We thank the Orchid Cancer Appeal, UK, for support of the research nurses involved. Please note there was no direct funding for this study.
The management of brain metastases in patients with germ cell tumors remains controversial. The authors assessed the outcome in this patient group after the introduction of GAMEC chemotherapy (14-day cisplatin, high-dose methotrexate, etoposide, and actinomycin-D with filgrastim support) and cessation of the routine use of cranial irradiation.
Data were recorded prospectively from 39 patients with germ cell tumors and concurrent brain metastases who received treatment before and after the advent of GAMEC after they relapsed on conventional cisplatin-based chemotherapy. Neurosurgery was offered to selected patients. Radiotherapy generally was used only as a salvage therapy after chemotherapy failure. The primary outcome measure was overall survival and was depicted using a Kaplan-Meier plot.
The 3-year overall survival rates were 38% for the whole cohort, 69% for those who presented with brain metastases at diagnosis (group 1), and 21% and 0% for those who developed metastases after initial chemotherapy (group 2) and while receiving chemotherapy (group 3), respectively. For the whole cohort, the median overall survival was 10.6 months (range, 5.5 months to not evaluable); and, for groups 1, 2, and 3 individually, the overall survival was not yet reached (range, from 7.4 months to not evaluable), 6.2 months (range, 2.1-15.3 months), and 2.7 months (range, from 0.6 months to not evaluable), respectively. The 3-year survival rate for those who received GAMEC chemotherapy was 56% compared with 27% for those who received chemotherapy pre-GAMEC.
Germ cell tumors are among the most common cancers in young men. The prognosis for patients with advanced germ cell tumors can be estimated according to the risk criteria set out by the International Germ Cell Consensus Classification. Patients with good-prognosis disease have a 5-year survival rate >90%, but this falls to approximately 50% for those in the poor-risk group. Patients with nonpulmonary visceral metastases, which include brain metastases, fall into the poor-prognosis group, but the 5-year survival rates for patients with brain metastases appears to be worse than the rates for others within this group: 33% compared with 35% for those with bone metastases and 49% for those with liver metastases. Brain metastases occur in approximately 10% of patients who have advanced, metastatic germ cell tumors. They are present in approximately 3% of patients at diagnosis, but autopsy studies indicate that this rises to 40% in patients who die from the disease.
In their retrospective review, Bokemeyer et al separated patients into 3 groups. Those who presented with brain metastases had the best prognosis (43%; group 1), those who developed metastases after the completion of chemotherapy had a poorer prognosis (group 2), and those who developed metastases during chemotherapy had the worst prognosis (<5%; group 3). A larger series reported similar overall survival (OS) rates in these groups but also identified an association between improved outcomes and the receipt of cisplatin-based and ifosfamide-based treatments (2-year survival rate, 67%) or intravenous cisplatin, vincristine, and methotrexate (1 g/m2) with additional intrathecal methotrexate alternating with actinomycin-D, cyclophosphamide, and etoposide at presentation (POMB/ACE).
Several other small series reported an association between improved survival and the presence of solitary rather than multiple brain metastases. All patients who did not achieve complete remission with chemotherapy went on to undergo surgery and receive cranial irradiation.[5, 6] Most retrospective studies have used cranial radiotherapy in combination with systemic chemotherapy to achieve the best outcome, but no randomized controlled studies have defined the optimal treatment for patients with brain metastases, and treatment strategies vary.[4, 7, 8] There have been suggestions that chemotherapy alone might be as effective in achieving complete remission and long-term survival in patients with brain metastases, especially for those who present with brain metastases at diagnosis. The numbers are small, but the chemotherapy regimen POMB/ACE, when supplemented with intrathecal methotrexate, led to 8 long-term survivors in a group of 10 patients. High-dose chemotherapy seems to provide higher OS rates.
With concerns about the risk of late treatment-related toxicities associated with cranial irradiation, such as cognitive decline in young long-term survivors, the routine use of cranial radiotherapy in patients with germ cell tumors and brain metastases should be questioned, although it remains part of the National Comprehensive Cancer Network practice guidelines. The relatively small number of patients with germ cell tumors and brain metastases creates difficulties for the design of prospective studies, and retrospective studies remain important in developing and optimizing treatment strategies for these patients.
In this single-institution study, we report outcome data for patients with advanced germ cell tumors and brain metastases who received treatment between 1993 and 2012. Initially, treatment was with standard platinum-based chemotherapy (cisplatin or carboplatin with etoposide and bleomycin); however, later, 2 different regimens were developed: GAMEC (14-day treatment with cisplatin, high-dose methotrexate [8 g/m2], actinomycin-D, and etoposide with growth factor support) and IPO (irinotecan, paclitaxel, and oxaliplatin) consolidated with high-dose chemotherapy for those who progressed after GAMEC. The introduction of these regimens coincided with a cessation in the routine use of cranial irradiation in patients who presented with brain metastases or those who were treated at first relapse with GAMEC. Therefore, we wished to determine the outcome in this particular patient group when routine cranial irradiation was stopped and a high-dose methotrexate regimen, without the use of intrathecal methotrexate, was introduced.
MATERIALS AND METHODS
Between 1993 and 2012, 39 consecutive patients with germ cell tumors and brain metastases who were treated at Barts Cancer Center, St. Bartholomew's Hospital London, were identified from patient records. Seventeen patients (44%) fell into group 1 (those who presented with brain metastases), 16 patients (41%) fell into group 2 (those who developed metastases after the completion of chemotherapy), and 6 patients (15%) fell into group 3 (those who developed metastases during chemotherapy). The median patient age was 29 years (range, 20-53 years). Further patient characteristics are provided in Table 1.
Thirty-seven patients were treated with primary chemotherapy. Before the year 2000, patients received standard cisplatin-containing chemotherapy regimens; but, after the year 2000, patients received GAMEC either as primary therapy (patients in group 1 or those patients with International Germ Cell Consensus Classification poor-prognosis disease) or as salvage therapy after relapse, regardless of the presence or absence of brain metastases. In terms of which group patients fell into, the demographics were comparable between those who received treatment in the pre-GAMEC era and those who received treatment in the GAMEC era. In total, 21 patients received GAMEC, either at presentation (n = 15) or after relapse (n = 6). Neurosurgery was offered to selected patients, mainly those with solitary brain metastases (n = 9). Radiotherapy was reserved for patients who failed salvage chemotherapy. Survival was calculated according to the Kaplan-Meier method, and survival curves were compared using the log-rank test.
The GAMEC regimen consisted of cisplatin 100 mg/m2 in weeks 1, 3, 6, and 8 plus 50 mg/m2 in weeks 2 and 4; actinomycin D 1 mg/m2 in weeks 1, 3, 6, and 8; high-dose methotrexate 8 g/m2 (with dose adjustments for renal impairment) in weeks 1, 3, 6, and 8; and etoposide 360 mg/m2 in weeks 1, 3, 6, and 8. The IPO regimen consisted of oxaliplatin 100mg/m2 day 1, irinotecan 200mg/m2 day 1, paclitaxel 80mg/m2 day 1,8 and 15 every 21 days. The IPO regimen consisted irinotecan 200 mg/m2 on day 1; paclitaxel 80 mg/m2 on days 1, 8, and 15; and irinotecan 200 mg/m2 on day 1 repeated every 3 weeks for a maximum of 4 cycles with high-dose consolidation using carboplatin at an area under the receiver operating characteristic curve of 21, topotecan 30 mg/m2, and thiotepa 500 mg/m2.
IPO followed by high-dose chemotherapy was used as a salvage regimen if the primary therapy was GAMEC. If there was evidence of progression on IPO but the patient was deemed fit enough to proceed to high-dose chemotherapy, then a combination of high-dose carboplatin and etoposide was used.
Outcome data are provided in Table 2. The 3-year median OS rate for the whole cohort was 38%, and it was 69% for group 1, 22% for group 2, and 0% for group 3 (Fig. 1). The median OS for the whole cohort and for groups 1, 2, and 3 individually was 10.6 months (range, from 5.5 months to not evaluable [NE]), not yet reached (range, from 7.4 months to NE), 6.2 months (range, 2.1-15.3 months), and 2.7 months (range, from 0.6 months to NE), respectively. One patient in group 2 developed fatal cerebral herniation within 6 hours of starting chemotherapy despite the use of steroids in the 2 days preceding chemotherapy. Three-year survival rates were 44% for the 27 patients who were treated after 2000 compared with 25% for those patients who were treated before the year 2000. The management of individual patients is detailed in Figures 2-4.
Table 2. Summary of Patient Outcomes for All Patients and Within Specific Groups
Group 1: Presented With Brain Mets
Group 2: Relapsed With Brain Mets
Group 3: Brain Mets During/After Treatment
Abbreviations: Mets, metastasis; NE, not evaluable; NYR, not yet reached; OS, overall survival.
No. progression-free (%)
No. progression-free with surgery/total no.
OS: Median [range], mo
OS rate, %
Median follow-up, y
Data surrounding the specific number of brain metastases were available for 38 of the 39 patients. Twenty patients had a solitary brain metastasis and 18 patients had ≥2 metastases, and the long-term OS rates for these groups were 51.5% and 25%, respectively (Fig. 3). In total, 9 patients underwent neurosurgery, including 6 patients who had a solitary metastasis. Only 1 patient in group 1 underwent neurosurgery, and he achieved a cure. In group 2, 6 of 15 patients underwent neurosurgery, and 3 of those patients were progression-free at 3 years; whereas 9 patients did not undergo surgery, and all died within 16 months. In group 3, 2 patients underwent neurosurgery, including 1 who had a single metastasis, and both patients died.
Regarding the pattern of failure 14 patients progressed overall. In 7 patients, progression was in the central nervous system (CNS) only—all of these patients had more than 1 cerebral metastasis. In 1 patient who had a single cerebral metastasis, progression occurred with the development of meningeal deposits and multiple cranial nerve palsies. In 5 patients, disease progression was both in the brain and outside the CNS; 3 of these patients only had a single cerebral metastasis. One patient with multiple cerebral metastases progressed outside the CNS alone.
Radiotherapy was offered to 11 patients, including 2 patients in group 1, 6 patients in group 2, and 3 patients in group 3. It was never offered as an initial therapy and only appears to have made a possible contribution to outcome in 2 patients. In both of those patients, there was no evidence of progression in the brain before radiotherapy.
Eleven patients received high-dose chemotherapy, including 5 patients in group 1 (see Fig. 2), and all have remained alive. One patient received high-dose chemotherapy after failure on GAMEC, 3 received it after failure on conventional-dose chemotherapy, and all have remained progression-free. In 1 patient, high-dose chemotherapy failed; and combined bleomycin, etoposide, and cisplatin after failure on high-dose chemotherapy was successful. Four patients were in group 2 (see Fig. 3), including 1 who received high-dose chemotherapy after failure on conventional-dose chemotherapy, and he is progression-free. The other 3 patients all progressed in group 3 (see Fig. 4).
High-dose chemotherapy seemed particularly useful on first relapse, specifically for patients who had failed on GAMEC. The combination of GAMEC followed by IPO and high-dose chemotherapy seemed to produce the greatest overall responses. It is noteworthy that only 1 patient from group 1 achieved a cure using conventional chemotherapy alone.
Germ cell tumors are uniquely chemosensitive and often curable, even with advanced, metastatic disease. However, in other chemosensitive tumors, such as small cell lung cancer, CNS recurrence occurs because of issues with poor chemotherapeutic penetration through the blood-brain barrier and can terminate a complete remission. It is unclear whether the CNS represents a sanctuary site in germ cell tumors, and the relevance of the blood-brain barrier in the treatment of patients with brain metastases remains controversial. Furthermore, there is no clear consensus regarding the optimal therapeutic strategy for these patients, although several groups have recommended treating solitary CNS metastases with either stereotactic radiosurgery or surgical resection followed by cisplatin-based chemotherapy and treating multiple CNS metastases with whole-brain radiation therapy and concomitant cisplatin-based chemotherapy.[3, 13] Others generally advocate treating with chemotherapy alone and then re-evaluating with CNS imaging. If complete remission is achieved, then additional treatment often is not advised, and close follow-up is recommended. However, in the context of small, residual CNS disease or a solitary lesion, subsequent surgical excision or stereotactic radiosurgery often is used. Whole-brain radiotherapy often is reserved for highly symptomatic patients with large CNS metastases or for those with persistent or recurrent metastases. It is noteworthy that, in patients who presented with brain metastases at the time of germ cell tumor diagnosis, Fossa et al demonstrated that systemic chemotherapy had the greatest impact on survival and that cerebral radiotherapy did not significantly influence outcome. However, cerebral radiotherapy and neurosurgery appeared to represent essential treatment modalities for patients in whom brain metastases were diagnosed after induction chemotherapy. Nonetheless, the description of a group of patients with newly diagnosed CNS germ cell tumors being successfully treated with chemotherapy alone provides evidence against there being a functional blood-brain barrier in this disease. Furthermore, it appears that, in half of patients, progression in the CNS is associated with concurrent progression outside the CNS, suggesting that systemic therapy should represent the mainstay of treatment in this patient subgroup. In the remaining 50% of patients who had progression within the CNS, many had more than 1 metastasis; and the question of whether selective radiation to these lesions may offer some benefit is clearly relevant to these patients, particularly for areas of the brain in which surgery is likely to lead to neurologic deficit.
Our study included all patients with germ cell tumors and concurrent brain metastases who were treated at our center between 1993 and 2012. During that period, treatment for poor-prognosis disease, as well as salvage therapy for relapse, was changed to GAMEC, an intensified, cisplatin-based regimen that incorporates high-dose methotrexate. The doses used are far higher than those used in the POMB/ACE schedule; and, because methotrexate can cross the blood-brain barrier, this regimen may negate the use of intrathecal chemotherapy. We have demonstrated that the introduction of high-dose methotrexate and the cessation of cranial irradiation across patient groups do not appear to be associated with poorer survival. It is interesting that the patients who progressed after receiving GAMEC did not progress in the brain alone, providing some evidence that, in these patients, the blood-brain barrier may not be the cause of treatment failure and that cranial irradiation may not have affected outcome. However, we also have demonstrated that high-dose salvage chemotherapy in patients who relapsed with brain metastases was clearly effective in some cases, and this may relate to the finding that the drugs used in such regimens can penetrate the blood-brain barrier at the doses used, but not at the lower doses administered in previous lines of therapy. Furthermore, high-dose chemotherapy was able to rescue patients who had failed on high-dose methotrexate in several cases. Whether the combinations used were particularly important is unclear, but the relative success post-GAMEC using 2 drugs with known brain penetration (topotecan and thiotepa) is encouraging.
A weakness in the current prognostic scores is that, for patients with brain metastases in whom relapse is determined by rising tumor markers, physicians can never be sure about the origin of such rises in markers. Furthermore, if there are persistent brain lesions, then it is not always clear whether those lesions are active or necrotic. Consequently, patients receiving whole-brain radiotherapy in this situation may be receiving it needlessly. The role of radiotherapy given with curative intent remains controversial, and there are increasing concerns about long-term side effects, such as progressive, multifocal leukoencephalopathy with ataxia; concentration difficulties; and sensory motor changes. The value of stereotactic radiotherapy remains undefined; and, although it is likely to lead to fewer long-term side effects, it may not treat micrometastatic disease within the CNS, which is an important consideration if a blood-brain barrier is believed to be present.
The 2 current prognostic systems based on the time to development of brain metastases and the number of metastases (single vs multiple) both were validated in this series. Our data demonstrate a long-term survival rate of 69% in patients who presented with brain metastasis (group 1) and received first-line chemotherapy with GAMEC, which is above those rates previously reported. For our patients in group 2, the long-term survival rate was 21%, which was similar to the rate reported by others for patients who had brain metastases at initial diagnosis.
At the time of this writing, 16 patients within this study were free from disease progression, and 11 had achieved this with chemotherapy alone. This suggests that GAMEC can achieve long-term survival on its own, without the need for additional treatment modalities. The relative failure of initial conventional chemotherapy to produce long-term survival in patients with brain metastases is disappointing. Many of these patients who were treated before the year 2000 will have received doses of etoposide and bleomycin that would be considered suboptimal today, and this may have accounted in part for the poor initial outcome.
Cisplatin-based chemotherapy remains the standard of care for patients with brain metastases. However, there are suggestions that high-dose chemotherapy or high-dose methotrexate-based chemotherapy may be superior regimes for these patients and that this strategy may be responsible for the improved outcome described for such patients in this study group. Large, systemic analyses and prospective trials investigating the optimal management of patients with germ cell tumors and brain metastases are still lacking. The rarity of this clinical situation means that conducting large clinical trials in this area is especially difficult, but such studies are required so that treatment for these patients can be sufficiently optimized.