Annals of Oncology 8: 555-559, 1997. © 1997 Kluwer Academic Publishers. Printed in the Netherlands.
Original article Mediastinal non-seminomatous germ cell tumours (MNSGCT) treated with cisplatin-based combination chemotherapy M. Hidalgo,1 L. Paz-Ares,1 F. Rivera,2 P. Lianes,1 G. Huidobro,3 A. Ruiz,1 M. Lopez-Brea,2 J. Sanz-Ortiz,2 J. Lopez Lopez,3 H. Cortes-Funes1 & J. M. Tabernero3 Divisions of Medical Oncology, ^Hospital Universilano '12 de Oclubre', Madrid. 2Hospital Unirersitario 'Marques de Valdecilla', Santander: Hospital Universitario de Sam Pau, Barcelona, Spain
Summary Background: Primary mediastinal non-seminomatous germ cell tumours (MNSGCT) constitute a rare malignancy. This study was performed to review our experience with cispatinbased chemotherapy in patients with MNSGCT. Patients and methods: Patients with MNSGCT treated with cisplatin-based combination chemotherapy between 1978-1995 in three university hospitals in Spain were retrospectively studied. Results: There were 25 males and two females with a median age of 26 years (range 4—71). Fifteen patients had disease confined to the mediastinum and 12 had metastatic disease. All patients were treated with cisplatin chemotherapy regimens (PVB: 7, BEP: 6, and other regimens 12) and considered for residual mass surgery (RMS) when indicated. Eleven patients (40.7%) were rendered disease-free with initial treatment: four with chemotherapy alone, one with surgery plus
adjuvant chemotherapy and six with chemotherapy plus RMS. Three of these patients relapsed at two, six and seven months. The remaining 16 had unfavourable reponses (five partial response, three no change, seven progressive disease and one toxic death) . Eleven patients received salvage treatment but none of them achieved a durable response. After a median follow-up of 77 months (range 1-168), 10 patients remain alive. Actuarial survival at five years is 31.7%. No patients in this series developed a haematological malignancy. Chromosomal analysis showed that 2 out of 10 patients (20%) had a 47XXY karyotype. Conclusions' Only patients who achieved disease-free status are likely to be cured. Therefore, new up-front strategies are needed for the treatment of MNSGCT. Key words: germ cell tumours, extragonadal, mediastinum, cisplatin-based chemotherapy
Introduction
Patients and methods
Extragonadal germ cell tumours (GCT) are rare, accounting for l%-5% of germ cell neoplasms, with the mediastinum the most common primary site [1]. Primary mediastinal non-seminomatous GCT (MNSGCT) usually appear in the third decade of life with a marked male preponderance. Their histology spectrum and serological features are similar to those of testicular primaries. However, the unique association of MNSGCT with haematological malignancies [2] and the Klinefelter syndrome [3] as well as their different natural history and prognosis suggest that they constitute a distinct clinico-pathological entity different from that of their testicular counterparts and therefore require specific treatment [4]. Because of the relatively small number of patients in reported series, the optimal approach to MNSGCT has not been defined. This retrospective study was undertaken to review our experience with malignant MNSGCT treated with modern cisplatin-based combination chemotherapy in three university hospitals in Spain.
We reviewed the medical records of all patients with MNSGCT treated from 1978 to 1995 in the three hospitals. The study criteria for inclusion were: histological diagnosis of non-seminomatous germinal cell cancer; involvement of mediastinal structures; an absence of testicular tumour determined by physical examination and ultrasound and primary treatment with cisplatin-based combination chemotherapy. Patients with poorly differentiated carcinoma of the mediastinum were included if they had elevated serum levels of alfa-fetoprotein (AFP) and/or a beta subunit of human chorionic gonadotrophin (BHCG). All patients were evaluated prior to treatment by medical history, physical exam, complete blood count and serum chemistry, chest X-ray and, since 1980, CTscan of thorax and abdomen, and tumour markers, AFP and BHCG. Patients were treated with a cisplatin combination chemotherapy of adequate platinum dose-intensity: PVB (9), BEP (6), POMB/ACE (4), VIP (6) and BOMP/EPI (2) [4-8]. Patients with partial response and negative tumour markers were scheduled for residual mass surgery (RMS), and those with viable residual cancer received further consolidation chemotherapy. Patients with nonresponding or relapsed tumours and optimal functional status and life expectancy were offered salvage chemotherapy with VIP (4), POMB/ ACE (1), VAB-6 (2) or paclitaxel (3) [7, 9,10], Response criteria were defined as follows: a complete remission (CR) was recorded when serum tumour markers normalised and complete resolution of tumour masses occurred. Partial remission (PR) was defined as a greater than 50% reduction in bidimensional tumour
556 measurements and at least a 90% decline in tumour markers. Patients who could be rendered free of tumour with chemotherapy and RMS were coded ashaving no evidence of disease (NED). Progressive disease was defined as a greater than 25% increase in bidimensional tumour measurements or in tumour markers levels. Stable disease was defined as a less than 50% decrease or a greater than 25% increase in bidimensional tumour measurements or stable tumour marker values Survival was calculated from the day of diagnosis to date of last follow-up or death. The product-limited survival was determined using the method of Kaplan-Meier. Potential relationships between outcome and dichotomous variables were compared with the log-rank test.
Results Twenty-seven patients were identified during the study. Only one patient had an upfront complete surgical resection. Fourteen patients had disease confined to the mediastinum, seven had lung disease and five had extrapulmonary disease (two brain and three liver). Nine patients had negative AFP levels at diagnosis, four patients had levels < 1000 ng/ml, four had values 100010000 ng/ml and six had values > 10000 ng/ml. In four patients AFP values at presentation were not known. BHCG was negative in 14 patients and eight patients had levels below 1000 mUI/ml, in five patients BHCG levels were not measured before treatment. Fifteen patients had LDH values more than 1.5 times the upper normal limit and none had levels over 10 times the normal limit. In seven patients, LDH levels were not recorded. Other patient characteristics are summarised in Table 1. Eleven patients (40.7%) were rendered disease-free with initial treatment. One patient had a complete surgical resection followed by adjuvant PVB. Four patients achieved a CR to chemotherapy alone and six had a PR with negative tumour markers and underwent RMS. There were viable tumour elements in the surgical specimens of three of these patients and they received consolidation chemotherapy including oral VP-16 for one year (one patient), POMB/ACE (one patient) [12] and high-dose chemotherapy (one patient) [12, 13]. Three of 11 patients relapsed at two, six and seven months while all of the others remain disease-free. Sixteen patients (59%) had unfavourable responses to upfront treatment. Five had a PR with persistence of elevated tumour markers, three had NC and seven had progressive disease during treatment. The other patient died of toxic effects of the first cycle of VIP. Salvage therapy with platinum combinations was given to 11 patients. One patient had a PR with viable tumour at RMS. All of the other patients developed tumour progression, after brief periods of stable disease in three of them. Two patients were treated with HDC with peripheral stem cell transplantation, one of them as consolidation of positive RMS with second-line chemotherapy, and the other following partial remission with initial chemotherapy. Both progressed 4 and 10 months after treatment. Paclitaxel, 200 mg/m2 every three weeks was given as second- or further-line to three patients. One patient achieved a PR lasting eight months while
Table 1. Patient characteristics. Age (years) Median Range Gender (M/F)° ECOG 0-2 3^t Histology Yolk Salk tumor Embryonal carcinoma + / - teratoma Choriocarcinoma Poorly differentiated Mixed Non-germ cell elements Royal Marden Hospital stageb IIIMB II1MC IVL1 IVL2 IVL3 IVH IVBr Indiana University Prognostic Group 0 Moderate disease Avanced disease Marker status d Low risk Intermediate risk High risk
26 4-71 25/2 23 4
5 14 2 2 4 2 2 13 2 3 2 3 2 9 18 7 14 6
" M/F: male/female. b IIIMB: mediastinal tumor involvement 2-5 cm in widest dimension; IIIMC as IIIMB but greater than 5 cm. IVL: metastatic lung disease; I < 3 nodules, II > 3 nodules less or equal to 2 cm in diameter, III > 3 and larger than 2 cm in diameter. IVBr' brain metastasis IVL: liver metastasis. c Moderate disease refer to patients with mediastinal involvement less than 50% of thoracic diameter or 5-10 lung nodules per lung field < 3 cm or solitary lung mas, avanced disease implies mediastinal mass > 50% thoracic diameter, > 10 lung nodules per lung field or greater 3 m or brain, bone or liver disease. d Low risk: AFT < 1000 ng/ml and BHCG < 1000 mUI/ml and LDH < 1 5 upper normal limit. Intermediate risk: 1000 ng/ml < AFT < 10000 ng/ml and 1000 mUI/ml < BHCG < 10000 mUI/ ml and 1.5 > LDH < 10 upper normal limit. High risk, any value greater than intermediate risk.
the other two developed disease progression. Table 2 summarises the patients' responses to treatment. After a median follow-up of 77 months, 10 patients (37.04%) are alive, including two with disease. Their actuarial eight-year survival is 33.1% and median survival 13.5 months (95% CI 7.3-19.6). Figure 1 shows the product limit estimation plot of survival. Comparison between outcome and patient baseline characteristics such as histology, tumour markers, extrapulmonary metastatic disease and treatments failed to identify any adverse prognostic characteristics. We performed peripheral blood mononuclear cell chromosomal analysis on 10 patients and two had trisomy 47 XXY (20%). No patient in this series developed a haematological malignancy or other second tumour.
557 Table 2. Treatment outcome.
Discussion
lignant non-germ cell elements, most commonly embryonal rhabdomyosarcoma, was previously described for GCT, and it appears to be much more common for mediastinal primaries [17]. Two patients had poorly differentiated carcinomas with elevated tumour markers. Nowadays, recognition of a specific cytogenetic anomaly i (12p) could help to establish a diagnosis of germ cell tumor in the absence of germ cell elements as determined by conventional pathologic techniques [18]. Twelve of the patients (44%) reported here had metastatic disease, and in five of them it was extrapulmonary. The proportion of patients with metastatic disease at diagnosis varies widely in the literature, ranging from 16% to 85% [14, 15]. This could explain thavourable results obtained by some investigators [15]. There is no doubt that cisplatin-based combination chemotherapy has dramatically changed the evolution of MNSGCT. In the current study, 11 of 27 patients treated with standard-dose cisplatin regimens achieved NED status and 10 are alive with a median follow-up of 77 months. The actuarial eight-year survival was 31.7%. These figures are in the same range as those of published studies (0%-73%) [19, 20], most of which included only a small number of patients (Table 3). Three published series include more than 25 patients. Six of 32 patients (19%) treated at the Memorial Sloan-Kettering Cancer Center achieved CR with chemotherapy alone and six additional patients were rendered NED with adjunctive surgery. The estimated five-year overall survival was 16% [16]. Investigators from Indiana University have reported data on 31 MNSGCT. Of 18 patients (58%) who achieved disease-free status, 11 of them with chemotherapy alone, 15 were alive at a median follow-up of 55 months (13-144). The estimated five-year overall survival was 47% [15].
MNSGCT have well characterised presenting features [14]. Five patients (18%) in our study had yolk sac tumour elements in the pathologic specimen, a figure which is low compared to those in other series [15, 16]. The prognostic implication of this histological variant remains undefined [16]. Two patients had mixed tumours containing rhabdomyosarcoma. The association of ma-
All of the long-term survivors in our series were rendered disease-free with chemotherapy alone or CT followed by aggressive residual mass surgery. We detected no difference in NED rates between the various chemotherapy regimens used, but the number of patients treated with each of them was very small. Taking into account that in large series patients treated with modern regimens did better than those who received lower doses
Primary therapy (27) CR surgery CR chemotherapy NED (necrosis & fibrosis) NED (cancer)
Salvage (11)
1 4 3 3
Total NED
1
11
5 3 7 1
PR positive markers NC PD Toxic death Total unfavourable responses
It 10
16
Abbreviations: CR - complete response; NED - no evidence of disease; PR - partial response; NC - no change; PD - progressive disease.
1,0 ,9 .8 ,7 ,6 ,5 ,4 ,3 .2 ,1 0,0
20
40
60
80
100
120
140
160
18C
Maths Figure I. Overall survival of the whole series.
Table 3. Reported series with ten or more MNSGCT patients treated with cisplatin-based chemotherapy. Author
Patients
Age (median)
Chemotherapy
CR (%)
NED (%)
Follow-up median (m)
Survival
Hematological malignancies
Toner [17] Nichols [16] Current series
32 31 27
25 25 26
6(19) 11 (35) 4(16)
12(38) 18(58) 11(41)
7
16% at 5 year 47% at 5 year 33% at 8 year
0 5(16%) 0
Goss[31] Bukowski [24] Gerl [32] Childs [22]
15 16 12 11
9
7 4(25) 6(50) 2(18)
8(53) 13(81) 8(67) 9(82)
70 81 96
6 alive 10 alive 47% at 7 year 73% at 5 year
1 (7%) 0 0 0
Logothetis [23]
11
7
VAB.VAB/EP, EBC PVB + Dox, BEP PVB, BOMP/EPI, BEP, POMB/ACE VAB-6, EP, PVB, CISCA, BEP PVB/BEP-Dox PVB, PI, ECBC BEP, PVB/BEP, BEV1P, CBOP/BEP PVB, CISCA
4(36)
4
4 alive
0
? 24 29
9
55 77
558
of cisplatin [18, 21], and the studies in high-risk GCT (testicular and extragonadal), the chemotherapy schedule of reference outside a clinical trial is BEP. Some authors suggested a potential role for alternating or rapidly cycling schemes based on their experience in recent years [20-22]. However, these results may also reflect the expertise of experienced centres. In fact, treatment in referral units has been shown to be a favourably independent predictor of survival for GCT patients [23]. About 30%-40% of patients with relapsing or refractory testicular GCT are currently being cured with salvage treatment. None of the patients in our series obtained prolonged remissions with salvage chemotherapy including cisplatin regimens, high-dose chemotherapy or new agents such as taxol. Recently, Saxman et al. reviewed 73 patients with extragonadal NSGCT treated with salvage chemotherapy at Indiana University. Only 2 of 42 patients (5%) were still NED at the time of the report [24]. Similar poor results in the salvage setting have consistently been reported by other groups, including those using high-dose chemotherapy [12, 13]. Primary mediastinal origin has been identified as an adverse prognostic factor in a multivariate analysis of GCT treated with salvage high-dose chemotherapy [25]. The precise prognostic implications of GCT arising in the mediastinum are still unclear. Some authors believe that the unfavourable prognosis of MNSGCT reflects advanced disease at presentation [26]. Others a priori assign these malignancies to poor-risk subgroups, emphasizing that there are intrinsic differences between them and their testicular counterparts, not only in clinical and biological characteristics but also in relation to response to chemotherapy [16]. The MSKCC prognostic model created for patients with testicular primary tumours was not useful for predicting the outcome of those with MNSGCT. Nichols also found no pre-treatment variable to be a good predictor. Similarly, we were not able to establish risk categories in this patient population. Even these three series are the largest in this context, and the relatively small numbers in each minimise the statistical power to detect reliable prognostic indicators. Due to the rarity of this entity, a single centre or group is unlikely to accumulate a large enough database for definition o a prognostic model. Alternatively, multiinstitutional cooperative efforts should be encouraged to approach this goal more realistically. The association between MNSGCT and hematologic malignancies, mainly acute megakaryocytic leukaemia, other acute nonlymphocytic leukemias and malignant histiocytosis [2] is now well established. This is probably a consequence of the multipotential differentiation capacity of the malignant germ cells. In fact, it has recently been shown that leukemic blasts from patients previously diagnosed with MNSGCT also carry the characteristic karyotypic abnormality i (12p) [27]. The blood disorders usually appear shortly (median time: six months) after the mediastinal tumour is diagnosed, and not infrequently at the same time. The exact frequency of devel-
opment of a hematologic neoplasm is unknown, but the absence of any case of it in the present series suggests that its incidence may be lower than that of other cohorts (Table 3). Numerous case reports in the literature point out the association between MNSGCT and the Klinefelter syndrome. In two prospective series from Indiana University and the MD Anderson, 5 of 22 patients (22%) and 4 of 19 patients (19%) had karyotypic evidence of the genetic abnormality. Two of 10 of our patients (20%) were found to have a 47 XXY karyotype. This result confirms that about 20% of MNSGCT patients have the Klinefelter syndrome. In summary, MNSGCT have distinct pathological, clinical and biological features. Despite the therapeutic improvements of the past two decades (cisplatin-based chemotherapy and adjunctive surgery), the prognosis remains unsatisfactory. Only patients who achieved a NED status at first therapeutic attempt are likely to be cured, and future efforts should be directed at identifying them at diagnosis. Patients not likely to do so, or those who relapse, should be candidates for new therapeutic strategies.
References 1. Nichols CR. Mediastinal germ cell tumors clinical features and biologic correlates Chest 1991; 99: 472-9. 2. Nichols CR, Roth BJ, Heerema N et al. Hematologic neoplasia associated with primary mediastinal germ-cell tumors. N Engl J Med 1990; 322: 1425-9. 3. Nichols CR, Heerema NA, Palmer C et al. Klinefelter syndrome associated with mediastinal germ cell neoplasia. J Clin Oncol 1987; 5: 1290-4. 4. Raghavan D, Boyer MJ. Malignant extragonadal germ cell tumours in adults. In Horwich A (ed): Testicular Cancer Philadelphia. Lippincott 1993; 297-317. 5. Wilhans SD, Birch R, Einhorn LH et al. Treatment of disseminated germ cell tumors with cisplatin, bleomycin, and either vinblastin or etoposide. N Engl J Med 1987; 316: 1435-40. 6. Rustin GJS, Newlands EW, Bagshave KD et al. Successful management of metastatic germ cell tumours in the brain. Cancer 1986; 57: 2108-13. 7. Loehrer PJ, Laver R, Roth BJ et al. Salvage therapy in recurrent germ cell cancer: Ifosfamide and cisplatin plus either vinblastin or etoposide. Ann Intern Med 1988; 109: 540-6. 8. Germa-Lluch JR et al. Intensive chemotherapy in poor prognosis non-seminomatous germ cell tumours of the testis. Eur Urol 1992; 21:287-93. 9. Bosl GJ, Gluckman R, Geller NL et al VAB-6: An effective chemotherapy regimen for patients with germ cell tumors. J Clin Oncol 1986; 4: 1493-9. 10. 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 1994; 12: 2277-83. 11. Cooper MA, Einhorn LH. Maintenance chemotherapy with daily oral etoposide following salvage therapy in patients with germ cell tumors. J Clin Oncol 1995; 13. 1167-9. 12. Broun ER, Nichols GR, Kneebone Pet al Long-term outcome of patients with relapsed and refractory germ cell tumors treated with high-dose chemotherapy and autologous bone marrow rescue. Ann Intern Med 1992; 117: 124-8. 13. Motzer RJ, Mazumdar M, Bosl GJ et al. High-dose carboplatin, etoposide, and cyclophosphamide for patients with refractory
559
14 15
16
17
18.
19.
20.
21 22.
23
germ cell tumors. Treatment results and prognostic factors for survival and toxicity. J Clin Oncol 1996; 14: 1098-105. Hainsworth JD, Greco FA. Extragonadal germ cell tumors and unrecognized germ cell tumors Semin Oncol 1992; 19: 119-27. Nichols CR, Saxman S, Willians SD et al. Primary mediastinal nonseminomatous germ cell tumors. A modern single institution experience. Cancer 1991; 65: 1641-6. Toner GC, Geller NL, Lin SYet al. Extragonadal and poor risk nonseminomatous germ cell tumors. Survival and prognostic features. Cancer 1991; 67: 2049-57. Ulbright T, Loehrer P, Roth L et al. The development of nongerm cell malignancies within germ cell tumors. A clinocopathologic study in 11 cases. Cancer 1984; 54: 1824-33. Motzer RJ, Rodriguez E, Reuter V et al. Molecular and cytogenetic studies in the diagnosis of patients with poorly differentiated carcinomas of unknown primary site. J Clin Oncol 1995; 13: 274-82. Kiffer JD, Sandeman TF et al. Primary malignant mediastinal germ cell tumors' A study of eleven cases and renew of the literature. Int J Radiat Oncol Biol Phys 1989; 17. 835-41. Childs WJ, Goldstraw P, Nicholls JE et al. Primary malignant mediastinal germ cell tumours: Improved prognosis with platinum-based chemotherapy and surgery Br J Cancer 1993; 671098-101. Logothetis CJ, Samuels ML, Selig DE et al Chemotherapy of extragonadal germ cell tumors. J Clin Oncol 1986; 3: 316-25. Bukowski RM,Wolf M, Kulander B et al. Alternating combination chemotherapy in patients with extragonadal germ cell tumors. A Southwest oncology group study. Cancer 1993, 71. 2631-8. Harding MJ, Paul J, Gilhs CR et al. Management of malignant teratoma: Does referral to a specialist unit matter? Lancet 1993; 341.999-1002.
24. Saxman SB, Nichols CR, Einhorn LH et al. Salvage chemotherapy in patients with extragonadal nonseminomatous germ cell tumors: The Indiana University experience. J Clin Oncol 1994; 12: 1390-3. 25. Beyer J, Kramar A, Mandanas R et al. High-dose chemotherapy (HDCT) as salvage treatment in germ cell tumors (GCT): A multivariate analysis of prognostic variables. J Clin Oncol 1996; 14: 2638-45. 26. Hainsworth JD, Einhorn LH, Willians SD et al. Advanced extragonadal germ cell tumors. Ann Intern Med 1987; 97: 7-11. 27. Ladanyi M, Samaniego F, Reuter VE et al. Cytogenetic and Inmunohistochemical evidence for the germ cell origen of a subset of acute leukemias associated with mediastinal germ cell tumors. J Natl Cancer Inst 1990; 82: 221-7. 28. Gerl A, Clemm C, Lamerz R et al. Cisplatin-based chemotherapy of primary extragonadal germ cell tumors. A single institution experience. Cancer 1996; 77: 526-32. 29. Goss PE, Schwertfeger L, Blackstein M et al. Extragonadal germ cell tumors. A 14 year Toronto experience. Cancer 1994; 73. 1971-9. Received 14 February 1997; accepted 15 April 1997.
Correspondence to • Luis Paz-Ares, MD, PhD Servicio de Oncologia Medica Hospital Univ '12 de Octubre' Avd Cordoba Km 5.4. 28041 Madrid Spain