Clin Transl Oncol DOI 10.1007/s12094-014-1179-5
ORIGINAL ARTICLE
Synchronous versus metachronous brain metastasis from testicular germ cell tumors (TGCT): an analysis from the Spanish Germ Cell Cancer Group data base R. Girones • J. Aparicio • P. Roure • J. R. Germa-Lluch • X. Garcı´a del Muro • S. Vazquez-Estevez • A. Saenz • J. Sastre • J. Arranz Arija • E. Gallardo • E. Gonzalez-Billalabeitia • A. Sanchez-Hernandez • J. Terrasa • A. Hernandez • C. Santander • E. Cillan • N. Sagastibelza • D. Almenar-Cubells • M. Lopez Brea J. P. Maroto • Spanish Germ Cell Cancer Group (SGCCG)
•
Received: 1 February 2014 / Accepted: 24 March 2014 Ó Federacio´n de Sociedades Espan˜olas de Oncologı´a (FESEO) 2014
Abstract Background Brain metastases of testicular germ cell tumor (TGCT) are a rare event. Prognostic is poor and there is not much evidence on optimal management of these patients. Patients and methods A review of case records of germ cell tumor patients within the Spanish Germ Cell Cancer Group data base from 1994 to 2012 was conducted. Results Thirty-three out of 6,200 cases (0.5 %). Nineteen patients (57 %) group 1: synchronous, 13 (40 %) group 2: metachronous and only one developed brain metastasis during cisplatin-based chemotherapy (excluded from the analysis). Median serum BHCG levels at initial diagnosis was higher in group 1, whereas elevated AFP serum levels were more common in group 2. Histology in the primary tumor: chorionic carcinoma for group 1 versus embryonal
carcinoma for group 2. Mainly solitary brain metastasis in group 2 (54 versus 21 %, respectively). The median overall survival from the diagnosis of central nervous system involvement was 16 months for group 1 (CI 95 % 13.9–18) and 23 months (95 % CI 0–165) for group 2 (log rank p = 0.84). Long-term survivors were practically identical in the two groups (38.9 % group 1 versus 38.5 % group 2). Regardless of the timing of brain metastasis, those patients that achieved complete response to the treatment had better survival (log rank p 0.003). Conclusion Although some distinctive clinical characteristics have been found between patients with synchronous versus metachronous brain metastasis from TGCT, the timing of brain metastasis did not seem to have prognostic influence, but due to the retrospective nature of the analysis and the results should be interpreted with caution.
R. Girones (&) Hospital Lluis Alcanyis Xa`tiva, Valencia, Spain e-mail: [email protected]
J. Arranz Arija Hospital Gregorio Maran˜on Madrid, Madrid, Spain
J. Aparicio Hospital La FE Valencia, Valencia, Spain P. Roure Hospital de Vic, Barcelona, Spain J. R. Germa-Lluch X. Garcı´a del Muro Institut Catala` D’Oncologia, Barcelona, Spain S. Vazquez-Estevez Hospital Lucus Augusti Lugo, Lugo, Spain
E. Gallardo Hospital Parc Taulı´, Sabadell, Spain E. Gonzalez-Billalabeitia Hospital Morales Meseguer, Murcia, Spain A. Sanchez-Hernandez Hospital Provincial de Castello´n, Castello´n, Spain J. Terrasa Hospital Son Espases, Palma, Spain
A. Saenz Hospital Clı´nico de Zaragoza, Saragossa, Spain
A. Hernandez Instituto Oncologico Guipuzcoa San Sebastia´n, San Sebastia´n, Spain
J. Sastre Hospital Clı´nico San Carlos Madrid, Madrid, Spain
C. Santander J. P. Maroto Hospital Sant Pau Barcelona, Barcelona, Spain
123
Clin Transl Oncol
Keywords Germ cell tumors Brain metastases Chemotherapy Neurosurgery Holocraneal radiotherapy Prognostic factors Survival Epidemiology Data base
Introduction Brain metastases are a rare phenomenon in germ cell tumors (GCTs), occurring in \2–3 % of all patients and in approximately 10 % of patients with advanced GCTs [1]. However, postmortem analysis series revealed that up to 40 % of patients who died for progressive disease had nonfounded brain metastasis [2]. Most of the available data about clinical implications and management of brain metastasis from TGCT come from small case series [3]. To add information for these selected patients, we evaluated the clinical characteristics, treatment and outcome of patients with brain metastases of testicular cancer from the Spanish Germ Cell Cancer Group data-base registry.
achieving response to previous cisplatin-based chemotherapy. The only case primary resistant to cisplatin-based chemotherapy who developed brain metastasis during the chemotherapeutic treatment was excluded of this analysis. Patients were treated according to the gold standard for systemic and brain locations throughout the study period. Statistical analysis This was a descriptive study based on date prospectively collected. It was designed to identify outcomes and managements of TGCT with brain metastases. SPSS (Statistical Package for the Social Sciences) version 15.0 was used for all of the analyses. For quantitative variables, measurements of central tendency (mean and median) and measures of spread (standard deviation and interquartile range) were used. For qualitative variables, absolute and relative frequency tables were employed. Graphical representation of the data obtained was used wherever possible. Mann– Whitney U test and Fisher’s exact test were used to assess the significance of the findings. Differences with a p \ 0.05 were considered significant.
Patients and methods Ethical considerations Methods This is an analysis of patients with brain metastasis registered prospectively in the data base of the Spanish Germ Cell Cancer Group registry. Our data base comprises diagnosis, clinical management and survival data of almost 6,200 patients collected since 1994.
The institution’s ethical review board approved this study. All of the patients provided written informed consent to be included in Germinal Spanish Germ Cancer Group data base at the time of diagnosis of testicular germ cell cancer, given their permission to collect their clinical data for future scientific purposes.
Patients Results A data collection form was sent to the investigators to clarify or implement confusing or missing data, no previously reported in the data base. Median follow-up was 16 months (1–228). All patients were classified into three groups based on the time of brain metastases diagnosis: group 1 included those patients who presented with brain metastases at the time of diagnosis of primary testicular cancer; group 2 was composed of patients who developed brain metastases with or without systemic disease after E. Cillan Hospital Servet Zaragoza, Saragossa, Spain N. Sagastibelza Hospital de Donostia, San Sebastia´n, Spain D. Almenar-Cubells Hospital Doctor Peset Valencia, Valencia, Spain M. Lopez Brea Hospital Ma´rques de Valdecilla, Santander, Spain
123
A total of 42 out of 6200 (0.6 %) cases registered in our data base developed brain metastasis. Reliable data were obtained from only 33 patients; in 19 patients (57 %), brain metastasis were identified at the time of diagnosis of primary tumor (group 1), whereas 13 patients (40 %) were identified at the time of recurrence (group 2). Among the latest group, nine patients (70 %) brain was the only site of recurrence. In the other four (30 %), lung metastasis accompanied brain recurrence. Of these four, only one case had also other sites of metastases. Clinical features of patients according to the time of presentation of brain metastasis Patient characteristics, including histology of the primary tumor, distribution and number of additional metastatic sites and levels of tumor markers at diagnosis of primary testicular cancer are listed in Table 1. All patients were
Clin Transl Oncol Table 1 Demographic characteristics
Table 2 Clinical presentation of brain metastases
Group 1 (n = 19)
Group 2 (n = 13)
Age at diagnosis of TGCT
31 (18–53)
29 (20–38)
Testicle side
18
13
Left
7 (36.8 %)
4 (30.8 %)
PS 0
0
2 (16 %)
11 (57.9 %)
9 (69.2 %)
PS 1 PS 2
5 (26 %) 8 (42 %)
3 (23 %) 3 (23 %)
18 (94.7 %)
13 (100 %)
Right Location at primary tumor Gonadal
Extragonadal 1 (5.3 %) Tumors markers [at diagnosis of TGCT (median; range)]
Group 1 (synchronous) (n = 19)
Performance status at time of diagnoses of brain metastases
6 (32 %)
5 (38 %)
Number of brain lesions
PS 3
2 (1–10)
1 (1–4)
30 (5–140)
31 (25–55)
AFP (2 unknown)
1,165 (0–19,814)
3,832 (0–23,300)
Size highest lesion (mm)
BHCG (2 unknown)
279.083 (65–956,400)
15,790 (0–117,800)
Brain symptoms at diagnosis
LDH (4 unknown)
720 (6–2,176)
680 (10–2,652)
Histology primary tumor (predominant) Seminoma
1 (5.3 %)
0
Chorionic carcinoma
8 (42.1 %)
3 (23.1 %)
Embryonal carcinoma
2 (10.5 %)
8 (61.5 %)
Mature teratoma
1 (5.3 %)
0
Immature teratoma
2 (10.5 %)
0
TGNS Endodermic sinus
3 (15.8 %) 3 (15 %)
1 (7.7 %) 0
Burn out
0
1 (7.7 %)
Group 2 (metachronous) (n = 13)
Yes
13 (64 %)
No 5 (26 %) Histology of brain metastases
12 (92 %) 1 (8 %)
Unknown
17 (90 %)
8 (62 %)
Same than primary tumor
2 (10 %): embryonal carcinoma
5 (38 %): embryonal carcinoma
Liver
18 (94.7 %)
Spleen
12 (63.2 %)
1 (7.7 %)
Stomach
1 (5.3 %)
0
Skin
1 (5.3 %)
0
Mediastinum
1 (5.3 %)
4 (30.8 %)
patients (38.5 %) as intermediate prognosis and five patients (38.5 %) as poor prognosis. Table 2 shows clinical characteristics of brain metastasis for both groups. In both groups, poor performance status (PS 2–3) and presence of neurological symptoms were related with brain metastasis, with poor performance and more brain symptoms at group 2. Significantly, more patients in group 2 had solitary brain metastasis; in seven cases, brain was the only site of recurrence. In group 2, median time to brain metastases appearance was 6 months since last platinum-based doses (3–22) (Table 3).
3 (23 %)
Treatment by groups
Lung metastases at initial diagnosis of TGTC Yes
20 (100 %)
No
11 (84.6 %) 2 (15.4 %)
Additional metastatic sites at initial diagnosis 12 (92.3 %)
Prognosis according IGCCCG at initial diagnosis Good Intermediate Poor
5 (38.5 %) 20 (100 %)
5 (38.5 %)
classified according to the IGCCCG [4]. We found significantly higher serum levels of BHCG (p 0.0000445) and LDH (p 0.005) in group 1, whereas those patients with metachronous metastasis had higher serum levels of AFP, but those did not reach statistical significance. The most common histology in the primary tumor was chorionic carcinoma for group 1 versus embryonal carcinoma for group 2. Based on the International Germ Cell Consensus Classification (IGCCC) [4] at the time of initial diagnosis, 100 % of patients in the synchronic group fulfilled criteria for poor prognosis. For metachronic brain metastases, three patients (23 %) were classified as good prognosis, five
The treatment for synchronic cases comprises orchiectomy followed by chemotherapy in 18 patients; only 1 patient received chemotherapy without orchiectomy. Chemotherapy consisted of cisplatin-based combinations (8 BEP, 6 BOMP-EPI, 1 BEP followed by TIP, 1 EP, 1 BOMP, 1 TIP followed by high-dose chemotherapy with stem cell support). After chemotherapy, five patients (26 %) underwent surgery for residual masses excision; other five (26 %) received irradiation of the residual mass. Treatment of brain metastasis in group 1 Among patients with synchronic metastasis, in 13 of them (68.4 %) whole brain radiotherapy was done as part of their primary treatment, before or simultaneously with chemotherapy. Stereotactic radiotherapy was not employed in any
123
Clin Transl Oncol Table 3 Characteristics of brain metastases in group 2 Group 2 (metachronous) (n = 13) Tumors markers at diagnosis of brain metastases: (median) AFT
132
BHCG
11,465
LDH
167
Time from first diagnosis to brain metastases, months Median
9
Range
1–22
Time from last cisplatin dose to brain metastases, months Median
6
Range
1–22
Lung metastases at diagnosis of brain metastases Yes 4 (30.8 %) No
9 (69.2 %)
Treatment of brain metastases CT ? RT
4 (31 %)
CT
3 (23 %)
QT ? RT ? NS
3 (23 %)
RT
1 (8 %)
QT ? NS
2 (15 %)
Outcome RC
6 (46 %)
RPNM
2 (15 %)
RPPM
0
PD
5 (39 %)
Median survival
Median OS 23 months 38.5 % Alive disease-free
CT chemotherapy, RT radiotherapy, NS neurosurgery, RC complete respo
Fig. 2 Treatment of brain metastasis in both groups. CT chemotherapy, RT radiotherapy, NS neurosurgery
After finishing all treatment 26 % of patients achieved complete response, 37 % partial response with negative tumor markers, 11 % partial response with positive tumor markers and 26 % did not respond to treatment (Figs. 1, 2, 3). Treatment of primary tumor in group 2
Fig. 1 Induction treatment by groups. BEP belomycin, etoposide, cisplatin, BOMP-EPI bleomycin, vincristine, methotrexate, cisplatin followed by etoposide, ifosfamide, cisplatin, ATSP high-dose chemotherapy regimen with stem cell transplantation, TIP paclitaxel, ifosfamide, cisplatin, BOMP bleomycin, vincristine, methotrexate, cisplatin, EP etoposide, cisplatin
patient. Intrathecal therapy or regimens that contained standard or high-dose methotrexate had not been used. No patient was derived to neurosurgery.
123
For primary tumor, orchiectomy and chemotherapy was the main treatment. BEP schedule (62 %) and BOMP-EPI (31 %) were the most used regimens. Two patients received consolidation with high-dose chemotherapy and ATSP (14 %). Surgery for residual disease was done in three patients and radiotherapy in only one. 23 % had complete response after induction treatment, 61 % partial response with negative markers and only one (7 %) had no response to treatment. Treatment for brain metastasis in group 2 Nine patients had brain metastases as the single site of recurrence (70 %). Only three of them (33 %) were resected; after surgery, one received chemotherapy (TIP: cisplatin, ifosfamide and paclitaxel), one holocraneal
Clin Transl Oncol
Fig. 4 Situation at last follow-up. ADF alive disease free, AWD alive with disease, DDF death disease free, DWD death with disease Fig. 3 Response to treatment by groups. CR complete response, PRNM partial response negative markers, PRPM partial response positive markers, NR no reponse
radiotherapy and the other one was resected after chemotherapy (GEMOX: gemcitabine–oxaliplatin) and holocraneal radiotherapy. The first two achieved completed response and were alive without disease. The other six patients (67 %) were treated with chemotherapy (2 TIP, 2 ACE, 1 VIP); one of them also with radiotherapy. Only two were alive without disease at last follow-up. For the four patients with other metastases at brain recurrence, surgery was offered only for one. This achieve complete response and was alive without disease. The others were treated with chemotherapy (2 TIP, 1 GEMOX) and radiotherapy. Only one remains alive. In this second group, 46 % of patients achieve complete response, 16 % of partial response with negative tumor markers, and 38 % of patients without therapy response (Figs. 1, 2, 3). Survival according to synchronic or metachronic subgroups At last follow-up, 33 % of patients in group 1 and 36 % in group 2 were alive without disease. 56 % patients had died due to disease progression in group 1 and 66 % in group 2 (Fig. 4).
The median overall survival for all patients was 1.385 years (95 % CI 0.116–2.655). Patients in group 2 showed a trend towards better overall survival (1.91 years versus 1.18, p: 0.857) (Fig. 4). A 2-year survival rates were 37.5 and 38.9 % for groups 1 and 2, respectively (Fig. 5). Prognostic factors for survival Survival rates in the cases with multiple brain metastases were significantly lower than those cases with a single metastases. Univariate and multivariate analysis of prognostic factors for survival was performed. Although elevation of B-HCG levels, extent and sites of additional metastatic disease, and histology of the primary tumor were not of significant influence, the presence of a single brain metastasis, low number of brain metastases (\3), diagnosis of brain metastasis at initial diagnosis and achievement of complete response were significant predictors for better outcome. In the multivariate analysis/less than three brain lesions and the achievement of complete response were prognostic factors for survival. On group 2, we applied the new IGCCC classification for recurrence [5]. 46 % (6) of patients on intermediate group; 46 % (6) on high group and 1 (8 %) very high group. Median survival for intermediate group was 97 months, with 33 % (2) still alive. Median survival for high risk group was 8 months, 33 % (2) still alive. The only one in the very
123
Clin Transl Oncol Fig. 5 Comparison survival between group 1 and 2
high group was alive with 105 months of survival. Classification by risk group was significantly related with survival. Causes of death in patients with brain metastasis Nineteen patients (60 %) have died at the end of the study period. Based on the available clinical data extracted from patients’ charts, we attempted to determine whether a patient had died of systemic tumor progression or due to progressive brain disease or other causes. On group 1, in six patients, the death could be attributed to massive systemic progression (55 %); six to uncontrolled brain lesions (27 %), one due to toxicity (9 %), and one for another reasons without evidence of disease (9 %; cause not reported in the chart). In group 2, the distribution was four deaths by systemic progression (50 %) and also four (50 %) for brain progression.
Discussion The current report of 33 patients with brain metastases from testicular cancer represents the largest Spanish series reported in the literature [6]. In most patients, brain metastases were associated with widespread systemic disease, particularly lung metastases. In our series, 100 % in group 1 had lung metastases. This is in accordance with other reports, which identified advanced tumor stage and multiple or bulky lung metastases as risk factors for the development of brain metastases. In group 2, the presence or absence of lung metastases were unrelated
123
to predict brain recurrence. Highly elevated BHCG levels have been reported as predictors for the presence of brain metastases [3]. The median HCG level in our patients was 279,083 U/ml in group 1 and 11,465 U/ml in group 2 at time of diagnosis of brain metastasis. These high levels may result from chorionic carcinomatous elements within the germ cell tumor and may be associated with a tendency towards early hematogenous spread. However, embrionary carcinoma was most related with brain metastases in group 2. This demonstrated that brain metastases are not exclusively associated with this histologic subtype. More surprisingly, one patient with seminoma had brain metastases at diagnosis. Few reports have described patients with seminoma and brain metastases. Bokemeyer et al. reported two patients. In our work, patient with seminoma histology had brain metastases at diagnosis (group 1) and died of toxicity. Based on these findings, routine CT scans of the brain should be recommended for all patients [7]. In our experience, brain metastasis at diagnosis its most easily to predict. Clinical factors described (lung metastasis, high BHCG, histology, bad prognosis group, etc.) were reproduced in group 1. Patients in group 2 were more difficult to predict, although we have found some association with other metastases than lung, choriocarcinoma and almost all patients had neurologic symptoms and bad performance status. We have not collected if brain metastases were detected by clinical or brain study was part of extension study. The cornerstone of treatment in both groups was chemotherapy. All patients treated with chemotherapy received adequate cisplatin-based regimens; mainly BEP. The use of BOMP-EPI is based on a multicenter study
Clin Transl Oncol
performed by the Spanish Germ Cell Cancer Group (GG) between 1985 and 1995 [8]. The BOMP-EPI regimen is an intensive alternating regimen with shortened intervals between the BOMP and EPI cycles and the subsequent increase in dose intensity for cisplatin. This protocol was applied to patients with poor prognosis. Thirty-eight patients were included, 5 % with brain metastases. The study found an estimated percentage of progression-free survival at 2 years of 58 % (95 % CI 42–74 %) and overall survival at 2 years of 64 % (95 % CI 49–80 %). Other series also reported multitude of chemotherapeutic schedules [2]; so it does not allow any meaningful comparison of the different chemotherapy combinations with regard to efficacy. The type of chemotherapeutic treatment for patients with testicular cancer and brain metastases has been the subject of controversy. It seems that almost all patients received chemotherapy plus another treatment strategy (radiotherapy or surgery). It is difficult to know which combination of modalities could be the best for these patients, due the nature of the case selection review. Our perception is that a single treatment modality could result in a worse outcome. The current series suggests an important beneficial role for neurosurgery and cerebral radiotherapy in patients who develop brain metastases after cisplatin-based induction chemotherapy; Fossa et al. [2] conclusions were similar for this group of patients. The main reason of death was systemic progression, supporting the use of systemic disease. However, randomized trials to compare different treatment modalities in this rare clinical situation will not likely not be performed. Patients with brain metastases belong to the worst prognostic groups [4, 5]. According to the results from investigators at Indiana University and others [9–13], patients who present with brain metastases of testicular cancer at initial diagnosis and who have not received prior systemic treatment are considered to have the best prognosis. Among our 19 patients in group 1, 31 % have survived more than 2 years after treatment. In contrast, patients who developed brain metastases after cisplatin-based chemotherapy have poor prognosis. At our series, surprisingly, this group had better prognosis, although no significant, and a higher percent had complete response. Patients in this group had brain as the only site of recurrence in a higher percentage that previously reported. This selection favors the use of local treatment in this clinical condition and, conversely, represents a ‘‘good prognosis’’ bias. Most of them were treated with systemic and also local therapy (radiotherapy or neurosurgery) and most than 1/3 could achieve complete response. If these patients had been
selected and we have a selection bias could not be excluded. However, it seems an opportunity to aggressive treatment in this bap prognosis group of patients. Bokemeyer et al. [3] excluded patients who relapse only with brain metastases. Our series confirms that more than 30–35 % of these patients can achieve long-term disease-free survival with adequate aggressive treatment of both brain metastases and systemic disease. Our series had the highest complete response rate and long-term survival reported at time on literature. Other series shown long-term survivors between 23 and 31 % [2]. However, series that reported longer percentage of patients alive included only 10 patients and patients with brain metastasis at diagnosis. In conclusions, long-term survival can be achieved in approximately 1/3 of patients with brain metastases. Chemotherapy remains the cornerstone of treatment. Owing to the retrospective nature of analysis, the results should be interpreted with caution. Conflict of interest
The authors declare no conflict of interest.
References 1. Raj S, Parkinson C, Williams M, Mazhar D. Management of brain metastases from germ cell tumors: do we know what we are doing? Future Oncol. 2008;4(1):1–4. 2. Fossa SD, Bokemeyer C, Gerl A, Culine S, Jones WG, Mead GM, et al. Treatment outcome of patients with brain metastases from malignant germ cell tumors. Cancer. 1999;85(4):988–97. 3. Bokemeyer C, Nowak P, Haupt A, Metzner B, Ko¨hne H, Hartmann JT, et al. Treatment of brain metastases in patients with testicular cancer. J Clin Oncol. 1997;15(4):1449–54. 4. International Germ Cell Cancer Collaborative Group. International germ cell consensus classification: a prognostic factor-based staging system for metastatic germ cell cancers. J Clin Oncol. 1997;15(2):594–603. 5. International Prognostic Factors Study Group, Lorch A, Beyer J, BascoulMollevi C, Kramar A, Einhorn LH, Necchi A, et al. Prognostic factors in patients with metastatic germ cell tumors who experienced treatment failure with cisplatin-based first-line chemotherapy. J Clin Oncol. 2010;28(33): 4906–11. 6. Rustin JS, Newlands S, Begent RH, Dent J, Bagshawe KD. Weekly alternating etoposide, methotrexate and actinomycin/vincristine and cyclophosphamide chemotherapy for the treatment of CNS metastases of choriocarcinoma. J Clin Oncol. 1989;7:900–3. 7. Gerl A, Clemm Ch, Kohl P, Schalhorn A, Wilmanns W. Central nervous system as sanctuary site of relapse in patients treated with chemotherapy for metastatic testicular cancer. Clin Expert Metastasis. 1994;12:226–30. 8. Germa`-Lluch JR, Garcı´a del Muro X, Tabernero JM, Sa´nchez M, Aparicio J, Alba E, et al. BOMP/EPI intensive alternating chemotherapy for IGCCC poorprognosis germ-cell tumors: The Spanish Germ-Cell Cancer Group experience (GG). Ann Oncol. 1999;10:289–93. 9. Lester SG, Morphis JG 2nd, Hornback NB, Williams SD, Einhorn LH. Brain metastases and testicular tumors: need for aggressive therapy. J Clin Oncol. 1984;2:1397–403. 10. Nonomura N, Nagahara A, Oka D, Mukai M, Nakai Y, Nakayama M, et al. Brain metastases from testicular germ cell tumors: a retrospective analysis. Int J Urol. 2009;16:887–93. 11. Stolinsky DC. Prolonged survival after cerebral metastasis of testicular carcinoma. Cancer. 1981;47:978–81. 12. Williams SD, Einhorn LH. Brain metastases in disseminated germinal neoplasms, incidence and clinical course. Cancer. 1979;44:1514–6. 13. Spears T, Morphis G 2nd, Lester G, Williams SD, Einhorn LH. Brain metastases and testicular tumors: long-term survival. Int J Radiat Oncol Biol Phys. 1991;32:17–22.
123
ORIGINAL ARTICLE
Synchronous versus metachronous brain metastasis from testicular germ cell tumors (TGCT): an analysis from the Spanish Germ Cell Cancer Group data base R. Girones • J. Aparicio • P. Roure • J. R. Germa-Lluch • X. Garcı´a del Muro • S. Vazquez-Estevez • A. Saenz • J. Sastre • J. Arranz Arija • E. Gallardo • E. Gonzalez-Billalabeitia • A. Sanchez-Hernandez • J. Terrasa • A. Hernandez • C. Santander • E. Cillan • N. Sagastibelza • D. Almenar-Cubells • M. Lopez Brea J. P. Maroto • Spanish Germ Cell Cancer Group (SGCCG)
•
Received: 1 February 2014 / Accepted: 24 March 2014 Ó Federacio´n de Sociedades Espan˜olas de Oncologı´a (FESEO) 2014
Abstract Background Brain metastases of testicular germ cell tumor (TGCT) are a rare event. Prognostic is poor and there is not much evidence on optimal management of these patients. Patients and methods A review of case records of germ cell tumor patients within the Spanish Germ Cell Cancer Group data base from 1994 to 2012 was conducted. Results Thirty-three out of 6,200 cases (0.5 %). Nineteen patients (57 %) group 1: synchronous, 13 (40 %) group 2: metachronous and only one developed brain metastasis during cisplatin-based chemotherapy (excluded from the analysis). Median serum BHCG levels at initial diagnosis was higher in group 1, whereas elevated AFP serum levels were more common in group 2. Histology in the primary tumor: chorionic carcinoma for group 1 versus embryonal
carcinoma for group 2. Mainly solitary brain metastasis in group 2 (54 versus 21 %, respectively). The median overall survival from the diagnosis of central nervous system involvement was 16 months for group 1 (CI 95 % 13.9–18) and 23 months (95 % CI 0–165) for group 2 (log rank p = 0.84). Long-term survivors were practically identical in the two groups (38.9 % group 1 versus 38.5 % group 2). Regardless of the timing of brain metastasis, those patients that achieved complete response to the treatment had better survival (log rank p 0.003). Conclusion Although some distinctive clinical characteristics have been found between patients with synchronous versus metachronous brain metastasis from TGCT, the timing of brain metastasis did not seem to have prognostic influence, but due to the retrospective nature of the analysis and the results should be interpreted with caution.
R. Girones (&) Hospital Lluis Alcanyis Xa`tiva, Valencia, Spain e-mail: [email protected]
J. Arranz Arija Hospital Gregorio Maran˜on Madrid, Madrid, Spain
J. Aparicio Hospital La FE Valencia, Valencia, Spain P. Roure Hospital de Vic, Barcelona, Spain J. R. Germa-Lluch X. Garcı´a del Muro Institut Catala` D’Oncologia, Barcelona, Spain S. Vazquez-Estevez Hospital Lucus Augusti Lugo, Lugo, Spain
E. Gallardo Hospital Parc Taulı´, Sabadell, Spain E. Gonzalez-Billalabeitia Hospital Morales Meseguer, Murcia, Spain A. Sanchez-Hernandez Hospital Provincial de Castello´n, Castello´n, Spain J. Terrasa Hospital Son Espases, Palma, Spain
A. Saenz Hospital Clı´nico de Zaragoza, Saragossa, Spain
A. Hernandez Instituto Oncologico Guipuzcoa San Sebastia´n, San Sebastia´n, Spain
J. Sastre Hospital Clı´nico San Carlos Madrid, Madrid, Spain
C. Santander J. P. Maroto Hospital Sant Pau Barcelona, Barcelona, Spain
123
Clin Transl Oncol
Keywords Germ cell tumors Brain metastases Chemotherapy Neurosurgery Holocraneal radiotherapy Prognostic factors Survival Epidemiology Data base
Introduction Brain metastases are a rare phenomenon in germ cell tumors (GCTs), occurring in \2–3 % of all patients and in approximately 10 % of patients with advanced GCTs [1]. However, postmortem analysis series revealed that up to 40 % of patients who died for progressive disease had nonfounded brain metastasis [2]. Most of the available data about clinical implications and management of brain metastasis from TGCT come from small case series [3]. To add information for these selected patients, we evaluated the clinical characteristics, treatment and outcome of patients with brain metastases of testicular cancer from the Spanish Germ Cell Cancer Group data-base registry.
achieving response to previous cisplatin-based chemotherapy. The only case primary resistant to cisplatin-based chemotherapy who developed brain metastasis during the chemotherapeutic treatment was excluded of this analysis. Patients were treated according to the gold standard for systemic and brain locations throughout the study period. Statistical analysis This was a descriptive study based on date prospectively collected. It was designed to identify outcomes and managements of TGCT with brain metastases. SPSS (Statistical Package for the Social Sciences) version 15.0 was used for all of the analyses. For quantitative variables, measurements of central tendency (mean and median) and measures of spread (standard deviation and interquartile range) were used. For qualitative variables, absolute and relative frequency tables were employed. Graphical representation of the data obtained was used wherever possible. Mann– Whitney U test and Fisher’s exact test were used to assess the significance of the findings. Differences with a p \ 0.05 were considered significant.
Patients and methods Ethical considerations Methods This is an analysis of patients with brain metastasis registered prospectively in the data base of the Spanish Germ Cell Cancer Group registry. Our data base comprises diagnosis, clinical management and survival data of almost 6,200 patients collected since 1994.
The institution’s ethical review board approved this study. All of the patients provided written informed consent to be included in Germinal Spanish Germ Cancer Group data base at the time of diagnosis of testicular germ cell cancer, given their permission to collect their clinical data for future scientific purposes.
Patients Results A data collection form was sent to the investigators to clarify or implement confusing or missing data, no previously reported in the data base. Median follow-up was 16 months (1–228). All patients were classified into three groups based on the time of brain metastases diagnosis: group 1 included those patients who presented with brain metastases at the time of diagnosis of primary testicular cancer; group 2 was composed of patients who developed brain metastases with or without systemic disease after E. Cillan Hospital Servet Zaragoza, Saragossa, Spain N. Sagastibelza Hospital de Donostia, San Sebastia´n, Spain D. Almenar-Cubells Hospital Doctor Peset Valencia, Valencia, Spain M. Lopez Brea Hospital Ma´rques de Valdecilla, Santander, Spain
123
A total of 42 out of 6200 (0.6 %) cases registered in our data base developed brain metastasis. Reliable data were obtained from only 33 patients; in 19 patients (57 %), brain metastasis were identified at the time of diagnosis of primary tumor (group 1), whereas 13 patients (40 %) were identified at the time of recurrence (group 2). Among the latest group, nine patients (70 %) brain was the only site of recurrence. In the other four (30 %), lung metastasis accompanied brain recurrence. Of these four, only one case had also other sites of metastases. Clinical features of patients according to the time of presentation of brain metastasis Patient characteristics, including histology of the primary tumor, distribution and number of additional metastatic sites and levels of tumor markers at diagnosis of primary testicular cancer are listed in Table 1. All patients were
Clin Transl Oncol Table 1 Demographic characteristics
Table 2 Clinical presentation of brain metastases
Group 1 (n = 19)
Group 2 (n = 13)
Age at diagnosis of TGCT
31 (18–53)
29 (20–38)
Testicle side
18
13
Left
7 (36.8 %)
4 (30.8 %)
PS 0
0
2 (16 %)
11 (57.9 %)
9 (69.2 %)
PS 1 PS 2
5 (26 %) 8 (42 %)
3 (23 %) 3 (23 %)
18 (94.7 %)
13 (100 %)
Right Location at primary tumor Gonadal
Extragonadal 1 (5.3 %) Tumors markers [at diagnosis of TGCT (median; range)]
Group 1 (synchronous) (n = 19)
Performance status at time of diagnoses of brain metastases
6 (32 %)
5 (38 %)
Number of brain lesions
PS 3
2 (1–10)
1 (1–4)
30 (5–140)
31 (25–55)
AFP (2 unknown)
1,165 (0–19,814)
3,832 (0–23,300)
Size highest lesion (mm)
BHCG (2 unknown)
279.083 (65–956,400)
15,790 (0–117,800)
Brain symptoms at diagnosis
LDH (4 unknown)
720 (6–2,176)
680 (10–2,652)
Histology primary tumor (predominant) Seminoma
1 (5.3 %)
0
Chorionic carcinoma
8 (42.1 %)
3 (23.1 %)
Embryonal carcinoma
2 (10.5 %)
8 (61.5 %)
Mature teratoma
1 (5.3 %)
0
Immature teratoma
2 (10.5 %)
0
TGNS Endodermic sinus
3 (15.8 %) 3 (15 %)
1 (7.7 %) 0
Burn out
0
1 (7.7 %)
Group 2 (metachronous) (n = 13)
Yes
13 (64 %)
No 5 (26 %) Histology of brain metastases
12 (92 %) 1 (8 %)
Unknown
17 (90 %)
8 (62 %)
Same than primary tumor
2 (10 %): embryonal carcinoma
5 (38 %): embryonal carcinoma
Liver
18 (94.7 %)
Spleen
12 (63.2 %)
1 (7.7 %)
Stomach
1 (5.3 %)
0
Skin
1 (5.3 %)
0
Mediastinum
1 (5.3 %)
4 (30.8 %)
patients (38.5 %) as intermediate prognosis and five patients (38.5 %) as poor prognosis. Table 2 shows clinical characteristics of brain metastasis for both groups. In both groups, poor performance status (PS 2–3) and presence of neurological symptoms were related with brain metastasis, with poor performance and more brain symptoms at group 2. Significantly, more patients in group 2 had solitary brain metastasis; in seven cases, brain was the only site of recurrence. In group 2, median time to brain metastases appearance was 6 months since last platinum-based doses (3–22) (Table 3).
3 (23 %)
Treatment by groups
Lung metastases at initial diagnosis of TGTC Yes
20 (100 %)
No
11 (84.6 %) 2 (15.4 %)
Additional metastatic sites at initial diagnosis 12 (92.3 %)
Prognosis according IGCCCG at initial diagnosis Good Intermediate Poor
5 (38.5 %) 20 (100 %)
5 (38.5 %)
classified according to the IGCCCG [4]. We found significantly higher serum levels of BHCG (p 0.0000445) and LDH (p 0.005) in group 1, whereas those patients with metachronous metastasis had higher serum levels of AFP, but those did not reach statistical significance. The most common histology in the primary tumor was chorionic carcinoma for group 1 versus embryonal carcinoma for group 2. Based on the International Germ Cell Consensus Classification (IGCCC) [4] at the time of initial diagnosis, 100 % of patients in the synchronic group fulfilled criteria for poor prognosis. For metachronic brain metastases, three patients (23 %) were classified as good prognosis, five
The treatment for synchronic cases comprises orchiectomy followed by chemotherapy in 18 patients; only 1 patient received chemotherapy without orchiectomy. Chemotherapy consisted of cisplatin-based combinations (8 BEP, 6 BOMP-EPI, 1 BEP followed by TIP, 1 EP, 1 BOMP, 1 TIP followed by high-dose chemotherapy with stem cell support). After chemotherapy, five patients (26 %) underwent surgery for residual masses excision; other five (26 %) received irradiation of the residual mass. Treatment of brain metastasis in group 1 Among patients with synchronic metastasis, in 13 of them (68.4 %) whole brain radiotherapy was done as part of their primary treatment, before or simultaneously with chemotherapy. Stereotactic radiotherapy was not employed in any
123
Clin Transl Oncol Table 3 Characteristics of brain metastases in group 2 Group 2 (metachronous) (n = 13) Tumors markers at diagnosis of brain metastases: (median) AFT
132
BHCG
11,465
LDH
167
Time from first diagnosis to brain metastases, months Median
9
Range
1–22
Time from last cisplatin dose to brain metastases, months Median
6
Range
1–22
Lung metastases at diagnosis of brain metastases Yes 4 (30.8 %) No
9 (69.2 %)
Treatment of brain metastases CT ? RT
4 (31 %)
CT
3 (23 %)
QT ? RT ? NS
3 (23 %)
RT
1 (8 %)
QT ? NS
2 (15 %)
Outcome RC
6 (46 %)
RPNM
2 (15 %)
RPPM
0
PD
5 (39 %)
Median survival
Median OS 23 months 38.5 % Alive disease-free
CT chemotherapy, RT radiotherapy, NS neurosurgery, RC complete respo
Fig. 2 Treatment of brain metastasis in both groups. CT chemotherapy, RT radiotherapy, NS neurosurgery
After finishing all treatment 26 % of patients achieved complete response, 37 % partial response with negative tumor markers, 11 % partial response with positive tumor markers and 26 % did not respond to treatment (Figs. 1, 2, 3). Treatment of primary tumor in group 2
Fig. 1 Induction treatment by groups. BEP belomycin, etoposide, cisplatin, BOMP-EPI bleomycin, vincristine, methotrexate, cisplatin followed by etoposide, ifosfamide, cisplatin, ATSP high-dose chemotherapy regimen with stem cell transplantation, TIP paclitaxel, ifosfamide, cisplatin, BOMP bleomycin, vincristine, methotrexate, cisplatin, EP etoposide, cisplatin
patient. Intrathecal therapy or regimens that contained standard or high-dose methotrexate had not been used. No patient was derived to neurosurgery.
123
For primary tumor, orchiectomy and chemotherapy was the main treatment. BEP schedule (62 %) and BOMP-EPI (31 %) were the most used regimens. Two patients received consolidation with high-dose chemotherapy and ATSP (14 %). Surgery for residual disease was done in three patients and radiotherapy in only one. 23 % had complete response after induction treatment, 61 % partial response with negative markers and only one (7 %) had no response to treatment. Treatment for brain metastasis in group 2 Nine patients had brain metastases as the single site of recurrence (70 %). Only three of them (33 %) were resected; after surgery, one received chemotherapy (TIP: cisplatin, ifosfamide and paclitaxel), one holocraneal
Clin Transl Oncol
Fig. 4 Situation at last follow-up. ADF alive disease free, AWD alive with disease, DDF death disease free, DWD death with disease Fig. 3 Response to treatment by groups. CR complete response, PRNM partial response negative markers, PRPM partial response positive markers, NR no reponse
radiotherapy and the other one was resected after chemotherapy (GEMOX: gemcitabine–oxaliplatin) and holocraneal radiotherapy. The first two achieved completed response and were alive without disease. The other six patients (67 %) were treated with chemotherapy (2 TIP, 2 ACE, 1 VIP); one of them also with radiotherapy. Only two were alive without disease at last follow-up. For the four patients with other metastases at brain recurrence, surgery was offered only for one. This achieve complete response and was alive without disease. The others were treated with chemotherapy (2 TIP, 1 GEMOX) and radiotherapy. Only one remains alive. In this second group, 46 % of patients achieve complete response, 16 % of partial response with negative tumor markers, and 38 % of patients without therapy response (Figs. 1, 2, 3). Survival according to synchronic or metachronic subgroups At last follow-up, 33 % of patients in group 1 and 36 % in group 2 were alive without disease. 56 % patients had died due to disease progression in group 1 and 66 % in group 2 (Fig. 4).
The median overall survival for all patients was 1.385 years (95 % CI 0.116–2.655). Patients in group 2 showed a trend towards better overall survival (1.91 years versus 1.18, p: 0.857) (Fig. 4). A 2-year survival rates were 37.5 and 38.9 % for groups 1 and 2, respectively (Fig. 5). Prognostic factors for survival Survival rates in the cases with multiple brain metastases were significantly lower than those cases with a single metastases. Univariate and multivariate analysis of prognostic factors for survival was performed. Although elevation of B-HCG levels, extent and sites of additional metastatic disease, and histology of the primary tumor were not of significant influence, the presence of a single brain metastasis, low number of brain metastases (\3), diagnosis of brain metastasis at initial diagnosis and achievement of complete response were significant predictors for better outcome. In the multivariate analysis/less than three brain lesions and the achievement of complete response were prognostic factors for survival. On group 2, we applied the new IGCCC classification for recurrence [5]. 46 % (6) of patients on intermediate group; 46 % (6) on high group and 1 (8 %) very high group. Median survival for intermediate group was 97 months, with 33 % (2) still alive. Median survival for high risk group was 8 months, 33 % (2) still alive. The only one in the very
123
Clin Transl Oncol Fig. 5 Comparison survival between group 1 and 2
high group was alive with 105 months of survival. Classification by risk group was significantly related with survival. Causes of death in patients with brain metastasis Nineteen patients (60 %) have died at the end of the study period. Based on the available clinical data extracted from patients’ charts, we attempted to determine whether a patient had died of systemic tumor progression or due to progressive brain disease or other causes. On group 1, in six patients, the death could be attributed to massive systemic progression (55 %); six to uncontrolled brain lesions (27 %), one due to toxicity (9 %), and one for another reasons without evidence of disease (9 %; cause not reported in the chart). In group 2, the distribution was four deaths by systemic progression (50 %) and also four (50 %) for brain progression.
Discussion The current report of 33 patients with brain metastases from testicular cancer represents the largest Spanish series reported in the literature [6]. In most patients, brain metastases were associated with widespread systemic disease, particularly lung metastases. In our series, 100 % in group 1 had lung metastases. This is in accordance with other reports, which identified advanced tumor stage and multiple or bulky lung metastases as risk factors for the development of brain metastases. In group 2, the presence or absence of lung metastases were unrelated
123
to predict brain recurrence. Highly elevated BHCG levels have been reported as predictors for the presence of brain metastases [3]. The median HCG level in our patients was 279,083 U/ml in group 1 and 11,465 U/ml in group 2 at time of diagnosis of brain metastasis. These high levels may result from chorionic carcinomatous elements within the germ cell tumor and may be associated with a tendency towards early hematogenous spread. However, embrionary carcinoma was most related with brain metastases in group 2. This demonstrated that brain metastases are not exclusively associated with this histologic subtype. More surprisingly, one patient with seminoma had brain metastases at diagnosis. Few reports have described patients with seminoma and brain metastases. Bokemeyer et al. reported two patients. In our work, patient with seminoma histology had brain metastases at diagnosis (group 1) and died of toxicity. Based on these findings, routine CT scans of the brain should be recommended for all patients [7]. In our experience, brain metastasis at diagnosis its most easily to predict. Clinical factors described (lung metastasis, high BHCG, histology, bad prognosis group, etc.) were reproduced in group 1. Patients in group 2 were more difficult to predict, although we have found some association with other metastases than lung, choriocarcinoma and almost all patients had neurologic symptoms and bad performance status. We have not collected if brain metastases were detected by clinical or brain study was part of extension study. The cornerstone of treatment in both groups was chemotherapy. All patients treated with chemotherapy received adequate cisplatin-based regimens; mainly BEP. The use of BOMP-EPI is based on a multicenter study
Clin Transl Oncol
performed by the Spanish Germ Cell Cancer Group (GG) between 1985 and 1995 [8]. The BOMP-EPI regimen is an intensive alternating regimen with shortened intervals between the BOMP and EPI cycles and the subsequent increase in dose intensity for cisplatin. This protocol was applied to patients with poor prognosis. Thirty-eight patients were included, 5 % with brain metastases. The study found an estimated percentage of progression-free survival at 2 years of 58 % (95 % CI 42–74 %) and overall survival at 2 years of 64 % (95 % CI 49–80 %). Other series also reported multitude of chemotherapeutic schedules [2]; so it does not allow any meaningful comparison of the different chemotherapy combinations with regard to efficacy. The type of chemotherapeutic treatment for patients with testicular cancer and brain metastases has been the subject of controversy. It seems that almost all patients received chemotherapy plus another treatment strategy (radiotherapy or surgery). It is difficult to know which combination of modalities could be the best for these patients, due the nature of the case selection review. Our perception is that a single treatment modality could result in a worse outcome. The current series suggests an important beneficial role for neurosurgery and cerebral radiotherapy in patients who develop brain metastases after cisplatin-based induction chemotherapy; Fossa et al. [2] conclusions were similar for this group of patients. The main reason of death was systemic progression, supporting the use of systemic disease. However, randomized trials to compare different treatment modalities in this rare clinical situation will not likely not be performed. Patients with brain metastases belong to the worst prognostic groups [4, 5]. According to the results from investigators at Indiana University and others [9–13], patients who present with brain metastases of testicular cancer at initial diagnosis and who have not received prior systemic treatment are considered to have the best prognosis. Among our 19 patients in group 1, 31 % have survived more than 2 years after treatment. In contrast, patients who developed brain metastases after cisplatin-based chemotherapy have poor prognosis. At our series, surprisingly, this group had better prognosis, although no significant, and a higher percent had complete response. Patients in this group had brain as the only site of recurrence in a higher percentage that previously reported. This selection favors the use of local treatment in this clinical condition and, conversely, represents a ‘‘good prognosis’’ bias. Most of them were treated with systemic and also local therapy (radiotherapy or neurosurgery) and most than 1/3 could achieve complete response. If these patients had been
selected and we have a selection bias could not be excluded. However, it seems an opportunity to aggressive treatment in this bap prognosis group of patients. Bokemeyer et al. [3] excluded patients who relapse only with brain metastases. Our series confirms that more than 30–35 % of these patients can achieve long-term disease-free survival with adequate aggressive treatment of both brain metastases and systemic disease. Our series had the highest complete response rate and long-term survival reported at time on literature. Other series shown long-term survivors between 23 and 31 % [2]. However, series that reported longer percentage of patients alive included only 10 patients and patients with brain metastasis at diagnosis. In conclusions, long-term survival can be achieved in approximately 1/3 of patients with brain metastases. Chemotherapy remains the cornerstone of treatment. Owing to the retrospective nature of analysis, the results should be interpreted with caution. Conflict of interest
The authors declare no conflict of interest.
References 1. Raj S, Parkinson C, Williams M, Mazhar D. Management of brain metastases from germ cell tumors: do we know what we are doing? Future Oncol. 2008;4(1):1–4. 2. Fossa SD, Bokemeyer C, Gerl A, Culine S, Jones WG, Mead GM, et al. Treatment outcome of patients with brain metastases from malignant germ cell tumors. Cancer. 1999;85(4):988–97. 3. Bokemeyer C, Nowak P, Haupt A, Metzner B, Ko¨hne H, Hartmann JT, et al. Treatment of brain metastases in patients with testicular cancer. J Clin Oncol. 1997;15(4):1449–54. 4. International Germ Cell Cancer Collaborative Group. International germ cell consensus classification: a prognostic factor-based staging system for metastatic germ cell cancers. J Clin Oncol. 1997;15(2):594–603. 5. International Prognostic Factors Study Group, Lorch A, Beyer J, BascoulMollevi C, Kramar A, Einhorn LH, Necchi A, et al. Prognostic factors in patients with metastatic germ cell tumors who experienced treatment failure with cisplatin-based first-line chemotherapy. J Clin Oncol. 2010;28(33): 4906–11. 6. Rustin JS, Newlands S, Begent RH, Dent J, Bagshawe KD. Weekly alternating etoposide, methotrexate and actinomycin/vincristine and cyclophosphamide chemotherapy for the treatment of CNS metastases of choriocarcinoma. J Clin Oncol. 1989;7:900–3. 7. Gerl A, Clemm Ch, Kohl P, Schalhorn A, Wilmanns W. Central nervous system as sanctuary site of relapse in patients treated with chemotherapy for metastatic testicular cancer. Clin Expert Metastasis. 1994;12:226–30. 8. Germa`-Lluch JR, Garcı´a del Muro X, Tabernero JM, Sa´nchez M, Aparicio J, Alba E, et al. BOMP/EPI intensive alternating chemotherapy for IGCCC poorprognosis germ-cell tumors: The Spanish Germ-Cell Cancer Group experience (GG). Ann Oncol. 1999;10:289–93. 9. Lester SG, Morphis JG 2nd, Hornback NB, Williams SD, Einhorn LH. Brain metastases and testicular tumors: need for aggressive therapy. J Clin Oncol. 1984;2:1397–403. 10. Nonomura N, Nagahara A, Oka D, Mukai M, Nakai Y, Nakayama M, et al. Brain metastases from testicular germ cell tumors: a retrospective analysis. Int J Urol. 2009;16:887–93. 11. Stolinsky DC. Prolonged survival after cerebral metastasis of testicular carcinoma. Cancer. 1981;47:978–81. 12. Williams SD, Einhorn LH. Brain metastases in disseminated germinal neoplasms, incidence and clinical course. Cancer. 1979;44:1514–6. 13. Spears T, Morphis G 2nd, Lester G, Williams SD, Einhorn LH. Brain metastases and testicular tumors: long-term survival. Int J Radiat Oncol Biol Phys. 1991;32:17–22.
123
