|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
Originally published as JCO Early Release 10.1200/JCO.2004.04.008 on August 9 2004 © 2004 American Society of Clinical Oncology. Early Predicted Time to Normalization of Tumor Markers Predicts Outcome in Poor-Prognosis Nonseminomatous Germ Cell TumorsFrom the Genito-Urinary Group of the French Federation of Cancer Centers, Paris, France; and the Department of Genitourinary Medical Oncology, The University of Texas M.D. Anderson Cancer Center, Houston, TX. Address reprint requests to Karim Fizazi, MD, PhD, Department of Medicine, Institut Gustave Roussy, 39 rue Camille Desmoulins, 94800 Villejuif, France; e-mail: fizazi{at}igr.fr
PURPOSE: The prognostic relevance of the rate of decline of serum alpha-fetoprotein (AFP) and human chorionic gonadotropin (HCG) during the first 3 weeks of chemotherapy for nonseminomatous germ cell tumors (NSGCT) was studied in the context of the International Germ Cell Cancer Collaborative Group (IGCCCG) classification. PATIENTS AND METHODS: Data from 653 patients prospectively recruited in clinical trials were studied. Tumor markers were obtained before chemotherapy and 3 weeks later. Decline rates were calculated using a logarithmic formula and expressed as a predicted time to normalization (TTN). A favorable TTN was defined when both AFP and HCG had a favorable decline rate, including cases with normal values. RESULTS: The median follow-up was 50 months (range, 2 to 151 months). Tumor decline rate expressed as a predicted TTN was associated with both progression-free survival (PFS; P < .0001) and overall survival (OS; P < .0001). The 4-year PFS rates were 64% and 38% in patients from the poor-prognosis group who had a favorable and an unfavorable TTN, respectively. The 4-year OS rates were 83% and 58%, respectively. This effect was independent from the initial tumor marker values, the primary tumor site, and the presence of nonpulmonary visceral metastases: tumor marker decline rate remained a strong predictor for both PFS (hazard ratio = 2.5; P = .01) and OS (hazard ratio = 4.6; P = .002) in patients from the IGCCCG poor-prognosis group in multivariate analysis. CONCLUSION: Early predicted time to tumor marker normalization is an independent prognostic factor in patients with poor-prognosis NSGCT and may be a useful tool in the therapeutic management of these patients.
The integration of cisplatin-based chemotherapy and surgery in the first-line treatment of disseminated nonseminomatous germ cell tumors (NSGCT) has resulted in cure rates of approximately 80% during the past two decades.1,2 Since the 1980s, improved treatment has led to the need to stratify patients by prognosis. Because differences between the prognostic classification systems make it difficult to compare results of clinical trials, the International Germ Cell Cancer Consensus Group (IGCCCG) published in 1997 a consensus prognostic index for NSGCT.3 This index stratifies patients into good-, intermediate-, and poor-prognosis subgroups on the basis of three criteria: the primary tumor site, the levels of serum tumor markers, and whether extra-pulmonary visceral metastases are present. This index has been validated by the IGCCCG3 and by other groups.4-7 Patients allocated to the good-prognosis group have a probability of cure of greater than 90%, but patients in the intermediate-prognosis and poor-prognosis groups have 5-year survival rates of only 80% and 48%, respectively.3 In the last two decades, one of the most important challenges for oncologists have been to attempt to reduce toxicity in the good-risk group and to improve efficacy in the intermediate- and poor-risk groups.1,2,8,9 Therefore, it is of great importance to achieve better and earlier selection of patients who are deemed to be at risk of experiencing treatment failure with standard treatment in order to switch them to a more effective (although often more toxic) treatment. Moreover, because the amount of toxic effects in survivors of disseminated NSGCT may be important,8 especially in those receiving more than three cycles of the standard bleomycin, etoposide, and cisplatin (BEP) regimen, it is also important to identify earlier patients who are likely to achieve a cure with standard treatment among those who are initially classified into the intermediate- or the poor-risk groups, so that useless toxic effects from more intensive treatment are avoided. Serum levels of human chorionic gonadotropin (HCG), alpha-fetoprotein (AFP), or both are elevated in approximately 80% of patients with disseminated NSGCTs,10 and their levels as assessed before chemotherapy have been integrated into the IGCCCG classification.3 Tumor marker normalization is one of the primary objectives to reach in patients receiving primary chemotherapy, and the lack of normalization is often associated with an incomplete response. Therefore, many authors have attempted to correlate tumor marker half-life with complete response and/or survival.11-17 An association between prolonged HCG or AFP half-lives and outcome was found by some authors7,11,12,14,16 but was not confirmed by others.13,15,17 Limitations of some of these studies include the lack of power and absence of multivariate analysis, a variety of methods being used for the calculations of tumor marker decline rates, and tumor marker determinations performed relatively late after the start of chemotherapy (which may allow only a late treatment switch). Moreover, a transient elevation of tumor marker in the very first weeks after the initiation of chemotherapy (the so-called tumor marker surge) has been described in approximately 25% of patients.18,19 Therefore, we assessed whether the decline rate of tumor markers that is measured rather early in the course of chemotherapy and expressed as a predicted time to tumor marker normalization may be correlated with progression and survival in the context of the IGCCCG classification.
Eligibility Criteria Patients included in prospective clinical trials performed from 1987 to 2000 by the Genito-Urinary Group of the French Federation of Cancer Centers20,21 and The University of Texas M.D. Anderson Cancer Center6,22-24 were included. Patients from France were treated in two randomized trials: one comparing three cycles of BEP with four cycles of cisplatin and etoposide in good-prognosis patients20 and one comparing BEP with cisplatin, doxorubicin, cyclophosphamide, bleomycin, and vinblastine (CISCA/VB) in poor-prognosis patients.21 From 1987 to 1993, at The University of Texas M.D. Anderson Cancer Center, patients with a good prognosis were included in a phase III trial comparing two level of doses of the CISCA/VB regimen.6 Patients treated since 1993 were included in phase II trials according to their prognosis: a carboplatin, etoposide, and bleomycin trial for good-risk patients,24 a CISCA/VB trial for intermediate-risk patients,23 and a trial using a dose-dense regimen combining bleomycin, vincristine, cisplatin, doxorubicin, cyclophosphamide, methotrexate, vincristine, dactinomycin, and etoposide in poor-prognosis patients.22 Eligibility criteria were as follows: evidence of a disseminated NSGCT; conventional-dose platinum-based chemotherapy; baseline HCG, AFP, and lactate dehydrogenase (LDH) measurements available; and HCG and AFP values available after one cycle of chemotherapy (theoretical day 21 value). Baseline tumor markers were typically obtained in the week before chemotherapy was initiated. Theoretical day 21 value was actually obtained between days 18 and 28 after the initiation of chemotherapy. Institutional normal values for AFP and HCG did not vary much, so raw values were used. This was not the case for LDH, which was expressed as a function of the normal value. Patients with a normal HCG or AFP value at baseline were eligible for this study. The total number of patients was 701, of whom 653 patients were eligible (48 patients were not eligible because of insufficient data).
IGCCCG Classification
Tumor Marker Decline: Predicted Time to Normalization The decline rate was calculated for each patient with an abnormal marker value at baseline and each tumor marker M (HCG or AFP) using only two values: the baseline value (M0) and the value obtained after one cycle of chemotherapy (day 21 value; M1). The decline rate was calculated using a logarithmic transformation, and it was expressed as a theoretical number of weeks necessary to normalization that was called predicted time to normalization (TTN): TTNM = 3 a/b, with a = log(M0) – log(MN), b = log(M0) – log(M1), and MN, the normal value of each tumor marker M.
Results were then classified into four categories: AM, normal tumor marker value at M0 and M1; BM, elevated M0 and normal M1; CM, elevated M0 and TTNM less than TM; DM, elevated M0 and TTNM The median TTNs in the data set were 7.6 weeks and 5.7 weeks for AFP and HCG, respectively. Cutoff points were set at TAFP = 9 weeks for AFP and at THCG = 6 weeks for HCG, which correspond to the normalization of TM after three and two cycles of the standard BEP regimen, respectively. These estimates have a pragmatic and practical interpretation, because they imply that if markers are predicted to still be abnormal after three cycles of BEP for HCG and after two cycles of BEP for AFP, then the chemotherapy protocol may not be well adapted and should be changed as early as possible. These choices better protect against a false-positive result. For example, using MN = 10 (log10[10] = 1) as the normal value for HCG, a patient with an initial value of 1,000 IU/mL (log10[1,000] = 3) who decreases to 100 IU/mL (log10[100] = 2) at 3 weeks has an estimated time to normalization of TTNM = 3* (3 – 1)/(3 – 2) = 6 weeks. The corresponding estimate for half-life is equal to HLM = log(2)/slope, with the slope equal to b/3, so we can express HLM as 3*log(2)/b. Thus there is a direct relationship between TTN and half-life, because TTNM = HLM a/log(2). We can see that TTNM takes into account half-life as well as the distance the initial marker is away from the normal value. A marker decline rate is considered favorable when both HCG and AFP fall into either the A, B, or C categories. A marker decline is considered unfavorable when either HCG or AFP or both fall into the D category. Similar definitions are considered for a half-life greater than 7 days for AFP and greater than 3.5 days for HCG.
Postchemotherapy Outcome
Statistical Analysis
The median follow-up time was 50 months (range, 2 to 151 months). Overall, 81 patients died and 155 patients either experienced relapse or died. The most important prognostic factor was the IGCCCG classification, with 4-year OS rates of 95%, 86%, and 64%, respectively in the good-, intermediate-, and poor-risk groups (P < .0001). The corresponding PFS rates were 89%, 74%, and 44%, respectively (Table 2). Initial marker levels were significant for both PFS and OS, but this is expected because they are included in the definition of IGCCCG.
Half-Life The 193 patients (30%) with an estimated half-life of greater than 3.5 days for HCG (unsatisfactory half-life) had a 63% 4-year PFS rate, as compared with 86% and 61% for patients with normal HCG at 3 weeks and a half-life of less than 3.5 days, respectively (P < .001). However, this difference was essentially due to difference between a normal and an abnormal HCG value at week 3. No difference in PFS (P = .074) or OS (P = .204) was observed for patients with a satisfactory or unsatisfactory AFP half-life (Table 2).
Predicted TTN
Multivariate Analyses
Patients with an unfavorable TTN for either AFP or HCG had a 4-year PFS rate of 61%, as compared with 91% and 79% for patients with both AFP and HCG normal at day 21 and those with AFP and/or HCG with a short predicted TTN, respectively (P = .036; Table 3). A similar approach was examined for half-life, and the results were similar. The 4-year PFS rates were 86% for patients with a satisfactory marker half-life as compared with 69% for patients with an unsatisfactory marker half-life (P = .043; Table 3). The 4-year OS rate was 94% for patients with a favorable predicted TTN as compared with 74% for patients with an unfavorable TTN (P = .005; Cox model stratified on IGCCCG status and adjusted for baseline values of AFP and HCG; Table 3). The 4-year OS rate was 93% for patients with a satisfactory marker half-life as compared with 82% for patients with an unsatisfactory marker half-life, but the difference was not statistically significant (P = .131). Predicted TTN was then examined in the context of the IGCCCG classification in a multivariate analysis to evaluate the prognostic relevance of TTN adjusted for initial tumor marker values, primary site, and nonpulmonary metastasis disease, because these variables are part of the IGCCCG classification (Table 4). The results show that the predicted TTN remained a strong predictor for both PFS (hazard ratio = 2.5) and OS (hazard ratio = 4.6) in patients from the IGCCCG poor-prognosis group, even when initial AFP and HCG values, mediastinal primary site, and nonpulmonary metastatic disease were taken into account. These results were confirmed when AFP and HCG values before chemotherapy were studied using a continuous scale (hazard ratio = 2.4 and 3.8 for PFS and OS, respectively).
PFS and OS curves according to predicted TTN in each IGCCCG group are shown in Figures 1 and 2. The 4-year PFS rates were 64% and 38% in patients from the IGCCCG poor-prognosis group who had a favorable and an unfavorable predicted TTN, respectively (P = .007). The 4-year OS rates were 83% and 58%, respectively (P = .009). The 4-year PFS rates were 89% and 87% in the IGCCCG good-risk group and 73% and 76% in the intermediate-risk groups for patients with favorable versus unfavorable predicted TTN, respectively. Results were similar for 4-year OS rates, with 96% and 91% (good risk) and 85% and 86% (intermediate risk), respectively.
Figure 3 shows an abacus designed as an aid for the classification of patients according to predicted TTN. Favorable predicted TTN occurs below each diagonal line. Unfavorable predicted TTN occurs above each diagonal line. For example, at 3 weeks from baseline, an initial HCG value of 5,000 U/L needs to decrease to less than 224 U/L, and an initial AFP value of 5,000 ng/mL needs to decrease to less than 630 ng/mL in order to qualify for a favorable predicted TTN. For baseline AFP values of 10,000 ng/mL and 80,000 ng/mL, unfavorable predicted TTN will be obtained for 3-week AFP values greater than 1,000 ng/mL and 4,000 ng/mL, respectively. For baseline HCG values of 4,000 U/L, 9,000 U/L, and 100,000 U/L, unfavorable predicted TTN will be obtained for 3-week HCG values greater than 200 U/L, 300 U/L, and 1,000 U/L, respectively. For other values, the formula presented in Patients and Methods will need to be programmed into a hand calculator or a spreadsheet computer program.
This study was based on a multi-institution experience of 653 patients, and we show that the decline rate of HCG and AFP as assessed by the predicted TTN has a prognostic impact on both PFS and OS, especially in patients with a poor-risk metastatic NSGCT according to the IGCCCG classification. In this category of patients, the 4-year OS rates were 83% and 58% for patients with a favorable and unfavorable TTN, respectively, thus providing the opportunity to intensify treatment in patients with an unfavorable tumor marker decline. In contrast, patients with a favorable decline may be kept on standard treatment, thus avoiding unnecessary toxic side effects. The prognostic impact was independent of the tumor marker levels at the start of chemotherapy in multivariate analysis. Tumor marker decline can be assessed visually on an abacus or using software. Recently, authors from the Memorial Sloan-Kettering Cancer Center also reported that tumor marker decline predicted complete response, PFS, and OS in a series of 415 patients.7 However, the method used in the Memorial Sloan-Kettering study requires multiple tumor marker monitoring. This is associated with two disadvantages: first, patients may not be compliant to have their tumor markers checked between cycles, so it would be difficult to calculate the tumor marker decline according to this method. Second, they assessed tumor marker decline on samples obtained from day 7 to day 56 after the start of chemotherapy, which allows classification of patients with either a satisfactory or an unsatisfactory decline late during the course of chemotherapy, after two to three cycles. In contrast, the method used in our study does not require multiple tumor marker monitoring. It allows the clinician to assess the prognosis as early as 3 weeks after the start of chemotherapy, that is, before starting the second chemotherapy cycle. However, to apply these methods, the baseline tumor marker values and those at 3 weeks should be as close as possible before the initiation of each chemotherapy cycle. Because it is our goal to use pragmatically the concept of tumor marker decline for better directed treatment of patients with a poor-risk NSGCT according to the IGCCCG, we included all unpretreated patients in this study, including those with normal tumor markers at baseline and those who achieved an early normalization after chemotherapy. Although this has been a subject of debate,7,13,15 we feel, like others,7 that the more patients can benefit from this tool, the better. Other groups have suggested that the IGCCCG classification might be improved to better assess the prognosis of individuals with advanced NSGCT. A classification and regression tree model analysis model was applied to a retrospective series of 332 patients with poor-prognosis disease according to the IGCCCG classification.28 This study suggested that different prognostic subsets could be identified among this category of patients. However, this technique entails a complex subclassification, and whether it could be practically useful in the clinical management of these patients requires validation. A study of 669 patients with advanced NSGCT and reported by the European Organization for Research and Treatment of Cancer also showed that an AFP surge is an independent predictive factor of progression, although it does not add to prediction of survival once the IGCCCG risk group is known.19 The analysis of this peak will have a much lesser impact on the predicted TTN than it would on the estimation of half-life. Predicting earlier response to therapy has emerged as a major challenge in oncology in recent years. Tumor markers are now used routinely in a number of neoplasms to predict the efficacy of treatment. For example, a decrease in prostate-specific antigen is used to predict efficacy after radical prostatectomy for localized disease as well as after chemotherapy in advanced disease.29 Early normalization of neuron-specific enolase is also an independent predictor of both a complete response and survival in small-cell lung cancer.30 Early decline of CA-125 also predicts efficacy of chemotherapy in ovarian cancer.31,32 That our study and the Memorial Sloan-Kettering Cancer Center study show that the AFP and HCG decline rate is an independent prognostic factor of outcome in patients with poor-risk advanced NSGCT shows that this parameter should also be taken into account in the clinic. Indeed, we have recently initiated a multi-institution international clinical trial in which patients with a poor-risk NSGCT according to the IGCCCG receive a first cycle of BEP. The tumor marker decline rate is assessed at day 21 using the method reported here: patients with a favorable decline receive three additional cycles of BEP (and the trial should prospectively confirm that a cure rate of greater than 80% can be achieved in this category). In contrast, patients with an unfavorable decline are randomly assigned to receive either standard treatment (three additional BEP) or a more intensive regimen that includes new drugs that have been selected based on their efficacy in the refractory setting. In conclusion, this study provides evidence that the early assessment of tumor marker decline as assessed by the predicted TTN method is a prognostic indicator of both PFS and OS in patients with advanced NSGCT and a poor-prognosis classification according to the IGCCCG. Statistical significance was retained after adjustment to initial tumor marker levels, mediastinal primary, and the presence of nonpulmonary visceral metastases.
The following individuals also contributed to this work: C. Beaumont-Raymond, P. Beuzeboc, P. Biron, B.N. Bui, A. Caty, C. Chevreau, R. Delva, P. Fargeot, F. Fruge, L. Geoffrois, A. Goupil, Z. Haddad, Hauteville, J.F. Héron, P. Kerbrat, F. Kohser, C. Linassier, J.P. Malhaire, C. Martin, Y. Merrouche, N. Mottet, M. Mousseau, J. Pény, C. Perez, C. Platini, F. Rolland, C. Théodore, A. Thyss, S. Walter, and T.N. Wiltz-Everett.
The authors indicated no potential conflicts of interest.
Presented in part at the 38th Annual Meeting of the American Society of Clinical Oncology, Orlando, FL, May 18-21, 2002. Authors’ disclosures of potential conflicts of interest are found at the end of this article.
1. Bosl GJ, Motzer RJ: Testicular germ-cell cancer. N Engl J Med 337:242-253, 1997 2. Culine S, Kramar A, Biron P, et al: Chemotherapy in adult germ cell tumors. Crit Rev Oncol Hematol 22:229-263, 1996[Medline]
3. International Germ Cell Consensus Classification: A prognostic factor-based staging system for metastatic germ cell cancers. J Clin Oncol 15:594-603, 1997
4. 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 8:477-483, 1997 5. Hinton SW, Catalano PJ, Einhorn LW, et al: Re-analysis of the intergroup trial (E3887) of cisplatin and etoposide and either bleomycin or ifosfamide in the treatment of disseminated germ cell tumors using the International Germ Cell Consensus Classification. Proc Am Soc Clin Oncol 19:341a, 2000 (abstr 1342)
6. Fizazi K, Do KA, Wang X, et al: A 20% dose reduction of the original CISCA/VB regimen allows better tolerance and similar survival rate in disseminated testicular non-seminomatous germ-cell tumors: Final results of a phase III randomized trial. Ann Oncol 13:125-134, 2002
7. Mazumdar M, Bajorin DF, Bacik J, et al: Predicting outcome to chemotherapy in patients with germ cell tumors: The value of the rate of decline of human chorionic gonadotrophin and alpha-fetoprotein during therapy. J Clin Oncol 19:2534-2541, 2001
8. Fizazi K, Chen I, Logothetis CJ: Germ cell tumor survivors: The price for cure. Ann Oncol 13:187-189, 2002 9. Culine S, Droz JP: Comparative tolerability of chemotherapy regimens for germ cell cancer. Drug Saf 22:373-388, 2000[CrossRef][Medline] 10. Bosl GJ: Circulating tumor markers: Biologic markers of malignancies, in MacDonald JS, Haller DG, Mayer RJ (eds): Manual of Oncologic Therapeutics (ed 3). Philadelphia, PA, Lippincott, 1995, pp 49-54 11. Picozzi VJ Jr, Freiha FS, Hannigan JF, et al: Prognostic significance of a decline in serum human chorionic gonadotropin levels after initial chemotherapy for advanced germ-cell carcinoma. Ann Intern Med 100:183-186, 1984
12. Toner GC, Geller NL, Tan C, et al: Serum tumor marker half-life during chemotherapy allows early prediction of complete response and survival in nonseminomatous germ cell tumors. Cancer Res 50:5904-5910, 1990
13. Stevens MJ, Norman AR, Dearnaley DP, et al: Prognostic significance of early serum tumor marker half-life in metastatic testicular teratoma. J Clin Oncol 13:87-92, 1995 14. Gerl A, Lamerz R, Clemm C, et al: Does serum tumor marker half-life complement pretreatment risk stratification in metastatic nonseminomatous germ cell tumors? Clin Cancer Res 2:1565-1570, 1996[Abstract] 15. De Wit R, Sylvester R, Tsitsa C, et al: Tumour marker concentration at the start of chemotherapy is a stronger predictor of treatment failure than marker half-life: A study in patients with disseminated non-seminomatous testicular cancer. Br J Cancer 75:432-435, 1997[Medline] 16. Gerl A, Lamerz R, Mann K, et al: Is serum tumor marker half-life a guide to prognosis in metastatic nonseminomatous germ cell tumors? Anticancer Res 17:3047-3049, 1997[Medline] 17. Inanc SE, Meral R, Darendeliler E, et al: Prognostic significance of marker half-life during chemotherapy in non-seminomatous germ cell testicular tumors. Acta Oncol 38:505-509, 1999[CrossRef][Medline] 18. Horwich A, Pecham MJ: Transient marker elevation following chemotherapy for germ-cell tumours of the testis. Cancer Treat Rep 70:1329-1339, 1986[Medline] 19. De Wit R, Collette L, Sylvester R, et al: Serum alpha-fetoprotein surge after the initiation of chemotherapy for non-seminomatous testicular cancer has an adverse prognostic significance. Br J Cancer 78:1350-1355, 1998[Medline] 20. Culine S, Kerbrat P, Bouzy J, et al: Are 3 cycles of bleomycin, etoposide and cisplatin (3 BEP) or 4 cycles of etoposide and cisplatin (4 EP) equivalent regimens for patients (pts) with good-risk metastatic non seminomatous germ cell tumors (NSGCT)? Preliminary results of a randomized trial. Proc Am Soc Clin Oncol 18:309a, 1999 (abstr 1188) 21. Droz JP, Culine S, Bouzy J, et al: Preliminary results of a randomized trial comparing bleomycin, etoposide, cisplatin (BEP) and cyclophosphamide, doxorubicin, cisplatin/vinblastine, bleomycin (CISCA/VB) for patients (pts) with intermediate- and poor-risk metastatic non seminomatous germ-cell tumors (NSGCT). Proc Am Soc Clin Oncol 20:173a, 2001 (abstr 690) 22. Fizazi K, Do KA, Wang X, et al: Alternating dose-dense chemotherapy in patients with high volume disseminated non-seminomatous germ cell tumors. Br J Cancer 86:1555-1560, 2002[CrossRef][Medline] 23. Daliani DD, Amato RJ, Jones DM, et al: Cisplatin, cyclophosphamide, adriamycin alternating with vinblastine, bleomycin (CISCA/VB) combination chemotherapy (CRx) for the treatment of moderate volume (MV) metastatic non-seminomatous germ cell tumors (NSGCT). Proc Am Soc Clin Oncol 16:332a, 1997 (abstr 1184) 24. Prow DM, Amato RJ, Jones DM, et al: Carboplatin, etoposide and bleomycin (CEB) combination chemotherapy for the treatment of small volume (SV) metastatic non-seminomatous germ cell tumor of the testis (NSGCT). Proc Am Soc Clin Oncol 16:340a, 1997 (abstr 1217)
25. Fizazi K, Tjulandin S, Salvioni R, et al: Viable malignant cells after primary chemotherapy for disseminated nonseminomatous germ cell tumors: Prognostic factors and role of postsurgery chemotherapy—Results from an international study group. J Clin Oncol 19:2647-2657, 2001 26. Logothetis CJ, Samuels ML, Trinidade A, et al: The growing teratoma syndrome. Cancer 50:1629-1635, 1982[CrossRef][Medline] 27. André F, Fizazi K, Culine S, et al: The growing teratoma syndrome: Results of therapy and long-term follow-up of 33 patients. Eur J Cancer 36:1389-1394, 2000
28. Kollmannsberger C, Nichols C, Meisner C, et al: Identification of prognostic subgroups among patients with metastatic IGCCCG poor-prognosis germ-cell cancer: An explorative analysis using cart modeling. Ann Oncol 11:1115-1120, 2000
29. Bubley GJ, Carducci M, Dahut W, et al: Eligibility and response guideline for phase II clinical trials in androgen-independent prostate cancer: Recommendations from the prostate-specific antigen working group. J Clin Oncol 17:3461-3467, 1999 30. Fizazi K, Cojean I, Pignon JP, et al: Normal serum neuron-specific enolase (NSE) value after the first cycle of chemotherapy: An early predictor of complete response and survival in small-cell lung cancer. Cancer 82:1049-1055, 1998[CrossRef][Medline]
31. Rustin GJ, Nelstrop AE, McClean P, et al: Defining response of ovarian carcinoma to initial chemotherapy according to serum CA 125. J Clin Oncol 14:1545-1551, 1996
32. Rustin GJS, Nelstrop AE, Bentzen SM, et al: Selection of active drugs for ovarian cancer based on CA-125 and standard response rates in phase II trials. J Clin Oncol 18:1733-1739, 2000 Submitted April 1, 2003; accepted April 20, 2004.
Related Editorial
This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2004 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
TOP
ABSTRACT
INTRODUCTION
TM or elevated M0 and increased value at day 21. 
.05 was considered statistically significant. The relative influence of different prognostic factors on survival was estimated with Cox’s proportional hazards model. Overall survival (OS) was calculated from the date of enrollment to the last date of follow-up. Progression-free survival (PFS) was calculated from the date of inclusion into the trial to the first date of progression, relapse, death, or last follow-up for patients alive who never experienced disease progression or relapse. For univariate analyses, initial AFP (normal, < 1,000, < 10,000, and 
