Search tips
Search criteria 


Logo of annoncLink to Publisher's site
Ann Oncol. 2016 May; 27(5): 875–879.
Published online 2016 February 9. doi:  10.1093/annonc/mdw045
PMCID: PMC4843188

Prognostic factors in patients with poor-risk germ-cell tumors: a retrospective analysis of the Indiana University experience from 1990 to 2014



Based on the risk stratification from the International Germ Cell Cancer Collaborative Group (IGCCCG), only 14% of patients with metastatic germ-cell tumors (GCT) had poor-risk disease with a 5-year progression-free survival (PFS) rate of 41% and a 5-year overall survival (OS) rate of only 48%. This analysis attempts to identify prognostic factors for patients with poor-risk disease.

Patients and methods

We conducted a retrospective analysis of all patients with GCT diagnosed and treated at Indiana University from 1990 to 2014. Clinical and pathological characteristics were available for all patients and all of them were treated with cisplatin–etoposide-based chemotherapy. Cox proportional hazards models were used to target significant predictors of disease progression and mortality. A significance level of 5% was used in the analysis.


We identified 273 consecutive patients with poor-risk GCT (PRGCT). Median follow-up time was 8 years (range 0.03–24.5). The 5-year PFS and OS rates were 58% [95% confidence interval (CI) 51% to 63%] and 73% (95% CI 67% to 78%), respectively. In multivariate survival analyses, multiple risk factors were associated with disease progression, including liver metastasis, brain metastasis, primary mediastinal nonseminomatous GCT (PMNSGCT), and elevation in logarithmic β-hCG. Significant predictors of mortality were PMNSGCT [hazard ratio (HR) 4.63, 95% CI 2.25–9.56; P < 0.001], brain metastasis (HR 3.30, 95% CI 1.74–6.23; P < 0.001), and increasing age (HR 1.03, 95% CI 1.01–1.06; P = 0.02).


Patients with PMNSGCT, brain metastasis, or with increasing age are at higher risk of death than their counterparts. This contemporary cohort (1990–2014) of 273 patients with PRGCT had improved PFS and OS outcomes than those from the historical IGCCCG group of patients (1975–1990).

Keywords: germ-cell tumor, poor-risk germ-cell tumor, poor-risk testicular cancer, primary mediastinal nonseminomatous germ-cell tumor


Although germ-cell tumors (GCTs) represent only 1% of male cancers, they are the most common cancer in young men between 15 and 35 years of age [1]. The incidence has risen over the last century [2]. GCTs are classified into seminomas and nonseminomatous germ-cell tumors (NSGCT). Pure seminomas account for ~50% cases with the rest compromised of NSGCT and tumors with combined seminomatous and nonseminomatous elements [3].

Various prognostic classification models have been proposed to estimate survival in patients with metastatic disease. An International Collaborative Group was established to bring consensus on this rare disease and the results of this group were published in 1997. Today, GCTs are classified by the International Germ Cell Cancer Collaborative Group (IGCCCG) into good, intermediate, and poor risks [4]. Good-risk GCT represents 60% of all metastatic cases with a 5-year progression-free survival (PFS) rate of 89% and a 5-year overall survival (OS) rate of 92%. Intermediate-risk GCT represents 26% of all cases with a 5-year PFS rate of 75% and a 5-year OS rate of 80%. The poor-risk GCT (PRGCT) category represents 14% of patients with a 5-year PFS rate of 41% and a 5-year OS rate of 48%. This dataset was accumulated from 1975 till 1990 and included treatment regimens that substituted carboplatin for cisplatin as well as nonetoposide regimens such as cisplatin, vinblastine, and bleomycin. Today, these older regimens are known to be inferior to the current standard of care [5].

PRGCT is defined by one or more of the following criteria: primary mediastinal nonseminomatous germ-cell tumor (PMNSGCT) or by testicular or retroperitoneal primary NSGCT with nonpulmonary visceral metastasis (NPVM), AFP >10 000 ng/ml, or β-hCG >50 000 mIU/ml. The heterogeneity of tumor aggressiveness in the PRGCT category has not been well established. Having further prognostic factors among the poor-risk group would provide oncologists with additional information regarding the patient's chance of cure with frontline therapy. Moreover, this information would help them in identifying patients that are at higher risk for treatment failure or relapse and that might be suitable for closer follow-up and/or novel approaches.

patients and methods

We conducted a retrospective analysis of patients with metastatic GCTs who were treated at Indiana University between 1990 and 2014 and identified patients with poor-risk disease according to IGCCCG criteria [4]. All patients were treated with standard cisplatin–etoposide-based chemotherapy or on prospective clinical trials [5]. Eligibility criteria for the study consisted of the following: any primary site, nonseminoma histology subtype, poor-risk IGCCCG category, diagnosed during the time frame 1990–2014, initial diagnosis, and treatment received at Indiana University. Patients who were referred to our center for second opinion after having received prior therapy were excluded from this analysis. Although serum LDH elevation was considered an independent adverse prognostic factor in the IGCCCG classification, this was not used as a sole criterion for risk stratification in this study as LDH can be a nonspecific in many instances.

After institutional review board approval was obtained, clinical data were collected from physical charts and electronic medical records. This included patient and disease characteristics, chemotherapy regimens, surgical resections, and survival outcomes. This study utilized the secure web-based, Title 21 CFR Part 11 compliant Research Electronic Data Capture (REDCap) system for data input.

statistical analysis

Analyses were carried out using SAS software, version 9.4. Patient and disease characteristics were tabulated for the population being studied. The primary aim for the present analysis was to determine the PFS and OS of patients with PRGCT at 5 years and to identify important clinical factors for disease progression and survival. PFS, OS, and follow-up period were calculated from the date of chemotherapy initiation to the date of relapse, date of death, and date of last follow-up, respectively. The PFS and OS rates were estimated using Kaplan–Meier methods. A significance level of 5% was used in the analysis.

The following variables were included in the analysis: age, location of primary tumor (testes/retroperitoneal versus mediastinal), presence or absence of NPVM (liver, brain, or bone metastasis), serum tumor marker levels of log-transformed AFP and β-hCG. Age was used as a continuous variable in this analysis. Log-rank tests were used to compare the survival functions between patients with different clinical profiles for each of the independent variables. We also analyzed PFS and OS in patients with single IGCCCG criterion for poor-risk disease versus patients with two or more IGCCCG criteria for poor-risk disease. Multivariate Cox proportional hazards models were used to identity important clinical factors in predicting PFS and OS and estimate the hazard ratio (HR) for each variable while adjusting for other covariates. No model reduction was applied and all potential prognostic factors were assessed in multivariable models.


patient and disease characteristics

Overall, 273 consecutive patients with IGCCCG PRGCT were diagnosed and received initial therapy at Indiana University between 1990 and 2014. Median age at diagnosis was 27 years (range 13–56 years). Median follow-up time was 8 years (range 0.03–24.5 years).

Patient and disease characteristics at the start of chemotherapy are listed in Table Table11.

Table 1.
Patient and disease characteristics at the beginning of chemotherapy

treatment and disease outcomes

Treatment decisions were based on clinical characteristics and marker amplitude. Frontline chemotherapy consisted of four cycles of bleomycin–etoposide–cisplatin (BEP) in 178 (65%) patients or four cycles of etoposide–ifosfamide–cisplatin (VIP) in 60 (22%) patients. Other frontline chemotherapy included four cycles of etoposide–ifsofamide–cisplatin–vinblastine–bleomycin (VIP/VB) in 23 (9%) patients or two cycles of BEP followed by high-dose chemotherapy×2 in 12 (4%) patients as part of clinical trials. These regimens were found to be equivalent in clinical trials but with various toxicity profiles [68].

In this study, the 5-year PFS for BEP×4 was 58% [95% confidence interval (CI) 49% to 66%] and was 57% (95% CI 49% to 65%) for other chemotherapy regimens (P = 0.9). The 5-year OS for BEP×4 was 73% (95% CI 64% to 80%) as compared with 74% (95% CI 65% to 80%) for other chemotherapy regimens (P = 0.71) (Supplemetary Figures S1 and S2, available at Annals of Oncology online).

One hundred sixty of 273 patients (59%) were continuously disease-free after first-line chemotherapy with a median follow-up time of 8 years. The 5-year PFS was 58% (95% CI 51% to 63%) and the 5-year OS was 73% (95% CI 67% to 78%). Survival outcomes in this study are depicted in Figure Figure11.

Figure 1.
Study flow chart of 273 patients with poor-risk germ-cell tumor.

One hundred three patients (38%) had postchemotherapy retroperitoneal lymph node dissection (PCRPLND). Pathology of PCRPLND included GCT in 13 patients, teratoma in 53 patients, necrosis in 35 patients, and malignant transformation of teratoma in 2 patients. Ninety-three patients (34%) had thoracic surgery with pathology including GCT in 26, teratoma in 32, necrosis in 27, and malignant transformation of teratoma in 8 patients.

Second-line chemotherapy consisted of two cycles of high-dose carboplatin and etoposide [9] in 45 patients or other salvage standard dose chemotherapy in 37 patients. Forty-eight patients (18%) had salvage surgery defined by postchemotherapy surgery in the setting of rising markers. Among these patients, 22 had no evidence of disease at last follow-up and 26 were dead of disease progression.

Seven patients (3%) experienced a late relapse, defined by relapse more than 2 years after initiating chemotherapy. Four patients (2%) had a second primary testicular cancer.

univariate analysis

We carried out univariate analysis using Kaplan–Meier estimates and log-rank tests to assess the impact of clinical factors on PFS and OS. Patients with NPVM had worse PFS than patients without NPVM: 5-year PFS 52% (95% CI 43% to 60%) versus 63% (95% CI 54% to 70%) (P = 0.05). There was, however, no significance in OS analysis. Brain metastasis had the worst impact on survival with 5-year PFS 24% (95% CI 13% to 38%) and 5-year OS 51% (95% CI 34% to 66%) (supplementary Material S3, available at Annals of Oncology online). However, this represented only 41 patients.

Patients with PMNSGCT had significantly worse OS compared with testicular/retroperitoneal primary sites (Figure (Figure2):2): 5-year OS 58% (95% CI 45% to 69%) versus 79% (95% CI 72% to 84%) (P < 0.001).

Figure 2.
Overall survival for germ-cell tumor patients with primary mediastinal nonseminomatous germ-cell tumor versus primary testicular/retroperitoneal.

Patients with multiple IGCCCG criteria for poor-risk disease had worse 5-year PFS compared with patients with a single IGCCCG criterion (Figure (Figure3):3): 48% (95% CI 38% to 57%) versus 64% (95% CI 56% to 71%) (P = 0.01). This did not translate into a worse OS: 5-year OS 69% (95% CI 59% to 77%) versus 76% (95% CI 68% to 82%) (P = 0.17).

Figure 3.
Progression-free survival for patients with single IGCCCG criteria versus multiple IGCCCG criteria of poor-risk disease.

multivariate analysis

We then carried out multivariate analysis to evaluate the impact among the predictors on PFS and OS while adjusting for other covariates. Factors in this multivariate model included: liver metastasis, brain metastasis, bone metastasis, PMNSGCT, age as a continuous variable, and logarithmic AFP and β-hCG. The adjusted HRs of these factors for PFS and OS are provided in Table Table2.2. Using this multivariable model, patients with liver metastasis, brain metastasis, PMNSGCT, or elevation in logarithmic β-hCG had worse PFS. Patients with brain metastasis, PMNSGCT, or with increasing age had worse OS.

Table 2.
Multivariate analysis for PFS and OS


Several prognostic models have been proposed for patients with GCTs. The IGCCCG risk stratification is the most utilized model in clinical practice [4].

In PRGCT patients treated at our center from 1990 to 2014, the 5-year PFS was 58% (95% CI 51% to 63%) and 5-year OS was 73% (95% CI 67% to 78%). These results appear superior to the results of the IGCCCG: 5-year PFS 41% and 5-year OS 48%. The IGCCCG evaluated patients who received heterogeneous treatment regimens and therapeutic strategy known today to be inferior to the current standard of care [5]. Moreover, salvage curative chemotherapy was not often available at the time of IGCCCG (1975–1990) and, hence, PFS and OS were very similar in that analysis. The results of other studies investigating poor-risk disease are depicted in Table Table33.

Table 3.
Poor-risk disease studies

This analysis identifies key prognostic factors in patients with PRGCT. Liver metastasis, brain metastasis, PMNSGCT, and elevation in logarithmic β-hCG were predictors of progression. Patients with PMNSGCT, brain metastasis, or with increasing age had worse OS, and this reflects higher risk for failure of salvage treatment at least in the PMNSGCT subgroup [12].

In 2000, Kollmannsberger et al. [13] reported an explorative analysis of patients with PRGCT. In this multi-institution study, patients with PMNSGCT plus lung metastasis had the worst PFS and patients with PMNSGCT plus NPVM had inferior OS compared with all other subgroups. This study was based on a median follow-up time of 23 months and did not evaluate many clinical factors as sole predictors of outcome.

This study included a large sample size of consecutive patients evaluated and treated at a tertiary care center with long follow-up. However, there are limitations to report. We assessed certain clinical factors and their impact on prognosis. There were likely other factors that were not evaluated in this analysis. Patients in this single institution study were treated at a referral center for metastatic GCT. Survival data for patients treated at other institutions might differ. Furthermore, several identified subgroups contained a limited number of patients and hence the sample size for certain subgroups was small when conducting the multivariate analysis.

Although patients with PRGCT have improved outcomes in the modern era, they continue to pose a therapeutic challenge with several trials suggesting that a therapeutic plateau has been reached [14, 15]. Further advances will require investigating the biology of these rare tumors and including novel agents that can overcome the mechanism of resistance to treatment.


LHE supported in part by the National Cancer Institute (1 RO1 CA157823-)1A1.


LHE is a Livestrong Foundation Distinguished Professor of Medicine. All remaining authors have declared no conflicts of interest.

Supplementary Material

Supplementary Data:


1. Einhorn EH. Testicular cancer: an oncological success story. Clin Cancer Res 1997; 3: 2630–2632. [PubMed]
2. Huyghe E, Matsuda T, Thonneau P Increasing incidence of testicular cancer worldwide: a review. J Urol 2003; 170: 5–11. [PubMed]
3. Oosterhuis JW, Looijenga LH Testicular germ-cell tumours in a broader perspective. Nat Rev Cancer 2005; 5: 210–222. [PubMed]
4. International Germ Cell Consensus Classification: a prognostic factor-based staging system for metastatic germ cell cancers. International Germ Cell Cancer Collaborative Group. J Clin Oncol 1997; 15: 594–603. [PubMed]
5. Williams SD, Birch R, Einhorn LH et al. Treatment of disseminated germ-cell tumors with cisplatin, bleomycin, and either vinblastine or etoposide. N Engl J Med 1987; 316: 1435–1440. [PubMed]
6. Hinton S, Catalano PJ, Einhorn LH et al. Cisplatin, etoposide and either bleomycin or ifosfamide in the treatment of disseminated germ cell tumors: final analysis of an intergroup trial. Cancer 2003; 97: 1869–1875. [PubMed]
7. Blanke C, Loehrer PJ, Nichols CR, Einhorn LH A phase II trial of VP-16, ifosfamide, cisplatin, vinblastine, and bleomycin in advanced germ-cell tumors. Am J Clin Oncol 1996; 19: 487–491. [PubMed]
8. Motzer RJ, Nichols CJ, Margolin KA et al. Phase III randomized trial of conventional-dose chemotherapy with or without high-dose chemotherapy and autologous hematopoietic stem-cell rescue as first-line treatment for patients with poor-prognosis metastatic germ cell tumors. J Clin Oncol 2007; 25: 247–256. [PubMed]
9. Einhorn LH, Williams SD, Chamness A et al. High-dose chemotherapy and stem-cell rescue for metastatic germ-cell tumors. N Engl J Med 2007; 357: 340–348. [PubMed]
10. Gundgaard MG, Lauritsen J, Mortensen MS et al. Germ cell cancer (GCC): long-term survival after treatment with bleomycin (B), etoposide (E), and cisplatin (P) in a large cohort. ASCO Meeting Abstracts 2013; 31: 4533.
11. Olofsson SE, Tandstad T, Jerkeman M et al. Population-based study of treatment guided by tumor marker decline in patients with metastatic nonseminomatous germ cell tumor: a report from the Swedish-Norwegian Testicular Cancer Group. J Clin Oncol 2011; 29: 2032–2039. [PubMed]
12. Hartmann JT, Einhorn L, Nichols CR et al. Second-line chemotherapy in patients with relapsed extragonadal nonseminomatous germ cell tumors: results of an international multicenter analysis. J Clin Oncol 2001; 19: 1641–1648. [PubMed]
13. 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 2000; 11: 1115–1120. [PubMed]
14. Nichols CR, Catalano PJ, Crawford ED et al. Randomized comparison of cisplatin and etoposide and either bleomycin or ifosfamide in treatment of advanced disseminated germ cell tumors: an Eastern Cooperative Oncology Group, Southwest Oncology Group, and Cancer and Leukemia Group B Study. J Clin Oncol 1998; 16: 1287–1293. [PubMed]
15. Fizazi K, Pagliaro L, Laplanche A et al. Personalised chemotherapy based on tumour marker decline in poor prognosis germ-cell tumours (GETUG 13): a phase 3, multicentre, randomised trial. Lancet Oncol 2014; 15: 1442–1450. [PMC free article] [PubMed]

Articles from Annals of Oncology are provided here courtesy of Oxford University Press