Treatment of epidural spinal cord involvement from germ cell tumors with chemotherapy


  • Presented in part at the American Society of Clinical Oncology Annual Meeting, Chicago, Illinois, June 4-8, 2010.



Germ cell tumors (GCTs) are chemosensitive, and epidural spinal cord compression (ESCC) from GCT may be amenable to treatment with chemotherapy (CT) only. This retrospective study compares the clinical outcome of GCT patients with ESCC treated with CT or radiotherapy (RT) + CT.


All patients with a histologic diagnosis of GCT from 1984 to 2009 were included in this study. Patients with ESCC were identified. Age, clinical features, histology, treatment, and outcome were analyzed.


The authors identified 1734 patients with GCT, of whom 29 (1.7%) had ESCC. The median age of these 29 patients was 32 years. The ESCC was treated with CT only in 16, RT + CT in 11, and 2 patients received palliative care only. The ESCC was more extensive in the RT + CT than the CT group. Patients who received RT + CT had a higher proportion of failed prior CT regimens, a higher percentage of nonseminomatous GCT, T-spine involvement, multilevel epidural disease, and bony vertebral metastases. Median overall survival after diagnosis of ESCC was not reached for those treated with CT alone versus 15 months for those receiving RT + CT (P = .02). There was also a significant difference in survival in patients receiving first-line therapy (n = 15), where median overall survival was not reached in the CT group (n = 11), compared with 22 months in the RT group (n = 4) (P = .04).


GCTs rarely involve the epidural compartment. Patients with ESCC who are likely to have chemosensitive disease can receive CT alone as definitive treatment. Cancer 2011. © 2010 American Cancer Society.

Germ cell tumors (GCTs) are uncommon, accounting for only 1% to 2% of all malignancies in men, but are the most frequent malignancy in men between the ages of 15 and 35 years.1, 2 Vertebral bone metastasis, spinal cord compression, or even spinal cord involvement by GCT is rare and only described in case reports for seminomas3-9 and nonseminomatous germ cell tumors (NSGCTs).10-14 In most other solid tumors, epidural spinal cord compression (ESCC) is treated with high-dose corticosteroids with radiation therapy (RT) or surgical decompression. However, ESCC from GCT may be treated effectively by chemotherapy (CT) only.15 The aim of this retrospective analysis is to describe a cohort of patients with ESCC from GCT and to compare survival and neurologic outcome after treating these patients with CT alone in comparison with RT followed by CT (RT + CT).


The Memorial Sloan-Kettering Cancer Center clinical database was used to identify patients with a histologic diagnosis of GCT (search terms: seminoma, NOS; seminoma-metastatic NOS; seminoma, anaplastic type; seminoma-anaplastic, meta; spermatocytic seminoma; germ cell tumor, nonseminomatous; mixed germ cell tumor; teratoma; choriocarcinoma; choriocarcinoma-met) from 1984 to 2009. Patients treated before 1984 with CT alone were already described in an earlier publication and not included in this study.15 Patients with ESCC were identified by the International Classification of Diseases, Ninth Edition code. Age, race/ethnicity, clinical features, histology, treatment, and outcome were analyzed.

The White spinal metastatic cancer grade16 (I, ambulatory; II, nonambulatory but some motor function; III, paraplegia with no motor function) and the Constans spinal metastases classification17 (I, pain or minor neurological symptoms with normal social and professional activities; II, mild neurological symptoms with normal life but interruption of professional activities; III, moderate neurological symptoms like paraparesis, sphincter disturbances, columnar pain with active life possible; IV, serious neurological symptoms like paraplegia and complete sphincter deficit; V, medullary syndrome of spinal cord transection) were used to quantify ESCC-related clinical deficits retrospectively. These grades were assigned based on the treating neurologist's examination done at ESCC diagnosis and during the patient's treatment. The diagnostic spinal magnetic resonance imaging (MRI) scan was reviewed for all patients for whom films were available, and the degree of ESCC was rated on a scale from 0 to 3, based on axial T2-weighted magnetic resonance images; this scale assesses the degree of subarachnoid space obliteration and ESCC and has been used previously.18

Fisher exact test was used to calculate significant levels for tumor type, ESCC location and severity on MRI, vertebral versus bone metastases, multilevel disease, prior failed CT, clinical grading scale, and improvement of ESCC on MRI after treatment. Survival was estimated using the Kaplan-Meier method and was compared between patients receiving CT alone versus RT + CT using the log-rank test. This study was approved by the Memorial Sloan-Kettering Cancer Center Institutional Review Board.


We identified 1734 patients with GCT who met study criteria. In the course of their disease, 393 (23%) patients developed lung metastases, 109 (6%) mediastinal metastases, 147 (8%) gastrointestinal metastases, 83 (5%) bone metastases, and 29 (1.7%) ESCC, of whom 13 had ESCC at presentation. The median age of these 29 patients was 32 years (range, 14-59 years) (Table 1). Ten patients had a seminoma and 19 a NSGCT. The ESCC was treated with CT in 16 patients, 11 received RT + CT, and 2 patients received palliation only (Table 1). Of the 27 patients who received active treatment for ESCC, films were available for review in 23 patients, and imaging reports confirmed ESCC in the remaining 4. Four had ESCC located in the cervical spine, 14 in the thoracic spine, and 9 in the lumbar spine. In 5 of the 27 patients, additional multilevel epidural disease without ESCC was also present (Table 2). Seventy-six percent of patients had severe ESCC (grade 2-3), 41% had vertebral metastases, and 44% had a paravertebral mass invading the neuroforamen causing ESCC (Table 3).

Table 1. Patient Characteristics
CharacteristicAll Patients (%)CTRT+CTPalliative
  1. CT indicates chemotherapy; RT, radiotherapy; NSGCT, nonseminomatous germ cell tumor.

Patients, No.2916112
Age median y, range32 (14-59)33 (18-45)32 (14-59)31 (28-34)
Histology (%)    
 Seminoma10 (34)7 (44)3 (27)0
 NSGCT19 (66)9 (56)8 (73)2 (100)
Table 2. Clinical Characteristics
CharacteristicAll Patients (%)CTRT+CT
  1. CT indicates chemotherapy; RT, radiotherapy; ESCC, epidural spinal cord compression.

Patients, No.271611
Median time to ESCC, mo151224
Clinical signs   
 Back pain26 (96)15 (94)11 (100)
 Ambulatory23 (85)14 (88)9 (82)
 Motor deficits10 (37)6 (38)4 (36)
 Sensory deficits9 (33)4 (25)5 (45)
 Reflex abnormalities9 (33)4 (25)5 (45)
 Incontinence5 (19)3 (19)2 (18)
 C-spine4 (15)3 (19)1 (8)
 T-spine14 (52)6 (38)8 (74)
 L-spine9 (33)7 (44)2 (18)
 Multilevel epidural disease5 (19)05 (45)
Failed prior CTs   
 017 (62)12 (75)5 (46)
 1-25 (19)3 (19)2 (18)
 ≥35 (19)1 (6)4 (36)
Table 3. Degree, Etiology, and Clinical Classification of ESCC
CharacteristicAll Patients (%)CTRT+CT
  1. ESCC indicates epidural spinal cord compression; CT, chemotherapy; RT, radiotherapy; MRI, magnetic resonance imaging; WSMCG, White spinal metastatic cancer grade; CSMC, Constans spinal metastases classification.

Patients, No.271611
Degree of ESCC on MRI   
  01 (4)1 (6)0
  11 (4)1 (6)0
  211 (40)7 (44)4 (36)
  310 (36)5 (31)5 (45)
 No images4 (15)2 (13)2 (18)
ESCC etiology   
 Vertebral metastases11 (41)5 (31)6 (55)
 Paravertebral12 (44)9 (56)3 (27)
 No images4 (15)2 (13)2 (18)
ESCC on MRI after treatment   
  Improved15 (56)11 (69)4 (36)
  Same3 (11)2 (12)1 (9)
  Not assessable9 (33)3 (19)6 (55)
 I21 (78)13 (81)8 (73)
 II4 (15)2 (13)2 (18)
 III2 (7)1 (6)1 (9)
 I9 (33)6 (38)3 (27)
 II9 (33)4 (25)5 (45)
 III6 (22)4 (25)2 (18)
 IV3 (12)2 (13)1 (9)

Of the patients receiving active treatment, 17 (8 with NSGCT) had no prior CT, 5 (4 with NSGCT) had 1-2 prior CT regimens, and 5 (all NSGCT) had >3 prior regimens. Of the 17 patients with no prior CT, 13 had ESCC at presentation, 2 had prior RT to the hemipelvis, 1 had orchiectomy only, and 1 denied prior treatment. Patients without ESCC at presentation developed ESCC a median of 86.5 months after initial diagnosis in the CT group compared with a median of 36 months in the RT + CT group (P = .8).

Patients who received RT + CT had a higher proportion of failed CT regimens (36% vs 6%; P = .13), which is likely related to the higher percentage of NSGCT (73% vs 56%; P = .4); T-spine involvement (74% vs 38%; P = .07); multilevel epidural disease (45% vs 0%; P = .006); and bony vertebral metastases (55% vs 31%; P = .21) compared with patients who received CT alone.

The patients receiving CT alone were similar in their clinical presentation to those who received RT + CT with respect to back pain, ambulation, motor deficits, and incontinence, but they had fewer sensory deficits and reflex abnormalities (Table 2). The White spinal metastatic cancer grade had a comparable distribution in both groups, whereas the Constans spinal metastases classification had more patients with severe grades (III, IV) in the CT group (38% vs 27%; P = .69) (Table 3). The degree of ESCC was similar in the CT and RT + CT groups (grade 3: 31% vs 45%; P = .69). In the CT group, 56% of the ESCCs were caused by a paravertebral mass and neuroforaminal infiltration, whereas in the RT + CT group, 55% were caused by bony vertebral metastasis (Table 3).

Back pain and ambulation improved in all patients after treatment (Table 4). Incontinence improved in all 3 patients receiving CT alone, whereas neither of the RT + CT patients with incontinence recovered. ESCC on MRI improved in 69% of patients receiving CT, compared with 36% in the RT + CT group (P = .13), but pretreatment or follow-up imaging was not available for review in 55% of patients receiving RT + CT, compared with 19% in the CT group (Table 3).

Table 4. Clinical Outcome and Survival
CharacteristicAll PatientsCTRT+CT
  1. CT indicates chemotherapy; RT, radiotherapy; ESCC, epidural spinal cord compression; NR, not recorded.

Patients, No.271611
 Back pain resolved26/2616/1611/11
 Ambulation restored4/42/22/2
 Incontinence restored3/53/30/2
Survival, mo   
 Median after ESCC26NR15
 Median overall13813843

In the CT alone group, 14 patients (88%) received a platinum-based CT regimen for their ESCC with bleomycin, etoposide, and cisplatin (BEP) being the most commonly used regimen in 9 (56%) patients (Table 5). In the RT + CT group, 8 patients received platinum-based therapy, only 2 of whom received BEP and 1 patient each received paclitaxel + ifosfamide, temozolomide, and sunitinib (9% each) after RT (Table 5). Eight patients in the RT + CT group had received CT for their GCT before ESCC development, all of whom received platinum-based CT (5 of 8 received BEP [63%]). Fifteen patients received first-line CT for their ESCC, of whom 11 received CT alone and 4 received RT + CT; all received platinum-based CT (in CT group 9 of 11 [81%] and in RT + CT group 2 of 4 [50%] received BEP).

Table 5. Chemotherapeutic Regimen Used to Treat ESCC in Patients Receiving CT
 CT (%)RT+CT (%)
  1. ESCC indicates epidural spinal cord compression; CT, chemotherapy; RT, radiotherapy; BEP, bleomycin, etoposide and cisplatin; TIP, paclitaxel, ifosfamide, cisplatin; VIP, high dose cisplatin, etoposide, ifosfamide; CAV, cyclophosphamide, doxorubicin, vincristine; FOLFOX, folinic acid, 5-fluorouracil, oxaliplatin.

Patients, No.1611
BEP9 (56)2 (18)
TIP3 (19)1 (9)
VIP2 (13)1 (9)
Paclitaxel+cisplatin+ifosfamide1 (6)
CAV1 (6)
Etoposide+cisplatin1 (9)
Cisplatin+ifosfamide1 (9)
Paclitaxel+ifosfamide1 (9)
FOLFOX2 (18)
Temozolomide1 (9)
Sunitinib1 (9)

Median survival from GCT was 138 months for those treated with CT alone versus 43 months for those receiving RT + CT (P = .65). Median survival after diagnosis of ESCC was not reached in the CT group but was only 15 months in the RT group (P = .02) (Fig. 1). When analyzing the 13 patients who presented with ESCC, all 10 patients who received CT alone responded. Fifteen patients received first-line CT at the time of ESCC; median overall survival for those patients was not reached in the CT group (n = 11) compared with 22 months in the RT + CT group (n = 4) (P = .04) (Fig. 2). Twenty-two patients received ≤2 prior CT regimens, 15 in the CT and 7 in the RT + CT group. Analysis of these patients, excluding patients heavily pretreated with CT, revealed a median overall survival time of 138 months in the CT and 38 months in the RT + CT group (P = .23) (Fig. 3).

Figure 1.

Kaplan-Meier survival curve after diagnosis of epidural spinal cord compression (ESCC) is shown. The dashed line represents patients receiving chemotherapy (CT) (n = 16), and the solid line represents patients receiving radiation therapy (RT+CT) followed by chemotherapy (n = 11) (P = .02).

Figure 2.

Kaplan-Meier survival curve after diagnosis of epidural spinal cord compression (ESCC) and treatment with first-line therapy is shown. The dashed line represents patients receiving chemotherapy (CT) (n = 11), and the solid line represents patients receiving first-line radiation therapy followed by chemotherapy (RT+CT) (n = 4) (P = .04).

Figure 3.

Kaplan-Meier survival curve in patients with ≤2 prior chemotherapy regimens is shown. The dashed line represents patients receiving chemotherapy (CT) (n = 15), and the solid line represents patients receiving radiation therapy followed by chemotherapy (RT+CT) (n = 7) (P = .23).


Early detection and aggressive treatment of ESCC are essential for a good neurologic outcome. More than 20,000 new patients per year develop ESCC in the United States.19 Treatment objectives are preservation and improvement of neurologic function, pain relief, and maintenance of spinal stability. The treatment of ESCC is often palliative and includes high-dose glucocorticoids, surgical decompression, and RT. A recent randomized trial of ESCC from solid tumors demonstrated a superior outcome with surgical resection followed by RT versus RT alone; however, there were no patients with GCT included in that study.20

ESCC is rare in patients with GCT; in 1 study of 140 GCT patients, none had ESCC.21 Furthermore, there are no specific data addressing therapy of ESCC in this population. In chemosensitive malignancies, such as GCT, lymphoma, neuroblastoma, or Ewing sarcoma, CT could be considered the primary treatment modality for ESCC, providing the expected response would be sufficiently robust as to relieve compression on the spinal cord rapidly.22 CT also treats all sites of active disease, sparing patients a delay in systemic treatment and the myelotoxicity of spinal RT. Our patient cohort represents the largest group with ESCC caused by GCT described so far in the literature, and supports the use of CT as primary treatment for ESCC in these patients.

The first report that cisplatin-based CT could treat ESCC from GCT described 3 patients from Memorial Sloan-Kettering Cancer Center.15 Our current series did not include these patients, as they were treated before the MRI era. We sought to expand this experience during the period of enhanced diagnostics with spine MRI. In our 27 patients, we could demonstrate the effectiveness of CT in the management of GCT with epidural involvement. Patients receiving CT alone recovered their neurologic function, even when there was a high degree of ESCC associated with significant neurologic deficits. These data indicate that CT worked sufficiently quickly that spinal cord compression was relieved promptly and no neurologic deterioration occurred during treatment. In our patient population, RT + CT was used in patients with multilevel epidural disease who were also more heavily pretreated. These patients had more chemoresistant disease and developed ESCC later in the course of their malignancy, so it is unsurprising that patients receiving CT alone had a survival advantage after the diagnosis of ESCC. However, patients who received CT alone as first therapy as well as those who had received ≤2 chemotherapeutic regimens before the development of ESCC had longer survival than comparable patients treated with RT + CT; this reached statistical significance only in the first-line therapy patients, likely because of the small patient numbers. These patients predictably responded to CT alone for their ESCC, and definitive systemic therapy was not delayed for spinal RT, which could have compromised overall outcome in patients receiving combined therapy.

There are several limitations to our study, including the small patient number, its retrospective nature, and inherent patient selection bias. However, the critical message is clear: patients receiving first-line CT or first or second platinum-based salvage regimens have excellent relief of neurologic symptoms and superior survival when treated with CT alone. This recommendation is based on our limited retrospective data, but a prospective study is unlikely given the rarity of ESCC in GCT. On the basis of these results, CT alone should be the treatment of choice in newly diagnosed GCT patients who present with ESCC and GCT patients who have received no more than 2 prior regimens if the disease was previously chemosensitive. Clinical judgment is needed in patients who have received more than 2 prior CT regimens, particularly if they have progressed after high-dose CT. In these cases, ESCC is most likely to represent chemoresistant disease, and RT + CT may be the more appropriate therapy.


The authors made no disclosures.