Original article. J. Pont, 1 ' 2 C. Bokemeyer, 3 A. Harstrick, 4 F. Sellner, 5 H. Greinix 6 & F. Stoiber 1 ' 7

Transcription

1 Annals of Oncology 8:,. Kluwer Academic Publishers. Printed in the Netherlands. Original article Chemotherapy for germ cell tumors relapsing after highdose chemotherapy and stem cell support: A retrospective multicenter study of the Austrian Study Group on Urologic Oncology J. Pont, ' C. Bokemeyer, A. Harstrick, F. Sellner, H. Greinix & F. Stoiber ' Study Group on Urologic Oncology (AUO) of the Austrian Society of Urologists, Medical Department III (Oncology), Kaiser Franz Josef Spital, Vienna, Austria; * Medical Services Department, Unit II, EberhardKarlsUniversitdt, Tubingen; * University Medical Center Essen, Medical Service, German Cancer Center, Essen, Germany; s Department of Surgery, Kaiser Franz Josef Spital, University Medical Center, Medical Department I (Bone Marrow Transplantation Service), Vienna; ^Department of Urology, Krankenhaus der Barmherzigen Schwestern, Linz, Austria * See page for list of other study participants Summary Background: In most patients with advanced refractory germ cell tumors undergoing highdose chemotherapy with stem cell support (HDCT) the disease progresses after HDCT. This study was designed to shed light on the unestablished role of posthdct chemotherapy. Patients and methods: In a retrospective multicenter study data of evaluable patients from nine centers subjected to posthdct chemotherapy for progression of their germ cell tumors were collected in a questionnaire survey and analyzed for treatment response and survival. Results: Of patients pretreated by HDCT, 8 (%) were subjected to posthdct chemotherapy for disease progression. Remission was achieved in (%) and markernegative remission in eight (%). The median survival time was weeks, weeks for responders and weeks for nonresponders. Only one of evaluable patients achieved sustained complete remission. Remissions significantly correlated with the posthdct interval, the use of ifosfamide and the combination regimens of cisplatin + etoposide i.v. or ifosfamide and of paclitaxel + ifosfamide or cisplatin. On univariate analysis a longer posthdct interval, the use of cisplatin, paclitaxel and ifosfamide and the combined use of paclitaxel + ifosfamide and/or cisplatin significantly improved the chances of survival. On multivariate analysis only treatment with paclitaxel and ifosfamide retained independent prognostic significance for survival. Conclusions: One third of the patients considered to be candidates for further chemotherapy once progressive after HDCT went into remission with a gain in survival time. Sustained remissions may occur, but are rarely seen. Paclitaxel and ifosfamide appear to be the most effective drugs in these heavily pretreated patients. Key words: cisplatinrefractory patients, germ cell cancer, ifosfamide, paclitaxel, posthdct chemotherapy Introduction Of the germ cell tumor patients failing standard firstline treatment, only % will benefit from currently established secondline treatment modalities by lasting remissions []. Depending on preceding prognostic variables [], most of them must be expected to progress even after highdose carboplatinbased chemotherapy plus hematopoietic progenitor cell support (HDCT). Counseling patients on further treatment following HDCT is a demanding task: The patients are all young and heavily pretreated; the chances of remission with a useful quality of life are poor []; but many want to go on with antitumor treatment. Phase II trials in refractory patients following salvage therapy suggested daily oral etoposide [, ] and more recently paclitaxel [] to be useful treatment options in this patient population. The benefits of chemotherapy for patients relapsing after HDCT has not been established to date. As this issue will become increasingly important, but prospective studies are unfeasible because of the expectedly small number of patients, a retrospective analysis was conducted. Patients and methods Data collection Thirty centers in Austria, Germany and Switzerland likely to have germ cell tumor patients after HDCT on their followup programs were contacted in writing. Of the 0 replying centers, 0 had patients receiving posthdct chemotherapy for disease progression on their data files. Nine of them, i.e., six in Austria and three in Germany with a total number of posthdct germ cell tumor patients, contributed data of cases with posthdct chemotherapy. Data were collected with a standardized questionnaire. Items requested were patient identification, type and number of prehdct cisplatinbased chemotherapy courses, date of HDCT and posthdct response evaluation, site

2 0 of posthdct disease progression; time, type, dosage and number of posthdct chemotherapies including response evaluation; factors prompting discontinuation of chemotherapeutic modalities; condition at last followup or date and cause of death. The completed questionnaires were checked for plausibility, completeness and consistency of the data by J.P., but no further checks, i.e., laboratory studies, histology or radiology reviews, or other source data verification, were performed. Only patients with complete data sets were included in the analysis. Data were entered in a personal computer and data entry was checked by J.P. and F.S. Definitions Complete remission (CR) was defined as clinical CR (= total disappearance of the tumor by all radiological studies as well as normalization of human chorionic gonadotropin (HCG) and alfafetoprotein (AFP) levels with chemotherapy alone or pathologic CR (= complete resection of nonviable malignancy or of mature teratoma without any evidence of residual disease). Markernegative partial remission (PRm) was defined as reduction of any radiologically measurable disease and normal HCG and AFP levels. Markerpositive partial remission (PRm+) was defined as reduction of any radiologically measurable disease by greater than 0%, but elevated HCG and /or AFP levels or a greater than 0% reduction of elevated tumor markers. Progressive disease (PD) was defined as increase of any radiologically measurable tumor mass by greater than % or a greater than 0% increase in elevated tumor markers. change (NC) was defined as response neither qualifying as PRm nor as PRm+ or PD. Response was defined as CR, PRm or PRm+. response was defined as NC or PD. Before the response to any chemotherapeutic modality was evaluated, patients had to have completed at least one full chemotherapy course with a survival time of weeks or more. Highdose chemotherapy. Patients evaluated in this study had received HDCT with either carboplatin + etoposide + ifosfamide (CEI) [8], carboplatin + etoposide + cyclophosphamide (CEC) [] or cisplatin + etoposide + ifosfamide (HDPEI) [0] with hematopoietic progenitor or stem cell support per protocol as described in [80]. Interval was defined as the time in weeks between the (last) HDCT and the next posthdct chemotherapy. Survival was defined as the survival time in weeks counted from the beginning of posthdct chemotherapy. Patients Inclusion criteria. Patients with germ cell tumors were eligible, if they had undergone systemic posthdct chemotherapy for disease progression and if a complete set of the data requested in the questionnaire was available for them. Of the patients, two were excluded: In one posthdct chemotherapy had been administered locally (intrathecally) rather than systemically, while the other one was lost to followup so that survival was not evaluable. This left a total of evaluable cases. All of them had nonseminomatous germ cell cancers. Detailed patient data are shown in Table. Data analysis and statistics As preceding variables (site of primary tumor, number of prehdct chemotherapy courses, HDCT regimen, response to HDCT, interval between HDCT and posthdct treatment, metastasizing pattern) may have been more important for the treatment outcome in these heavily pretreated patients than the type of posthdct chemotherapy, these were also correlated with response and survival. Correlating the different chemotherapy regimens applied with response and survival did not appear to be useful for various reasons: Patients included in this retrospective studyreceiveda variety of post HDCT chemotherapy regimens (Table ); up to six different regimens with a variable number of treatment courses were administered; the time to response varied (first to third chemotherapy regimens); the doses applied differed; and the number of patients was small. To assess the efficacy of the cytostatic drugs used patients were, therefore, allocated to two groups, i.e., those receiving a single drug or drug combination and those who did not. The resultant paired groups were subjected to univariate and multivariate analysis and compared for the best response to posthdct chemotherapy and for survival time. Response was analyzed by the chisquare test and by Fischer's exact test. For survival curves the KaplanMeier plot and the logrank test were used []. Multivariate survival analysis was done with Cox's proportional hazards regression model []. P values of less than 0.0 were considered to indicate statistical significance. Results Of the germ cell tumor patients who had received HDCT in the nine participating centers, 8 (%) underwent systemic posthdct chemotherapy for disease progression. The evaluable patients received one to six different chemotherapy regimens (median, two regimens) with one to nine courses (median, two courses). In only two patients treatment was discontinued because of drug toxicity; in all others disease progression prompted drug withdrawal or replacement by another regimen. Table Patient characteristics. Age (years) Primary tumor Testicular Extragonadal Chemotherapy courses before HDCT Highdose regimen CEC CEI HDPEI Response to HDCT CR PRm PRm+ NC PD Interval (weeks) npulmonary visceral metastases, = 0 0 Range 08 0 Median Abbreviations: HDCT highdose chemotherapy; CEC carboplatin + etoposide + cyclophosphamide []; CEI carboplatin + etoposide + ifosfamide [8]; HDPEI highdose cisplatin + etoposide + ifosfamide [0]; CR complete remission; PRm markernegative partial remission; PRm+ markerpositive partial remission, NC no change; PD progressive disease. 0 8

3 Table. Response to posthdct chemotherapy regimens. Daily oral etoposide Paclitaxel ajone Paclitaxel + cisplatin Paclitaxel + ifosfamide Paclitaxel + cisplatin + ifosfamide Cisplatin ajone Cisplatin + etoposide Cisplatin + etoposide + ifosfamide Cisplatin + gemcitabine POMB/ACE or POMB CISCA Carboplatin + etoposide +/ bleomycin +/ vinblastine Carboplatin + etoposide + ifosfamide Vinorelbine + ifosfamide + epirubicin Vindesine + ifosfamide + epirubicin Vinorelbine Vinblastine + bleomycin +/ methotrexate Vinblastine + methotrexate + actinomycin D Epirubicin +/ etoposide Gemcitabine Trofosfamide Oxaliplatin n Re response sponse _ 0 mycin; ACE actinomycin D, cyclophosphamide, etoposide; CISCA cisplatin, cyclophosphamide, adriamycin. Response In of the patients (%) posthdct chemotherapy achieved remissions: These were CR in four patients, PRm in four patients and PRm+ in nine patients. Of the responders, 0 responded to the first posthdct regimen, while seven did not, but were found to show objective evidence of a response to the second or third regimens. The correlation of the best response to posthdct chemotherapy with preceding variables (primary tumor, prehdct chemotherapy courses, HDCT regimen, response to HDCT, interval between HDCT and post HDCT chemotherapy, metastasizing pattern) is shown in Table. Of all these variables, only the interval between HDCT and posthdct chemotherapy for disease progression was significantly correlated with remissions: PostHDCT chemotherapy within three months of completing HDCT achieved remissions in no more than one of patients. By contrast, retreatment after more than one year was associated with a remission rate of four in six cases. The chemotherapy regimens used and the responses to them are shown in Table. As explained in 'Patients and methods', the great variety of chemotherapy regimens applied in this retrospective analysis necessitated patient grouping for a statistical analysis of the efficacy of chemotherapy: patients receiving a given drug or drug combination were compared with those not receiving Table. Best response to posthdct chemotherapy and preceding variables. Primary tumor Testicular Extragonadal Chemotherapy courses before HDCT < > Highdose chemotherapy regimen CEC CEI HDPEI Response to HDCT Response response Interval < weeks weeks weeks weeks > weeks npulmonary visceral metastases n Re re x p sponse sponse Abbreviations: HDCT highdose chemotherapy; CEC carboplatin, etoposide, cyclophosphamide [], CEI carboplatin, etoposide, ifosfamide [8]; HDPEI highdose cisplatin, etoposide, ifosfamide [0]. the drug or drug combination for the best response achieved. The pertinent data are shown in Table. Of the single drugs evaluated, only ifosfamide reached statistical significance. Among the drug combinations, cisplatin + etoposide, cisplatin + ifosfamide and cisplatin + etoposide + ifosfamide were significantly correlated with response. Paclitaxel in combination with ifosfamide and /or cisplatin showed the highest significance. Survival The median KaplanMeier survival estimates of the patients evaluated in this study was weeks. At the time of data analysis seven patients were still alive at to 0 weeks. Of these, two had progressive disease, four were in markerpositive partial remission (PRm+) and one in sustained complete remission without any treatment for months at the time this manuscript was written. Patients responding to posthdct chemotherapy for disease progression had a highly significant survival benefit with a projected median survival of weeks in responders and of weeks in nonresponders (logrank 8., P = ). Of the preceding variables evaluated by univariate analysis (Table ), only the interval between HDCT and posthdct chemotherapy reached statistical significance. Like the correlation of the response with the

4 Table. Best response to posthdct chemotherapy and drugs used. cytostatic Table. Survival after posthdct chemotherapy and cytostatic drugs used; KaplanMeier estimates, univariate analysis. Treated with Daily oral etoposide Cisplatin Paclitaxel lfosfamide Etoposide i.v. Cisplatin + etoposide Cisplatin + ifosfamide Cisplatin + ifosfamide + etoposide Paclitaxel + ifosfamide or + cisplatin n Response z P Treated with Daily oral etoposide Cisplatin Paclitaxel Ifosfamide Etoposide Cisplatin + etoposide Cisplatin + ifosfamide Cisplatin + ifosfamide + etoposide Paclitaxel + ifosfamide or + cisplatin Median survival (weeks) response Logrank P Table. Survival after posthdct chemotherapy and preceding variables; KaplanMeier estimates, univariate analysis. Median Log P survival rank (weeks) Primary tumor Testicular Extragonadal Chemotherapy courses before HDCT < > High dose chemotherapy regimen. 0. CEC CEI HDPEI Response to HDCT Response response 0 Interval < weeks weeks weeks weeks > weeks n pulmonary visceral metastases Abbreviations: HDCT highdose chemotherapy; CEC carboplatin, etoposide, cyclophosphamide []; CEI carboplatin, etoposide, ifosfamide [8]; HDPEI highdose cisplatin, etoposide, ifosfamide [0]. preceding variables, the survival in patients needing posthdct chemotherapy less than three months after HDCT was particularly poor. Table shows the survival data obtained by univariate analysis for patients receiving a given cytostatic drug or drug combination and for those who did not. Of the single drugs applied, cisplatin, paclitaxel and ifosfamide correlated significantly with survival, while daily oral etoposide did not. Among the drug combinations, only paclitaxel combined with ifosfamide or cisplatin reached significance. Of all the factors potentially affecting survival, i.e., preceding variables and cytostatic drugs used, only treatment with paclitaxel and ifosfamide retained independent prognostic significance (P = and 0.00, respectively) on multivariate analysis. While prognostically significant for survival on univariate analysis, the interval failed to retain independent prognostic significance on multivariate analysis. Discussion Most patients with metastatic germ cell tumors who relapse after firstline therapy or achieve incomplete responses show disease progression after conventionaldose salvage chemotherapy [] and HDCT, the only curative thirdline approach which is considered state

5 oftheart for testicular germ cell tumors responding incompletely to firstline treatment []. If they are ineligible for surgical salvage [] or palliative radiotherapy, the last resort for them would be to resume cytostatic chemotherapy. The primary objective of our survey was to find out how many patients undergo posthdct chemotherapy and how many of them appear to benefit from it in order to see whether posthdct chemotherapy makes any sense at all in this population. We found that posthdct chemotherapy was administered to no less than one quarter of all patients in the participating centers. In light of the posthdct relapse rate of more than 0% [], this is about one third of all patients relapsing after HDCT. As the other twothirds are apparently no longer willing or able to undergo chemotherapy, this represents a positive selection. The other objective of our study was to identify prognostic outcome variables for posthdct chemotherapy. For the outcome of HDCT progressive disease before HDCT, mediastinal primary tumor, refractory disease to conventionaldose cisplatin and (HCG >,000 U/l were defined as prognostic variables []. In our study the response to the previous highdose chemotherapy and the site of the primary tumor did not have prognostic significance for the outcome of post HDCT chemotherapy. Tumor markers before post HDCT chemotherapy were not included in our analysis. On univariate analysis only the interval between HDCT and posthdct chemotherapy, which probably is equivalent to progressionfree survival, correlated significantly with both response and survival. This agrees well with the clinical experience available. The benefits are poorest for patients who need posthdct chemotherapy within no less than three months after HDCT. Unlike the use of ifosfamide and paclitaxel for post HDCT chemotherapy, the interval did, however, not retain independent prognostic significance on multivariate analysis of the prognostic factors for survival. Of the many cytostatic drugs tested in cisplatinrefractory patients, only etoposide [], ifosfamide [8] and paclitaxel [] showed useful antitumor activity. The efficacy of these drugs in refractory patients who had previously undergone HDCT is still poorly understood. What little evidence is available was derived from studies in cisplatinrefractory patients who rarely had been given prior HDCT. Miller and Einhorn [] and Greco et al. [] reported 0 patients who had received daily oral etoposide for disease progression after HDCT. Of these, three responded and complete remission was sustained for three years in one of them. In phase II studies of paclitaxel in cisplatinrefractory patients Bokemeyer et al. [] and Gerl et al. [] found three of patients with disease progression after HDCT to respond, while Motzer et al. [] did not observe any response in any one of the five patients treated with posthdct paclitaxel. To the best of our knowledge data on posthdct ifosfamide treatment are not available. The response rate to daily oral etoposide was four in cases in our study. This agrees fairly well with reported rates for cisplatinrefractory patients [,, ]. Daily oral etoposide is widely used in these patients because of its relatively good tolerance and its suitability for outpatient treatment. However, daily oral etoposide did not show any prognostic significance for both response and survival in our study. Singledrug treatment with paclitaxel produced a response in no more than four of patients. Adding up the figures for all combinations of paclitaxel with other cytostatic drugs increased the response rate to eight in cases. For the combined use of paclitaxel with ifosfamide and /or cisplatin the response rate was four in five cases. As a result, the use of this combination was found to be a significant prognostic factor for response. Paclitaxel in this combination was also shown to have significance for survival. On multivariate analysis paclitaxel was retained as one of two independent prognostic factors for survival. The only patient in sustained pathologic complete remission after posthdct chemotherapy had received eight courses of paclitaxel + cisplatin at a dose of / mg/m. At the time this manuscript was submitted he had been in remission for 0 months counted from the beginning of posthdct chemotherapy and for months after completing treatment. These data appear to be related to the observation that paclitaxel may reverse acquired cisplatin resistance of germ cell tumors [], as was shown in in vitro studies [0]. It suggests once more that phase II and phase III trials with paclitaxel combinations are justified in poorrisk germ cell cancer patients [, ]. While correlated with survival on univariate analysis, the use of cisplatin for posthdct chemotherapy did not retain independent prognostic significance on multivariate analysis. Although ifosfamide had already been administered for HDCT in more than half of our patients, its use for posthdct chemotherapy was still significantly correlated with response and survival both on univariate and multivariate analysis. The composition of the preceding HDCT regimen did not show any statistically significant correlation with the outcome of posthdct chemotherapy. Our data do not show whether ifosfamide is only active in combination with cisplatin and/or paclitaxel or also as a single agent, because it was only administered in combination regimens to the patients in our study. The demands made on the heavily pretreated patients by resuming chemotherapy should be carefully weighed against what little chances they have for tumor control. Therefore, the likely benefits of another chemotherapy must be assessed in light of the performance status, expected toxicity and the cumulative toxicity to which the patients had been exposed in the past. The retrospective design of our study ruled out a reliable assessment of the performance status and an analysis of

6 toxicities. But it was remarkable to see that drug toxicity was only recorded twice as the factor prompting discontinuation of treatment. The cause of death in all nonsurvivors was the underlying germ cell tumor. Drugrelated deaths were not documented. With all due reservations befitting a retrospective study we conclude that one third of the patients who are still eligible for posthdct chemotherapy because of disease progression after HDCTwill benefit by remission and prolonged survival. This benefit is all the more likely to occur the longer the interval between HDCT and posthdct chemotherapy. Paclitaxel and ifosfamide combined with each other and with cisplatin turned out to be the most effective drugs. Acknowledgement The authors wish to thank Ulrich J. Pont for his assistance in preparing the manuscript. * Other study participants Klinikum GroBhadern, University of Munich, Germany: A. Gerl; Krankenhaus der Elisabethinen, Linz, Austria: O. Krieger; Wilhelminenspital, Vienna, Austria: Irene KOhrer; University Medical Center, Medical Department, Graz, Austria: Marianne Schmid. References. Einhorn LH. Salvage therapy for germ cell tumors. Semin Oncol, : Suppl :.. Beyer J, Kramar A, Mandanas R et al. Highdose chemotherapy as salvage treatment in germ cell tumors: A multivariate analysis of prognostic variables. J Clin Oncol ; : 8. Miller JC, Einhorn LH. Phase II study of daily oral etoposide in refractory germ cell tumors. Semin Oncol 0; (Suppl ):.. Motzer RJ, Bajorin DF, Schwartz LH et al. Phase II trial of paclitaxel shows antitumor activity in patients with previously treated germ cell tumors. J Clin Oncol ; : 8. Bokemeyer C, Schmoll HJ, Natt F et al. Preliminary results of a phase I/II trial of paclitaxel in patients with relapsed orcisplatinrefractory testicular cancer. J Cancer Res Clin Oncol ; 0:.. Bokemeyer C, Beyer J, Metzner B et al. Phase II study of paclitaxel in patients with relapsed or cisplatinrefractory testicular cancer. Ann Oncol ; :.. Bokemeyer C, Hartmann JT, Kuczyk MA et al. The role of paclitaxel in chemosensitive urological malignancies: current strategies in bladder cancer and testicular germcell tumors. World J Urol ; : 8 Siegert W, Beyer J, Strohscher I et al. Highdose treatment with carboplatin, etoposide, and ifosfamide followed by autologous stemcell transplantation in relapsed or refractory germ cell cancer: A phase I/II study. The German Testicular Cancer Cooperative Study Group. J Clin Oncol ; :.. Linkesch W, Greinix HT, Hocker P et al. Longterm follow up of phase I/II trial of ultrahigh carboplatin, VP, cyclophosphamide with ABMT in refractory or relapsed NSGCT. Proc Am Soc Clin Oncol ; : (Abstr ). 0. Bokemeyer C, Harstrick A, Ruther U et al. Sequential highdose VIPchemotherapy plus peripheral stem cell (PBSC) support for advanced germ cell cancer. Ann Oncol ; (Suppl ): (Abstr 0). Armitage P, Berry G. Statistical Methods in Medical Research, third ed. Oxford, UK: Blackwell Scientific.. Cox DR. Regression models and life tables. J R Stat Soc (B) ; : 80.. Motzer RJ. Selecting patients with cisplatinresistant germ cell tumors for highdose chemotherapy. J Clin Oncol ; :.. Murphy BR, Breeden ES, Donohue JP et al. Surgical salvage of chemorefractory germ cell tumors. J Clin Oncol ; :.. Greco FA, Johnson DH, Hainsworth JD. Chronic oral etoposide. Cancer ; : 0.. Gerl A, Wilmanns W. Antitumor activity of pachtaxel after failure of highdose chemotherapy in a patient with late relapse of a nonseminomatous germ cell tumor. AntiCancer Drugs ; : 8.. Williams SD, Einhorn LH, Greco FA et al. VP salvage therapy for refractory germinal neoplasms. Cancer 80; : Wheeler BM, Loehrer PJ, Williams SD et al. Ifosfamide in refractory germ cell tumors. J Clin Oncol 8; : 8.. Motzer RJ, Chou TC, Schwarzt L et al. Paclitaxel in germ cell cancer. Semin Oncol ; :. 0. Reed E, Parker RJ, Dabholkar M. Taxol effect on cisplatindna adduct repair in human ovarian cancer cells. Second National Cancer Institute Workshop on Taxol, Alexandria, VA (Abstr).. Droz JP, Culine S, Biron P, Kramar A. Highdose chemotherapy in germcell tumors. Ann Oncol ; : 00. Received June ; accepted September. Correspondence to: Dr. Jorg Pont. Medizinische Abteilung mit Onkologie Kaiser Franz Josef Spital KundratstraQe AllOOWien Austria