A population-based study of treatment guided by tumor marker decline in. 603 patients with metastatic non-seminomatous germ cell tumor - A report

Transcription

1 A population-based study of treatment guided by tumor marker decline in 603 patients with metastatic non-seminomatous germ cell tumor - A report from the Swedish-Norwegian Testicular Cancer Group (SWENOTECA) Sven-Erik Olofsson 1, Torgrim Tandstad 2, Mats Jerkeman 1, Olav Dahl 3, Olof Ståhl 1, Olbjørn Klepp 2, Roy M. Bremnes 4, Gabriella Cohn-Cedermark 5, Carl W. Langberg 6, Anna Laurell 7, Arne Solberg 2, Ulrika Stierner 8, Rolf Wahlqvist 9, Hans Wijkström 10, Eva Cavallin-Ståhl 1 1 Department of Oncology and Department of Cancer Epidemiology, Lund University Hospital, Lund University, Lund, Sweden, 2 Department of Oncology, St. Olav University Hospital, Trondheim, Norway, 3 Section of Oncology, Institute of Medicine, University of Bergen and Department of Oncology, Haukeland University Hospital, Bergen, Norway, 4 Department of Oncology, Institute of Clinical Medicine University of Tromsø and University Hospital of North Norway, Tromsø, Norway, 5 Department of Oncology-Pathology, Karolinska Institute and University Hospital, Stockholm, Sweden, 6 Cancer Centre, Ullevål University Hospital, Oslo, Norway, 7 Department of Oncology, Uppsala University Hospital, Uppsala, Sweden, 8 Department of Oncology, Sahlgrenska University Hospital, Göteborg, Sweden, 9 Department of Urology, Oslo University Hospital, Oslo, Norway, 10 Department of Urology, Karolinska University Hospital, Huddinge, Stockholm, Sweden Acknowledgement: The Swedish Cancer Society, Gunnar Nilsson Foundation for Cancer Research, Nordic Cancer Union Department of Oncology, Lund University Hospital, Lund, Sweden sven-erik.olofsson@ltkronoberg.se, Telephone: , Fax:

2 Running head: Treatment based on tumormarker decline in metastatic NSGCT Data presented in part at ASCO Orlando 2009 Disclaimers: None 2

3 Abstract Purpose: From 1995 to 2003, 603 adult patients from Sweden and Norway with metastatic testicular non-seminomatous germ cell tumor, NSGCT, were included prospectively in a population-based protocol with strict guidelines for staging, treatment and follow up. The basic strategy was to individualize treatment according to initial tumor marker response. Methods: Initial treatment for all patients was two courses of standard BEP, with tumor markers analyzed weekly. Satisfactory response was defined as a half-life (t1/2) for AFP, 7 days and/or for β-hcg, 3 days. Patients with prolonged marker t1/2 received intensification with addition of ifosfamide (BEP-if/PEI) in step 1. If still unsatisfactory response, the treatment was intensified with high-dose chemotherapy with stem cell rescue, step 2. Results: 99% (603 of 610) of all patients with metastatic testicular NSGCT in the population were included in the protocol. Median follow up was 8.2 years. 77% of the patients were treated with BEP alone, median 4 courses, 18% received intensification step 1 and 5% step 2. Grouped according to International Germ Cell Consensus Classification, 10-year overall survival was 94.7% in good prognosis, 90.1% in intermediate prognosis and 67.4 in poor prognosis patients, respectively. Conclusion: With detailed treatment protocols and a dedicated collaborative group of specialists, treatment results comparable to those reported from large single institutions can be achieved at a national level. With the treatment principles used in SWENOTECA IV, the survival of intermediate prognosis patients is remarkable and close to that of good prognosis patients. 3

4 Introduction Testicular cancer is the most common cancer among men between years of age with approximately 500 new cases diagnosed every year in Sweden and Norway. Since the introduction of cisplatin in the treatment of metastatic testicular non-seminomatous germ cell cancer (NSGCT), the cure rate has been high. Still, however, a notable fraction of patients with metastases die of this disease. For decades, several prognostic classifications have been proposed, but in 1997 these were replaced by the International Germ Cell Consensus Classification (IGCCC) 1 which has been universally applied since then. The IGCCC, based on clinical parameters present prior to therapy, divides patients with metastatic testicular cancer into three groups with good, intermediate or poor prognosis, respectively. The tumor marker decline during chemotherapy has, in retrospective analyses, demonstrated a strong impact on outcome 1-5. The rationale behind the SWENOTECA IV protocol of 1995 was that risk factors present at staging were not considered specific enough to be used for treatment intensification and decision-making. Hence, this SWENOTECA IV protocol was designed for early identification of patients, in whom the response to two standard chemotherapy courses was inadequate and to provide intensified treatment to these individuals. SWENOTECA (Swedish Norwegian testicular cancer project) was initiated 1981 as a collaboration between Swedish and Norwegian health services, providing mutual cancer care programs, clinical research projects and quality assurance. Herein we summarize the SWENOTECA IV treatment strategy with respect to prognostic groups and outcome. 4

5 Study design SWENOTECA IV is a prospective population-based non-randomized study in metastatic testicular non-seminomatous germ cell cancer (NSGCT). The individually response-adapted therapy was based on tumor marker decline and, in marker negative patients, on radiological assessment. An overview of the treatment principles is shown in Figure 1. The study was approved by the Swedish and Norwegian Medical Ethics Committees. Patients and clinical methods Patients All adult patients diagnosed with metastatic NSGCT between July 1, 1995 and December 31, 2003 in Sweden and Norway, with the exception of one Norwegian centre, were included. Exclusion criteria were extragonadal tumor and previous treatment for contralateral testicular tumor. Patients with NSGCT in retroperitoneal lymph node metastases without viable cancer in the ablated testicle, but only fibrosis/scarring, were included. All Swedish centres report to the SWENOTECA database which is cross-checked with the Swedish national cancer registry once a year. This ensures that all Swedish patients diagnosed with NSGCT are included in the material. Five Norwegian centres stage and treat patients with testicular cancer, four of which report to the SWENOTECA database. Completeness of the Norwegian data was ensured by thorough searches made in hospital databases and the findings cross-checked with the SWENOTECA database. The survival status of all patients was checked against the national population registry in Sweden as of May 1, 2009 and in Norway September 1, Staging Staging was performed according to the Royal Marsden Hospital staging system (RMH) 6. The staging procedures included physical examination, serum tumor marker concentrations 5

6 (AFP, β-hcg, LDH) before and after orchiectomy and computed tomography scan (CT) of the thorax, abdomen and pelvis. For patients without obvious metastases, but with elevated tumor markers after orchiectomy, the markers were followed weekly and definitive staging was postponed until normalized or definitively increased. Initially, risk grouping was performed only according to the Medical Research Council system (MRC) 7 but since 1997, classification according to the IGCCC 8 has also been performed. Patients registered before 1997 were retrospectively classified according to the IGCCC. Treatment principles Initial chemotherapy Treatment strategy is presented in Figure 1 and chemotherapy regimens in Table 1. All patients were initially treated with 2 cycles of BEP. In case of severe renal impairment, not due to tumor obstruction, cisplatin was replaced by carboplatin. In patients with severely impaired respiratory function, bleomycin was replaced by ifosfamide, PEI. Tumor marker decline was calculated for day 5/6, 15/16 and 22. Response evaluation was performed after 2 BEP. In case of satisfactory response, BEP was continued. Patients with small volume disease initially, normal tumor markers after 1 cycle of BEP, and complete remission (CR on CT) after 2 cycles of BEP received a total of 3 cycles. The remaining patients who responded satisfactorily, received 4 cycles of BEP. After 4 cycles, a second evaluation was performed with CT and tumor markers. If tumor markers still were elevated, but declining satisfactorily, one cycle of PEI was given before surgery. Intensification step 1 Unsatisfactory tumor marker decline was considered as insufficient response to treatment and 6

7 the therapy was intensified by addition of ifosfamide. After the first cycle of BEP-If, a peripheral stem cell harvest was performed and tumor marker decline was calculated for day 5/6, 15/16 and 22. Patients responding satisfactorily were treated with another 1 to 2 cycles. Normally, two cycles of chemotherapy were given after tumor marker normalization. Intensification step 2 An unsatisfactory tumor marker decline during treatment with BEP-If was considered a failure, and the treatment was further intensified by HDSCT with stem cell support. Two cycles were planned, given with an interval of 6-7 weeks. Response evaluation AFP and β-hcg were analyzed days 1, 5/6, and 15/16 in each chemotherapy course. The values were plotted in a semi logarithmic graph for AFP and β-hcg, respectively. Two consecutive marker values demonstrating an inadequate marker decrease (i.e. β-hcg T½ > 3 days, AFP T½ > 7 days), indicated a treatment failure. For the first marker evaluation, values from day 15/16 of the first course and days 1, 5/6 and 15/16 of the second course were used, taking into account a possible surge, i.e. a serum tumor marker increase at the initiation of therapy 9. The marker status was followed through all treatment courses. For marker negative patients post-orchiectomy, the evaluation was based on tumor volume reduction calculated by the product of two perpendicular axial diameters on CT. If the tumor responded with a shrinkage of >25% after the two initial chemotherapy courses, this was considered an adequate response Post chemotherapy surgery In patients with retroperitoneal lymph node metastases larger than two cm at staging, a 7

8 retroperitoneal lymph node dissection (RPLND) was recommended for all after primary chemotherapy. Residual tumors in other locations should also be resected if possible. In patients where the resected specimen contained only necrotic debris/fibrosis or teratoma, no further treatment was given, whereas patients with remaining vital cancer in the resected tissue received adjuvant chemotherapy with two cycles of PEI. Patients intensified to step 1 (BEP- If/PEI) pre-surgery received one cycle of adjuvant HDSCT.. Toxicity Acute toxicities, gastrointestinal, hematological, renal, pulmonary, neurological and infectious were all graded according to the WHO toxicity criteria. Follow up After completion of treatment, patients were followed according to a standardized schedule with visits every two months the first two years, every three months the third year, every six months years four to five and every 12 months years 6 to 10. Endpoints and statistical methods Overall survival (OS) was defined as time from orchiectomy to death from any cause or last follow-up (FU). Progression-free survival (PFS) was defined as time from orchiectomy to progression of disease or relapse death from any cause or last FU. A contra lateral tumor was not considered a relapse. For patients who never reached a disease free status, time to progression was set to one day. Cancer specific survival (CSS) was defined as time from orchiectomy to death from NSGCT, or treatment or to last FU/death from other causes. Survival curves were estimated using the Kaplan-Meier method using the log-rank test for comparisons. SPSS 17.0 was used for the statistical analyses. 8

9 Results From a total of 610 patients were diagnosed with metastatic testicular NSGCT, 418 in Sweden and 192 in Norway. For one patient medical records were lost, four patients were diagnosed at autopsy and two patients were neither staged nor treated (severe mental retardation; 91 years and dementia). These seven patients (1.1%) are not included in the analyses. Patient characteristics and therapy-related data are presented in Table 2. The median age of the 603 included patients was 28 years and 5.3 % were >50 years. Ninety four percent were considered tumor-free after primary treatment, comprising conventional dose chemotherapy (BEP only or with the addition of ifosfamide) +/- surgery in 95% of the patients and high dose chemotherapy with stem cell support (HDSCT) +/- surgery in 5% (Figure 2). After a median FU time of 99 months (range , 32% followed >10 years), 52 patients relapsed (9%). The median time to recurrence was eight months (range 1-118). There were 11 late relapses (LR), occurring more than two years after treatment termination, corresponding to a 10-year cumulative incidence of 2.7% (95%CI, ). The 10-year OS post relapse was 74.3% with a median FU time of 96 (range ) months. A total of 61 patients have died, 52 due to progressive disease (n=41), or treatment (n=11). Only one patient with LR died. Nine patients died of other causes without evidence of persistent NSGCT: heart disease (n=2), accidents (n=2), suicide (n=2), chronic alcoholism (n=2), multiple myeloma (1). Good prognosis patients According to the IGCC classification, 395 patients (65.5%) belonged to the good prognosis group. All but nine patients (97.7%) became tumor free after primary treatment including post- 9

10 chemotherapy surgery in 50% of the patients. The administered chemotherapy was BEP alone in the vast majority (88.6%) of the patients, <4 courses in 33.2%, 4 courses in 65.9% and >4 courses in 0.9% of the patients. There were 32 relapses, of which 22 (69%) were located in retroperitoneal nodes. Half of them (17) were mature teratomas and treated with surgery only, while the rest were treated with pre- or post surgical chemotherapy. None received high-dose chemotherapy. Four patients (1.3%) with relapse died. Nine (28%) were LR occurring after median 4.5 years (range ). Of these, all were cured, 56% with surgery alone. Of all good prognosis patients, 20 patients (5.1%) died, 14 due to NSGCT or treatment, six due to other causes. OS, CSS and PFS at 10 years were 94.7 %, 96.4% and 87.4 % respectively. Intermediate prognosis patients According to the criteria for intermediate prognosis, 114 (18.9%) patients were allocated to this group due to moderately elevated tumor markers (AFP ng/ml and/or β-hcg IU/L and/or LDH 1.5xN- 10xN) 8. In 10 patients, preorchiectomy values were used for risk group stratification: in seven patients chemotherapy was started before orchiectomy due to advanced disease and poor medical condition, and in three patients postorchiectomy values were missing. Nine of these 10 patients are alive in first complete remission. A total of 108 (95%) patients were disease free after primary treatment, which included chemotherapy with median four courses BEP (range 2-6) in 83 (77%), intensification step 1 with median two courses BEP-IF/PEI (range 1-4) in 22 (20%) and further intensification with HDSCT in three (2.8%) patients. Post-chemotherapy surgery was performed in 76 (71 %) patients. Relapses occurred in 10 (9%, 3 LR) patients within a median of six months (range 2-56). Death occurred in 11 (9.7%) patients, nine due to NSCGT or treatment and two due to other causes. The 10-year OS, CSS 10

11 and PFS were 90.0%, 91.7% and 84.9% respectively. OS and PFS in the intermediate risk group were not significantly different from the good prognosis group (p=0.066 and p=0.446). Poor prognosis Ninety-four (15.6%) patients were categorized as poor prognosis patients. The poor prognostic factor was poor tumor markers only (AFP 10,000 ng/ml or β-hcg 50,000 mu/ml or LDH 10 times the upper normal) in 38 patients while 56 had non-pulmonary visceral metastases +/- poor markers (Table 3). Of 71 (76 %) patients who became tumor free after primary treatment: eight (11%) were treated with BEP only, 39 (55%) were intensified with ifosfamide addition, and 24 (34%) received HDSCT. Post-chemotherapy surgery was performed in 58 (82 %) patients. Nine (13%) patients relapsed, after a median of two months (range: 1-12 months). Thirty patients (32%) died, 29 due to disease or treatment, one due to other causes. The 5- and 10-year OS, CSS and PFS were similar at 67%, 68% and 64% respectively. Patients with poor markers only had a significantly more favorable outcome in comparison to patients with non-pulmonary visceral metastases as the 10-year OS was 81.6% and 58.1% respectively (p=0.032). Toxicity The treatment toxicity corresponded to previous experience with standard BEP, ifosfamide containing chemotherapy and HDSCT. A total of 11 patients died due to treatment related complications; sepsis (n=3), myocardial infarction (1), intracerebral hemorrhage (2), kidney failure (1), liver cirrhosis (1), complications to surgery (1) and two not further specified. Two of the 31 patients who received HDSCT died due to treatment. 11

12 Discussion Herein, we report the clinical characteristics and long term treatment results of 603 patients diagnosed with metastatic testicular NSGCT July 1995 to December 2003 in Sweden and Norway. The results were obtained with a treatment protocol in which the efficacy of initial chemotherapy, mainly indicated by the rate of tumor marker decline, guided further therapy in the individual patient. The study was community-based, including 27 Norwegian and Swedish hospitals, but with surgical procedures for metastatic lesions concentrated to the university hospitals. The results are favorable with a 10-year OS and CSS of 90 and 91%, respectively, which is well comparable to results reported from large single institutions 10. The distribution of patients on IGCCC prognostic groups was as expected, with the majority in the good prognosis group (66%) followed by the intermediate (19%) and poor prognosis (16%) group. Compared to the original report from the IGCCCG 8 the proportion of good prognosis patients was somewhat larger (56 vs. 66 %) and that of intermediate prognosis patients somewhat smaller (28 vs. 19%). This might be explained by the fact that the patient population, on which the IGCCCG analyses were performed, were diagnosed and staged years earlier than our population. Furthermore patients with extragonadal NSGCT were excluded in our study. Our data on distribution between prognostic groups are comparable to that of the Spanish Germ-Cell Cancer Group 11, with patients diagnosed and staged during the same time period, and reflects the current situation in Sweden and Norway. Patients assessed to have a good prognosis according to the IGCC classification and treated with BEP chemotherapy have an excellent cure rate. Our treatment results in good prognosis 12

13 patients are consistent with the previously reported 5-year OS and PFS at 95% and 90%, respectively. These survival rates have been rather constant during the last 20 years 8,12. In good prognosis patients, it has recently been established that treatment with 3 courses of BEP is as effective as 4 courses 13. In the present series, the majority of good prognosis patients were treated with 4 courses (Table 2), which, in the light of current knowledge, might represent overtreatment. The outcome in the intermediate risk group was remarkable, with a 10-year OS of 90%, which compares favorably with results from a meta-analysis of 232 intermediate prognosis patients from ten series published between 1998 and 2004 (5-year OS of 83%) 12, and to a recent randomized trial with an estimated 5-year OS of 80% in the intermediate risk group 14. Frequently, intermediate and poor risk patients are pooled together and treated with similar protocols 14,15. In this series, the outcome of patients in these two risk groups was fundamentally dissimilar. On the other hand, we found no significant difference in OS and PFS between the good and intermediate risk groups in our series. The results in the poor prognosis group were unsatisfactory, as in previous series, exemplified by two recent randomized studies 16,17, underscoring the need for novel treatment strategies and therapeutic agents for this population. Interestingly, the outcome for patients with major marker elevation only was significantly superior to patients with non-pulmonary visceral metastasis, which has been noted previously 18. This may be a result of our treatment strategy directed at early detection of slow marker decline, but could also reflect an underlying difference in tumor biology between these two patient populations. 13

14 Primary extragonadal germ cell tumors (EGCT), accounting for 2-5% of all germ cell tumors, were not included in this SWENOTECA protocol. The majority (60-70%) of EGCTs are located in the mediastinum and are all categorized as poor prognosis in the IGCCC. These are biologically distinct from and have a worse prognosis than primary testicular GCT. The exclusion of this poor prognostic group may have contributed to a more favorable outcome in our IGCCC poor prognosis group. However, a GCT localized in the retroperitoneum, with no clinically obvious testicular tumor, is likely to be associated with an occult or burned-out testicular tumor 19. The treatment outcome is also similar to that of their testicular counterparts 20. In our study population, six patients (1%) presented with a retroperitoneal tumor and a fibrotic testicular scar (good 2; intermediate 3; poor prognosis 1). Thus, the possible exclusion of a limited number of patients with a retroperitoneal EGCT is not expected to have any impact on our treatment results in the intermediate and good prognosis groups. The SWENOTECA IV protocol was designed to tailor treatment for the individual according to efficacy of initial standard chemotherapy and to use tumor marker decline and radiological evaluation during initial standard chemotherapy as an early indicator of treatment efficacy. This strategy has shown to be feasible in a population-wide trial with near-complete coverage. Furthermore, in view of the favorable treatment results in both the intermediate prognosis group and in the poor prognosis patients with marker elevation only, the strategy of early detection of treatment failure and subsequent treatment intensification is worthy of further evaluation in the treatment of advanced NSGCT. Acknowledgements The authors thank Monika Andersson, for excellence in data management and secretarial assistance during the trial. 14

15 Table 1. Chemotherapy regimens Regimen Drug Daily Dosing (mg/m 2 ) Initial therapy BEP* Bleomycin Etoposide Cisplatin Step 1 BEP-if Bleomycin Etoposide Cisplatin Ifosfamide Mesna Mesna (po) IE IE Day 1,5, ,5, PEI Step 2 HDSCT cycle 1 Etoposide Cisplatin Ifosfamide Mesna Mesna (po) Cyclophosfamide Etoposide Carboplatin AUC= HDSCT cycle 2 Cyclophosfamide Thiotepa Carboplatin x2 AUC= * In case cisplatin was replaced by carboplatin, the dose was AUC=7 according to Calverts formula 21 Bleomycin was omitted after a cumulative dose of IU 15

16 Table 2. Characteristics, treatment, relapse and survival of 603 patients with metastatic nonseminomatous testicular cancer All Good Intermediate Poor Relapse prognosis prognosis prognosis N (%) 603 (100) 395 (65.5) 114 (18.9) 94 (15.6) 51 (9.0) Age median (range) 28 (16-70) 28 (16-69) 28 (16-70) 28 (16-62) 28 (16-65) AFP* median ng/ml (range) 19 ( ) 12 (1-890) 127 (1-7037) 321 ( ) 36 ( ) ß-HCG median IU/L (range) 18 ( ) 5 (0-4735) 240 ( ) ( ) 76 ( ) Marker negative Preorchiectomy N (%) 92 (15) 80 (20) 8 (7) 3 (3) 3 (3) Postorchiectomy N (%) 152 (26) 137 (35) 10 (10) 5 (6) 8 (5.3) Clinical stage Mk+/II/III/IV (%) 9/50/5/36 13/60/4/23 1/45/10/45 0/15/2/83 16/8/13/8 Clinical stage II A/B/C (%) 22/57/21 27/62/8 6/49/39 0/7/93 12/4/14 Primary treatment BEP N (%) Postchemo surgery (%) 436 (77) (51) 346 (89) (47) 83 (77) (68) 7 (10) 6 (86) 39 (75) (8.1) BEP+BEP-IF/PEI N (%) Postchemo surgery (%) 99 (18) (82) 38 (10) (76) 22 (20) (86) 39 (56) (85) 13 (12) (14) HDSCT N (%) 29 (5) 2 (1) 3 (3) 24 (34) 0 Tumor-free after primary therapy N (%) 564 (94) 386 (98) 108 (95) 70 (74) Relapse N (%) 52 (9.0) 32 (8.3) 10 (9.3) 9 (12.9) Time to relapse (m) median (range) 8 (1-118) 11 (2-118) 6 (2-56) 2 (1-12) Survival FU (m) median (range) 99 (24-162) 99 (48-162) 92 (24-159) 104 (49-159) 10-year OS % year CSS % year PFS % * missing, good/intermediate/poor: 1/0/2 missing, good/intermediate/poor: 1/0/0 According to MRC (10-148) Pairwise comparison: Good vs. Intermediate p=0.066, Good vs. Poor p=0.000, Intermediate vs. Poor

17 p=0.000 Pairwise comparison: Good vs. Intermediate p=0.045, Good vs. Poor p=0.000, Intermediate vs. Poor p=0.000 Pairwise comparison: Good vs. Intermediate p=0.446, Good vs. Poor p=0.000, Intermediate vs. Poor p=0.000 Table 3. Prognostic factors in 94 poor prognosis patients Poor prognosis due to Patients N (%) AFP ng/ml median (range) Markers only 38 (40) 7091 ( ) APF >10000 ng/ml β-hcg >50000* IU/L 19 (20) ( ) 18 (20) 21 (1-3782) β-hcg IU/L median (range) ( ) 200 ( ) ( ) LD > 10 x ULN year OS % 10-year CSS % Non-pulmonary visceral metastasis 56 (60) 135 ( ) One site 42 (45) 106 ( ) liver 21 (22) 409 ( ) brain 9 (10) 9 ( ) bone 3(3.2) 6 ( ) other** 9 (10) 38 ( ) >One site 14 (15) 237 ( ) ( ) 300 ( ) ( ) 3549 ( ) ( ) 300 ( ) ( ) *Four patients started chemotherapy before orchiectomy **Adrenal gland, pleura, pericardium, peritoneum, intraspinal, soft tissue in mediastinum and in pelvis 17

18 Figure legends Figure 1 Treatment principles Abbreviations: M±: tumor markers increased/normal; CR: complete remission on CT or MRI; RPLND: retroperitoneal lymph node dissection; FU: follow up; HDSCT: high dose chemotherapy with stem cell rescue; BEP, BEP-if, PEI, HDSCT: see Table 1 *according to MRC 7, ** see Post chemotherapy surgery Figure 2 Survival of all 603 metastatic NSGCT and according to the IGCC classification. Panel A, Overall survival; Panel B, Cause-specific survival; Panel C, Progression-free survival Figure 3 Proportions according to treatment intensity in patients who were tumor free after the primary treatment 18

19 Figure 1 Small volume disease* Large and very large* volume disase Good response = t½ AFP 7 days t½ β-hcg 3 days BEP1 Poor response = t½ AFP > 7 days t½ β-hcg > 3 days BEP1 M+ BEP2 If progress during BEP BEP-if If progress during BEP-if Individual treatment M- BEP2 <CR Good response Poor response CR BEP3 M- BEP3+4 M+ BEP-if1 Stem cell harvest Surgery** PEIx1 BEP-if/PEI2 FU Surgery** Good response Poor response No malignant cells FU Malignant cells PEIx2 No malignant cells Surgery** Malignant cells HDCT1 HDCT2 Surgery** before or after HDCT FU FU HDCT FU FU 19

20 Figure 2 a-c A B C 20

21 Figure 3 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% eel del 5% High dose chemotherapy with stem cell rescue Addition of ifosfamide (BEP-if, PEI=VIP) BEP 0% 0% All patients 21

22 Appendix Investigators at the following hospitals have participated in the study: Haukeland University Hospital, Bergen; Borås Hospital, Borås; Mälarsjukhuset Hospital, Eskilstuna; Gävle Hospital, Gävle; Sahlgrenska University Hospital, Göteborg; County Hospital Ryhov, Jönköping; Central Hospital, Karlstad; Linköping University Hospital, Linköping; Sunderby Hospital, Luleå; Lund University Hospital, Lund; University Hospital MAS, Malmö; Oslo University Hospital, Oslo; Stavanger University Hospital, Stavanger; Danderyd Hospital, Stockholm; Karolinska University Hospital Solna, Stockholm; Karolinska University Hospital Huddinge, Stockholm; Karolinska University Hospital Södersjukhuset, Stockholm; Sundsvall Hospital, Sundsvall; University Hospital of North Norway, Tromsø; St. Olav University Hospital, Trondheim; Umeå University Hospital, Umeå; Uppsala University Hospital, Uppsala; Västerås Hospital, Västerås; Växjö Central Hospital, Växjö; Ålesund Hospital, Ålesund; Örebro University Hospital, Örebro; Östersund Hospital, Östersund. 22

23 References 1. Schultz H, Sell A, Norgaard-Pedersen B, Arends J. Serum alpha-fetoprotein and human chorionic gonadotropin as markers for the effect of postoperative radiation therapy and/or chemotherapy in testicular cancer. Cancer. 1978;42: Lange PH, Vogelzang NJ, Goldman A, Kennedy BJ, Fraley EE. Marker half-life analysis as a prognostic tool in testicular cancer. J Urol. 1982;128: Vogelzang NJ, Lange PH, Goldman A, Vessela RH, Fraley EE, Kennedy BJ. Acute changes of alpha-fetoprotein and human chorionic gonadotropin during induction chemotherapy of germ cell tumors. Cancer Res. 1982;42: Picozzi VJ, Jr., Freiha FS, Hannigan JF, Jr., Torti FM. Prognostic significance of a decline in serum human chorionic gonadotropin levels after initial chemotherapy for advanced germ-cell carcinoma. Ann Intern Med. 1984;100: Murphy BA, Motzer RJ, Mazumdar M, et al. Serum tumor marker decline is an early predictor of treatment outcome in germ cell tumor patients treated with cisplatin and ifosfamide salvage chemotherapy. Cancer. 1994;73: Peckham MJ, McElwain TJ, Barrett A, Hendry WF. Combined management of malignant teratoma of the testis. Lancet. 1979;2: Peckham MJ, Bagshaw KD, Blandy JP. Prognostic factors in advanced non-seminatous germ-cell testicular tumours: Results of a multicenter study. Report from the Medical Research Council Working Party on Testicular Tumours. Lancet. 1985;1:8. 8. 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: Horwich A, Peckham MJ. Transient tumor marker elevation following chemotherapy for germ cell tumors of the testis. Cancer Treat Rep. 1986;70: Ehrlich Y, Brames MJ, Beck SD, Foster RS, Einhorn LH. Long-term follow-up of Cisplatin combination chemotherapy in patients with disseminated nonseminomatous germ cell tumors: is a postchemotherapy retroperitoneal lymph node dissection needed after complete remission? J Clin Oncol;28: Germa-Lluch JR, Garcia del Muro X, Maroto P, et al. Clinical pattern and therapeutic results achieved in 1490 patients with germ-cell tumours of the testis: the experience of the Spanish Germ-Cell Cancer Group (GG). Eur Urol. 2002;42: ; discussion

24 12. van Dijk MR, Steyerberg EW, Habbema JD. Survival of non-seminomatous germ cell cancer patients according to the IGCC classification: An update based on meta-analysis. Eur J Cancer. 2006;42: de Wit R, Roberts JT, Wilkinson PM, et al. Equivalence of three or four cycles of bleomycin, etoposide, and cisplatin chemotherapy and of a 3- or 5-day schedule in good-prognosis germ cell cancer: a randomized study of the European Organization for Research and Treatment of Cancer Genitourinary Tract Cancer Cooperative Group and the Medical Research Council. J Clin Oncol. 2001;19: Culine S, Kramar A, Theodore C, et al. Randomized trial comparing bleomycin/etoposide/cisplatin with alternating cisplatin/cyclophosphamide/doxorubicin and vinblastine/bleomycin regimens of chemotherapy for patients with intermediate- and poor-risk metastatic nonseminomatous germ cell tumors: Genito-Urinary Group of the French Federation of Cancer Centers Trial T93MP. J Clin Oncol. 2008;26: Fossa SD, Paluchowska B, Horwich A, et al. Intensive induction chemotherapy with C-BOP/BEP for intermediate- and poor-risk metastatic germ cell tumours (EORTC trial 30948). Br J Cancer. 2005;93: Droz JP, Kramar A, Biron P, et al. Failure of high-dose cyclophosphamide and etoposide combined with double-dose cisplatin and bone marrow support in patients with high-volume metastatic nonseminomatous germ-cell tumours: mature results of a randomised trial. Eur Urol. 2007;51: ; discussion Motzer RJ, Nichols CJ, Margolin KA, et al. Phase III randomized trial of conventional-dose chemotherapy with or without high-dose chemotherapy and autologous hematopoietic stem-cell rescue as first-line treatment for patients with poor-prognosis metastatic germ cell tumors. J Clin Oncol. 2007;25: van Dijk MR, Steyerberg EW, Stenning SP, Habbema JD. Identifying subgroups among poor prognosis patients with nonseminomatous germ cell cancer by tree modelling: a validation study. Ann Oncol. 2004;15: Scholz M, Zehender M, Thalmann GN, Borner M, Thoni H, Studer UE. Extragonadal retroperitoneal germ cell tumor: evidence of origin in the testis. Ann Oncol. 2002;13: Hartmann JT, Nichols CR, Droz JP, et al. Prognostic variables for response and outcome in patients with extragonadal germ-cell tumors. Ann Oncol. 2002;13: Calvert AH, Newell DR, Gumbrell LA, et al. Carboplatin dosage: prospective evaluation of a simple formula based on renal function. J Clin Oncol. 1989;7: