Advanced seminoma—Treatment results and prognostic factors for survival after first-line, cisplatin-based chemotherapy and for patients with recurrent disease

A single-institution experience in 145 patients

Abstract

BACKGROUND

Advanced seminoma is a rare clinicopathologic entity. To the authors' knowledge, very few sizeable reports published to date have studied the outcome of patients with advanced seminoma after first-line and salvage therapy, and few have dealt with prognostic factors initially or in patients with recurrent disease.

METHODS

The records of 145 men with advanced seminoma who were treated with cisplatin-based first-line chemotherapy regimens were reviewed. Six patient characteristics, including age, prior radiotherapy, primary tumor site, initial serum lactate dehydrogenase and human chorionic gonadotropin levels, and disease stage, were studied as initial prognostic factors. In patients with recurrent disease, outcome according to the site of recurrence and the salvage treatment was also reviewed.

RESULTS

A complete response was obtained in 130 patients (90%) after cisplatin-based first-line chemotherapy, and the 5-year overall survival rate was 81% (95% confidence interval [95% CI], 73–87%). Nonpulmonary visceral metastasis at diagnosis was the only initial adverse prognostic factor. Thirty-one patients (21%) developed recurrent disease. Recurrence in the liver or the central nervous system was a major adverse prognostic factor, with a 5-year overall survival rate of 7% (95% CI, 1–32%), compared with 58% (95% CI, 33–79%) in patients who had lymph node, lung, or bone recurrences. The only durable complete remission after a liver recurrence was obtained with high-dose chemotherapy followed by autologous stem cell transplantation. All 12 patients who were treated for primary mediastinal seminoma with cisplatin-based chemotherapy alone were long-term disease free survivors.

CONCLUSIONS

Overall, the prognosis of patients with advanced seminoma was good after cisplatin-based, first-line chemotherapy. Metastasis in the liver or the central nervous system, initially or at recurrence, is currently the only proven adverse prognostic factor. Cancer 2003;98:745–52. © 2003 American Cancer Society.

DOI 10.1002/cncr.11574

Pure seminomas are a well defined clinicopathologic entity accounting for approximately 50% of all testicular germ cell tumors. Eighty percent of all patients with pure seminomas present initially with clinical Stage I disease (testis involvement only) or with nonbulky Stage II disease (retroperitoneal lymphadenopathy measuring < 5 cm in greatest dimension), and > 90% of these patients are cured after undergoing radical inguinal orchiectomy followed by retroperitoneal and ipsilateral pelvic lymph node radiation therapy.1–3

Advanced seminoma is defined as either a testicular seminoma associated with retroperitoneal lymphadenopathy measuring > 5 cm in greatest dimension; or supradiaphragmatic lymphadenopathy; or a visceral metastasis; or primary extragonadal, mediastinal, or retroperitoneal disease. Radiation therapy cures only 50–65% of these patients,4–6 and cisplatin-based chemotherapy regimens currently are the standard first-line treatment for patients with advanced seminoma, with a complete response rate of 70–90%.7–12 However, a minority of patients with advanced seminoma are either refractory to or will develop recurrent disease after first-line, cisplatin-based chemotherapy.10, 12 Given the low incidence of advanced seminoma in general, especially disease requiring salvage therapy, very few studies have shed light on issues such as adverse prognostic factors, initially and at recurrence, or have provided an unequivocal percentage of patients with advanced seminoma who achieved a durable complete response after first-line cisplatin-based chemotherapy. We report our experience with 145 patients who were treated for advanced seminoma in a single institution. Pretreatment prognostic factors were investigated to determine a more individualized and risk-based treatment. In patients who developed recurrent disease, outcomes according to the site of recurrence and the salvage treatment were also reviewed.

MATERIALS AND METHODS

Patients

From January, 1980 to December, 1999, 145 patients received cisplatin-based regimens as first-line chemotherapy for their advanced seminoma at the Institut Gustave Roussy. Patients were identified retrospectively from the institution's computerized data base. Results concerning 60 patients were reported previously by Culine et al.,12 and these data were reviewed with an updated follow-up. Ten patients in this series were treated on an Eastern Organization for Research and Treatment of Cancer prospective trial (Trial 30874) with the approval of the Institutional Review Board.13

All patients had a histologically proven pure seminoma as well as normal serum α-fetoprotein levels throughout the course of their disease.

Advanced seminoma was defined as Stage IIC (testicular seminoma with associated retroperitoneal lymphadenopathy measuring > 5 cm in greatest dimension), Stage IIIA/B (testicular seminoma with pulmonary metastasis and/or supradiaphragmatic lymphadenopathy), or Stage IIIC (testicular seminoma with nonpulmonary visceral metastasis) according to the American Joint Committee on Cancer-International Union Against Cancer 1997 classification system. Patients who were treated for primary extragonadal, mediastinal, or retroperitoneal seminomas also were included in this study.

First-line chemotherapy regimens have been described previously12 and include 4 cycles of etoposide 500 Mg/m² per cycle and cisplatin 100 mg/m² (EP) in 57 patients, 3 cycles of bleomycin, etoposide and cisplatin (BEP) in 29 patients, 4–6 cycles of cyclophosphamide, vinblastine, bleomycin, dactinomycin, and ciplatin (VAB-6) in 26 patients, 4 cycles of ifosfamide, vincristine, and cisplatin (HOP) in 17 patients and other cisplatin based chemotherapy combinations in 16 patients.

Response to first-line chemotherapy was evaluated either by surgery or by computed tomography (CT) scan. Surgery usually was performed for a persistent, stable, residual mass with a greatest tumor dimension > 3 cm on a follow-up CT scan scheduled 6 months after completion of chemotherapy. Since 1999, a dedicated positron emission tomography scan also has been used to evaluate response in 10 patients.

A complete response to chemotherapy alone was defined as the disappearance of all clinical, radiographic, and biochemical evidence of disease for a minimum of 4 weeks after the last cycle of induction chemotherapy. Surgical resection of necrotic tissue or fibrosis without pathologic evidence of viable seminoma also was considered a complete response to chemotherapy alone. A complete response to chemotherapy plus surgery was defined as the total excision of all residual masses with pathologic evidence of a viable seminoma in these masses. Stable or regressing residual masses with a greatest tumor dimension < 3 cm with normal serum tumor markers and no radiographic or clinical signs of progression after at least a 2-year follow-up were considered a favorable partial response to first-line chemotherapy. Seminoma that progressed within 1 month after the end of the last cycle of induction chemotherapy was considered refractory.

Survival comparisons were performed using 6 pretreatment patient characteristics: age at diagnosis (younger than 30 years vs. 30 years and older), prior treatment with radiotherapy, primary tumor site (gonadal vs. extragonadal), pretreatment serum lactate dehydrogenase (LDH) and human chorionic gonadotropin (HCG) levels, and disease stage according to the TNM classification. Evidence of recurrence was based on rising serum tumor markers associated with a growing mass on radiologic studies. Patients with histologic confirmation of viable seminoma on biopsy specimens of an incompletely excised residual mass also were candidates for salvage therapy.

In patients with recurrent seminoma, salvage treatment was comprised of either 4 cycles of chemotherapy with vinblastine, ifosfamide, cisplatin (VeIP14) in 12 patients; 3 cycles of VeIP followed by high-dose chemotherapy (HDCT) and autologous stem cell transplantation in 6 patients; and other treatment modalities (radiotherapy and/or surgery and/or chemotherapy) in 12 patients. High-dose combination chemotherapy regimens have been described previously and was comprised of carboplatin, etoposide, and cyclophosphamide in four patients; thiotepa and busulfan in one patient; and double intensification with the ifosfamide, carboplatin, and etoposide (ICE) regimen in one patient.15–17 At the end of salvage treatment, all patients had been followed regularly, and radiographic examinations and serum tumor marker determinations had been performed every 2 months during the first 2 years and at gradually increasing intervals thereafter. Survival comparisons according to both the site of recurrence and the salvage treatment were also performed.

Statistical Methods

Survival curves were generated using the Kaplan–Meier method. The duration of remission and survival were calculated from the first day of chemotherapy until recurrence, death, or the last follow-up.

A univariate analysis for comparisons between groups was conducted using the log-rank test. A multivariate analysis also was performed using a Cox proportional hazards regression model with a forward, stepwise variable selection procedure.

RESULTS

Patient Characteristics

Patient characteristics are listed in Table 1. All 145 patients were chemotherapy-naïve prior to the administration of cisplatin-based chemotherapy regimens. Forty-two patients had received prior radiotherapy for early-stage seminoma that encompassed the retroperitoneal lymph nodes in 38 patients and both the mediastinum and the retroperitoneal lymph nodes in 4 patients. The primary tumor site was the testis in 121 patients (83.4%) and the mediastinum in 14 patients, and the retroperitoneum in 10 patients. All patients with initial Stage IIIC disease had liver and/or central nervous system (CNS) metastases.

Table 1. Patient Characteristics

Characteristic	No. of patients	%
LDH: lactate dehydrogenase; UNL: upper normal limit; HCG: human chorionic gonadotropin; EP: etoposide; BEP: bleomycin, etoposide, and cisplatin; HOP: ifosfamide, vincristine, and cisplatin; VAB: vinblastine, actinomycin D, and bleomycin.
No. of patients	145	100
Age (yrs)
Median	39	—
Range	20–64	—
Primary tumor site
Testis	121	83.4
Mediastinum	14	9.7
Retroperitoneum	10	6.9
Stage (in patients with of testicular seminoma)
IIC	74	61.2
IIIA/B	34	28.1
IIIC	13	10.7
Prior radiotherapy
No	103	71.0
Yes	42	29.0
Prechemotherapy serum LDH value (n = 118 patients)
Normal	58	49.2
1–2 × UNL	24	20.3
2–3 × UNL	10	8.5
3–4 × UNL	9	7.6
> 4 × UNL	17	14.54
Prechemotherapy serum HCG value (n = 138 patients)
Normal	70	54.7
Elevated	58	45.3
First-line chemotherapy
EP (4 cycles)	57	39.3
BEP (3 cycles)	29	20.0
HOP (4 cycles)	17	11.7
VAB-6	26	17.9
Other	16	11.1
Recurrence
No	114	79.0
Yes	31	21.0
Long-term status
Free of disease	120	83.0
Died of disease	25	17.0

Response and Recurrence

All 145 patients were assessable for response. The median follow-up of surviving patients was 60 months (range, from 12 months to 19 years). No patients were refractory to first-line cisplatin-based chemotherapy. One hundred thirty patients (90%) achieved a complete response to chemotherapy alone. All 15 patients (10%) who achieved a favorable partial response remained free of disease progression after a median follow-up of 50 months (range, from 24 months to 13 years).

Thirty-one patients (21%) developed recurrent seminoma. The median time to recurrence was 8.8 months (range, 2–48 months). Sixteen patients developed a recurrence in the liver and/or the CNS, whereas the sites of recurrence were the lymph nodes, lungs, or bones in the other patients.

Survival

The 5-year (OS) rate for the total study population (145 patients) was 81% (95% confidence interval [95% CI], 73–87%) (Fig. 1). All surviving patients had no evidence of disease at the last follow-up. Twenty-five patients (17%) had died. The cause of death was progressive seminoma in 18 patients (12%). Four patients died of sepsis secondary to Grade 4 neutropenia after chemotherapy: Two of those patients had received prior lumbar-aortic and mediastinal radiotherapy for their seminoma, and other two patients had a concomitant medical history of acquired immunodeficiency syndrome and uncontrolled diabetes mellitus, respectively. One death was due to hepatic venoocclusive disease that occurred during intensive chemotherapy with busulfan and thiotepa administered after the patient developed recurrent seminoma in the lungs and liver.

Figure 1. Overall survival for 145 patients with advanced seminoma (n = 120 patients alive).

Two patients who presented with primary mediastinal seminoma died of treatment-induced toxicities: One patient had received initial radiotherapy to the mediastinum (35 grays [Gy] with standard dose fractionation followed by 2 fractions of 5 Gy each to the tumor bed) followed by 4 cycles of EP; the patient died 12 months later of radiation-induced mediastinal fibrosis. The second patient with no prior history of heart dysfunction had undergone complete excision of his mediastinal mass followed by adjuvant radiotherapy (35 Gy with standard dose fractionation) and chemotherapy with 4 cycles of vinblastine, ifosfamide, and cisplatin. He died of ventricular arrhythmia 3 months after the end of chemotherapy. The 5-year OS rate after a first recurrence was 33% (95% CI, 17–53%) (Fig. 2).

Figure 2. Overall survival after first recurrence in 30 patients with advanced seminoma (n = 11 patients alive).

All surviving patients had at least 18 months of follow-up after they received salvage chemotherapy. Patient outcome was found to be correlated with the site of metastasis and the salvage therapy regimen and is presented in Table 2. Thirteen of 16 patients (81%) who had a liver and/or CNS recurrence died of progressive disease. The only long-term (≥ 18 months) disease free survival (DFS) after a liver recurrence was obtained after two consecutive cycles of HDCT with the ICE regimen.

Table 2. Patients with Recurrent Seminoma: Outcome According to Salvage Treatment and the Site of Metastasis

Salvage treatment/site of metastases	No. of patients	Outcome
		DOD	NED
DOD: died of disease; NED: no evidence of disease; VeIP: vinblastine, ifosfamide, and cisplatin; HDCT: high-dose chemotherapy; ASCT: autologous stem cell transplantation; N/L/B: lymph node/lung/bone; CNS: central nervous system.
Four cycles of VeIP (n = 12 patients)
N/L/B	8	3/8	5/8
CNS/liver	4	4/4	0/4
Three cycles of VeIP plus HDCT and ASCT (n = 6 patients)
N/L/B	4	0/4	4/4
CNS/liver	2	1/2	1/2
Other treatment modalities (n = 12 patients)
N/L/B	3	3/3	0/3
CNS/liver	9	8/9	1/9

Prognostic Factors for Survival Initially and at Recurrence

Table 3 shows the 5-year DFS and/or OS rates for the individual clinical factors examined. There was no prognostic value for age at diagnosis (younger than 30 years vs. 30 years and older; P = 0.89), for primary tumor site (P = 0.96), or for pretreatment serum HCG values (P = 0.94). Prior radiotherapy to lumbar-aortic and ipsilateral pelvic lymph nodes was not associated with a worse prognosis (P = 0.74). However, four patients had received initial lumbar-aortic and mediastinal radiotherapy, and all of them experienced major hematologic toxicities, including two deaths due to sepsis during first-line chemotherapy.

Table 3. Univariate Analysis of 5-Year Disease-Free Survival and/or Overall Survival According to Initial Patient Characteristics

Characteristic	No. of patients	OS (%)	P value	DFS (%)	P value
OS: overall survival; DFS: disease-free survival; HCG: human chorionic gonadotropin; LDH: lactate dehydrogenase; UNL: upper normal limit.
Age at diagnosis
< 30 yrs	31	73	0.89	—	—
≥ 30 yrs	113	80	—	—	—
Prior radiotherapy
No	103	79	0.74	—	—
Yes	42	78	—	—	—
Primary site
Testis	121	84	0.96	—	—
Extragonadal	24	90	—	—	—
Initial serum HCG
Normal	70	80	0.94	—	—
Elevated	58	82	—	—	—
Initial serum LDH
<2 × UNL	82	85	0.2	80	0.12
≥ 2 × UNL	36	77	—	70	—
Disease stage
Stage IIC	74	85	—	78	—
Stage IIIA/B	34	77	0.004	71	0.024
Stage IIIC	13	48	—	35	—

No difference was found in DFS (P = 0.12) or OS (P = 0.2) for prechemotherapy serum LDH values < 2 times the upper normal limit or LDH values ≥ 2 times the upper normal limit. The results demonstrated that initial disease stage was the only prognostic factor for DFS and OS in both the univariate and multivariate analyses. The 5-year OS rate for patients with Stage IIC seminoma was 87% (95% CI, 76–93%), compared with 76% for patients with Stage IIIA/B seminoma (95% CI, 58–88%) and 48% for patients with Stage IIIC seminoma (95% CI, 22–74%) (Figs. 3, 4).

Figure 3. Disease free survival according to initial disease stage in patients with advanced seminoma.

Figure 4. Overall survival according to initial disease stage in patients with advanced seminoma.

Primary mediastinal seminoma was not associated with an adverse prognosis, because all 12 patients who received cisplatin-based chemotherapy achieved a complete response and had not developed recurrent disease after a median follow-up of 7 years. In patients who developed a recurrence after receiving first-line, cisplatin-based chemotherapy, liver and/or CNS metastases clearly were associated with a greater mortality rate compared with lung, bone, or lymph node involvement (P = 0.019) (Fig. 5).

Figure 5. Overall survival according to the site of recurrence in patients with advanced seminoma.

DISCUSSION

In this retrospective review of treatment outcomes in patients with advanced seminoma, first-line, cisplatin-based chemotherapy yielded a complete response or a favorable partial response in all study patients, and 83% of patients remained free of disease with a median follow-up of 60 months. Because advanced seminoma is relatively rare and because the cure rate obtained with cisplatin-based chemotherapy regimens is high, few studies have focused on prognostic factors, and studies that have reported on prognostic factors yielded varying results: Mencel et al.10 and Fossa et al.18 found that high prechemotherapy serum LDH levels (and also HCG levels in the study by Mencel et al.) were poor prognostic factors, whereas Peckham et al.19 and Loehrer et al.8 stated that prior radiation therapy to both retroperitoneal and mediastinal lymph nodes portended a worse prognosis. Other authors, such as Tjulandin et al.20 and Culine et al.,12 did not identify any poor prognostic factors. In 1997, an International Consensus Prognostic Classification for metastatic germ cell tumors11 identified two groups of patients with advanced seminoma based on the presence or absence of nonpulmonary visceral metastases, with 5-year OS rates of 72% and 86%, respectively. A recent meta-analysis21 that pooled data from 566 patients with advanced seminoma found that variables such as age older than 50 years, previous radiotherapy, visceral or pulmonary metastases, and 2 or more other metastatic sites were significant negative factors for progression free survival and OS in univariate analyses. Our retrospective study results are consistent with the findings of the International Germ Cell Consensus Classification: The site of metastasis at diagnosis is the only initial prognostic factor. In addition, prior extensive-field radiotherapy increased the risk of first-line, chemotherapy-induced toxicity, resulting in either death or a reduction in the dose intensity of chemotherapy and, consequently, potentially lower complete remission rates and a greater risk of recurrence.

The prechemotherapy serum LDH level had no prognostic value. This may be explained by the limited statistical power of the study. However, reelevation of serum LDH levels after the completion of first-line chemotherapy is a major sign of recurrence, because it exceeded the upper normal limit in 24 of 31 patients (77%) who developed recurrent disease.

Many studies have focused on the pathologic features of seminoma, and some investigators10, 12, 22–24 have pointed out that atypical seminomas may have a more aggressive clinical course, exhibiting different cytologic and immunohistochemical characteristics compared with those found in classic seminomas: a high mitotic index, nuclear pleomorphism, increased nuclear-to-cytoplasmic ratio, and cytoplasmic expression of low-molecular-weight keratin or the type I precursor to blood group antigens. Furthermore, Aubry et al.23 showed that the MAGE-A4 gene, a germ cell specific marker, is expressed uniformly in classical seminoma and is absent in anaplastic seminoma and in nonseminomatous germ cell tumors, supporting the hypothesis that anaplastic seminoma may be a borderline entity between seminomatous and nonseminomatous germ cell tumors. The presence of anaplastic seminoma may be a pretreatment factor that indicates a poor prognosis. A blinded review of primary tumor specimens from the series is ongoing in the Pathology Department at the Institut Gustave Roussy to investigate this hypothesis.

This series was singular, in that it included high percentages of patients who had liver and/or CNS metastases either initially or at recurrence: Thirteen patients (9%) had either liver and/or CNS metastases at presentation. Seven of those 13 patients developed recurrences, accounting for 23% of all recurrences. Fifty percent of all patients who developed recurrences had liver and/or CNS metastases, compared with 0% and 7% of patients in the 2 major studies on advanced seminoma reported in the literature.14, 25 This may explain the differences observed between those series and our results concerning disease recurrence after first-line chemotherapy (21%, vs. 9% for the Memorial Sloan Kettering Cancer Center [MSKCC] study10) and in the 5-year OS rate after a first recurrence (33%, vs. approximately 50% for both the MSKCC study25 and the Indiana University Medical Center study14).

The study results emphasize the limited efficacy of the current, standard, first-line and salvage chemotherapy regimens in achieving a durable complete response in patients with liver and/or CNS metastases: Fifty-four percent of patients with Stage IIIC seminoma developed recurrences after they received standard, first-line, cisplatin-based chemotherapy; and only 2 of 15 patients (13%) with CNS and/or liver recurrences achieved a durable complete response, compared with 58% of patients with lymph node, lung, and/or bone recurrences.

Seminoma remains a highly chemosensitive malignancy. Single-agent paclitaxel was investigated first in 1994 in Phase II trials as treatment for patients with recurrent germ cell tumors, and durable responses were reported.26 In addition, sequential, dose-intensive chemotherapy regimens consisting of paclitaxel, ifosfamide, carboplatin, and EP yielded a 41% durable response rate in patients with germ cell tumors who had unfavorable prognostic features for achieving a durable complete response to conventional-dose salvage programs.27 Prospective trials testing novel antineoplastic agents are warranted for patients with Stage IIIC seminoma to define the optimal first-line treatment modality. HDCT may be a very interesting treatment option for patients with recurrent, advanced seminoma, particularly in the presence of nonpulmonary visceral metastases, and even when lymph node, and/or pulmonary, and/or bone recurrences, because all four patients in the current study who had these features achieved a durable complete response after HDCT, compared with a 60% complete response rate after salvage treatment with four cycles of VeIP (Table 2). Although the first results from a large, prospective, multicentric European trial28 in patients with germ cell tumors did not demonstrate the superiority of HDCT over conventional ifosfamide and cisplatin-based salvage chemotherapy regimens in terms of DFS and OS, those results were inconclusive, because the trial included a very small percentage of patients with advanced seminomas who were not analyzed separately. Furthermore, a recent Phase II study by Rick et al.29 showed that HDCT can be curative in patients with advanced seminoma even if it is offered as a second salvage treatment.

In conclusion, advanced seminoma is a highly chemosensitive malignancy. The presence of liver or CNS metastases, initially or at recurrence, is associated with high treatment failure rates after standard first-line and salvage chemotherapy regimens. Large, multicentric studies are warranted to define new prognostic factors so that a more individualized treatment strategy can be proposed based on both clinical and pathologic patterns.

Acknowledgements

The authors thank Ms. Lorna St. Ange for editing.