastinal nonseminomatous germ cell tumors: results of modern therapy including cisplatin-based chemotherapy. J Clin Oncol. 1998;16: Schmoll HJ, Souchon R, Krege S, et al. European consensus on diagnosis and treatment of germ cell cancer: a report of the European Germ Cell Cancer Consensus Group (EGCCCG). Ann Oncol. 2004;15: Bokemeyer C, Nichols CR, Droz JP, et al. Extragonadal germ cell tumors of the mediastinum and retroperitoneum: results from an international analysis. J Clin Oncol. 2002;20: De Giorgi U, Rosti G, Papiani G, et al. The status of high-dose chemotherapy with hematopoietic stem cell transplantation in patients with germ cell tumor. Haematologica. 2002;87: Bokemeyer C, Kollmannsberger C, Meisner C, et al. First-line high-dose chemotherapy compared with standard-dose PEB/ VIP chemotherapy in patients with advanced germ cell tumors: a multivariate and matched-pair analysis. J Clin Oncol. 1999;17: Decatris MP, Wilkinson PM, Welch RS, et al. High-dose chemotherapy and autologous haematopoietic support in poor risk non-seminomatous germ-cell tumours: an effective firstline therapy with minimal toxicity. Ann Oncol. 2000;11: Rosti G, De Giorgi U, Wandt H, et al. First line high-dose chemotherapy for patients with poor prognosis extragonadal germ cell tumors: the experience of the European Bone Marrow Transplantation (EBMT) Solid Tumors Working Party. Bone Marrow Transplant. 2004;34: De Giorgi U, Demirer T, Wandt H, et al. Second-line highdose chemotherapy in patients with mediastinal and retroperitoneal primary non-seminomatous germ cell tumors: the EBMT experience. Ann Oncol. 2005;16: Bokemeyer C, Schleucher N, Metzner B, et al. First-line se-

7 High-Dose Chemotherapy for Mediastinal Germ Cell Tumors 1091 quential high-dose VIP chemotherapy with autologous transplantation for patients with primary mediastinal nonseminomatous germ cell tumours: a prospective trial. Br J Cancer. 2003;89: Simon R. Optimal two-stage designs for phase II clinical trials. Control Clin Trials. 1989;10: Kaplan EL, Meier P. Non parametric estimation from incomplete observations. J Am Stat Assoc. 1958;53: Dazzi C, Cariello A, Rosti G, et al. Peripheral blood progenitor cell (PBPC) mobilization in heavily pretreated patients with germ cell tumors: a report of 34 cases. Bone Marrow Transplant. 1999;23: Date H, Kiura K, Ueoka H, et al. Induction chemotherapy, surgical resection, and high-dose chemotherapy for mediastinal nonseminomatous germ-cell tumor. J Thorac Cardiovasc Surg. 2005;130: Chevreau C, Droz JP, Pico JL, et al. Early intensified chemotherapy with autologous bone marrow transplantation in first line treatment of poor risk non-seminomatous germ cell tumours. Preliminary results of a French randomised trial. Eur Urol. 1993;23: Schmoll HJ, Kollmannsberger C, Metzner B, et al. Long-term results of first-line sequential high-dose etoposide, ifosfamide, and cisplatin chemotherapy plus autologous stem cell support for patients with advanced metastatic germ cell cancer: an extended phase I/II study of the German Testicular Cancer Study Group. J Clin Oncol. 2003;21: Kaye SB, Mead GM, Fossa S, et al. Intensive induction-sequential chemotherapy with BOP/VIP-B compared with treatment with BEP/EP for poor-prognosis metastatic nonseminomatous germ cell tumor: a randomized Medical Research Council/European Organization for Research and Treatment of Cancer study. J Clin Oncol. 1998;16: Bower M, Newlands ES, Holden L, et al. Treatment of men with metastatic non-seminomatous germ cell tumours with cyclical POMB/ACE chemotherapy. Ann Oncol. 1997;8: Margolin KA, Doroshow JH, Frankel P, et al. Paclitaxel-based high-dose chemotherapy with autologous stem cell rescue for relapsed germ cell cancer. Biol Blood Marrow Transplant. 2005; 11: Christian JA, Huddart RA, Norman A, et al. Intensive induction chemotherapy with CBOP/BEP in patients with poor prognosis germ cell tumors. J Clin Oncol. 2003;21: