Association of tumor location, extent of resection, and neurofibromatosis status with clinical outcomes for 221 spinal nerve sheath tumors

Free access

OBJECT

Intradural extramedullary spine tumors represent two-thirds of all primary spine neoplasms. Approximately half of these are peripheral nerve sheath tumors, mainly neurofibromas and schwannomas. Given the rarity of this disease and, thus, the limited analyses of clinical outcomes, the authors examined the association of tumor location, extent of resection, and neurofibromatosis (NF) status with clinical outcomes.

METHODS

Patients were identified through a search of the University of California, San Francisco, neuropathology database and a separate review of current procedural terminology codes. Data recorded included patient age, patient sex, clinical presentation, presence of NF, tumor type, tumor location, extent of resection (gross-total resection [GTR] or subtotal resection [STR]), and clinical follow-up.

RESULTS

Of 221 tumors in 199 patients (mean age 45 years), 53 were neurofibromas, 163 were schwannomas, and 5 were malignant peripheral nerve sheath tumors. The most common presenting symptom was spinal pain (76%), followed by weakness (36%) and sensory abnormalities (34%). Mean symptom duration was 16 months. In terms of spinal location, neurofibromas were more common in the cervical spine (74% vs 27%, p < 0.001), and schwannomas were more common in the thoracic and lumbosacral spine (73% vs 26%, p < 0.001). Rates of GTR were lower for neurofibromas than schwannomas (51% vs 83%, p < 0.001), regardless of location. Rates of GTR were lower for cervical (54%) than thoracic (90%) and lumbosacral (86%) lesions (p < 0.001). NF was associated with lower rates of GTR among all tumors (43% vs 86%, p < 0.001). The mean follow-up time was 32 months. Recurrence/progression was more common for neurofibromas than schwannomas (17% vs 7%, p = 0.03), although the mean time to recurrence/progression did not differ according to tumor type (45 vs 53 months, p = 0.63). As expected, GTR was associated with lower recurrence rates (4% vs 22%, p < 0.001). According to multivariate analysis, cervical location (OR 0.239, 95% CI 0.110–0.520) and presence of NF (OR 0.166, 95% CI 0.054–0.507) were associated with lower rates of GTR. In a separate model, only GTR (OR 0.141, 95% CI 0.046–0.429) was associated with tumor recurrence.

CONCLUSIONS

Resection is an effective treatment for spinal nerve sheath tumors. Neurofibromas were found more commonly in the cervical spine than in other regions of the spine and were associated with higher rates of recurrence and lower rates of GTR than other tumor types, particularly in patients with NF Types 1 or 2. According to multivariate analysis, both cervical location and presence of NF were associated with lower rates of GTR. According to a second multivariate model, the only variable associated with tumor recurrence was extent of resection. Maximal safe resection remains ideal for these lesions; however, patients with cervical tumors or NF should be counseled about their increased risk for recurrence.

ABBREVIATIONSGTR = gross-total resection; NF = neurofibromatosis; NF1 = NF Type 1; NF2 = NF Type 2; STR = subtotal resection.

OBJECT

Intradural extramedullary spine tumors represent two-thirds of all primary spine neoplasms. Approximately half of these are peripheral nerve sheath tumors, mainly neurofibromas and schwannomas. Given the rarity of this disease and, thus, the limited analyses of clinical outcomes, the authors examined the association of tumor location, extent of resection, and neurofibromatosis (NF) status with clinical outcomes.

METHODS

Patients were identified through a search of the University of California, San Francisco, neuropathology database and a separate review of current procedural terminology codes. Data recorded included patient age, patient sex, clinical presentation, presence of NF, tumor type, tumor location, extent of resection (gross-total resection [GTR] or subtotal resection [STR]), and clinical follow-up.

RESULTS

Of 221 tumors in 199 patients (mean age 45 years), 53 were neurofibromas, 163 were schwannomas, and 5 were malignant peripheral nerve sheath tumors. The most common presenting symptom was spinal pain (76%), followed by weakness (36%) and sensory abnormalities (34%). Mean symptom duration was 16 months. In terms of spinal location, neurofibromas were more common in the cervical spine (74% vs 27%, p < 0.001), and schwannomas were more common in the thoracic and lumbosacral spine (73% vs 26%, p < 0.001). Rates of GTR were lower for neurofibromas than schwannomas (51% vs 83%, p < 0.001), regardless of location. Rates of GTR were lower for cervical (54%) than thoracic (90%) and lumbosacral (86%) lesions (p < 0.001). NF was associated with lower rates of GTR among all tumors (43% vs 86%, p < 0.001). The mean follow-up time was 32 months. Recurrence/progression was more common for neurofibromas than schwannomas (17% vs 7%, p = 0.03), although the mean time to recurrence/progression did not differ according to tumor type (45 vs 53 months, p = 0.63). As expected, GTR was associated with lower recurrence rates (4% vs 22%, p < 0.001). According to multivariate analysis, cervical location (OR 0.239, 95% CI 0.110–0.520) and presence of NF (OR 0.166, 95% CI 0.054–0.507) were associated with lower rates of GTR. In a separate model, only GTR (OR 0.141, 95% CI 0.046–0.429) was associated with tumor recurrence.

CONCLUSIONS

Resection is an effective treatment for spinal nerve sheath tumors. Neurofibromas were found more commonly in the cervical spine than in other regions of the spine and were associated with higher rates of recurrence and lower rates of GTR than other tumor types, particularly in patients with NF Types 1 or 2. According to multivariate analysis, both cervical location and presence of NF were associated with lower rates of GTR. According to a second multivariate model, the only variable associated with tumor recurrence was extent of resection. Maximal safe resection remains ideal for these lesions; however, patients with cervical tumors or NF should be counseled about their increased risk for recurrence.

ABBREVIATIONSGTR = gross-total resection; NF = neurofibromatosis; NF1 = NF Type 1; NF2 = NF Type 2; STR = subtotal resection.

Primary spinal cord tumors represent approximately 5% of primary CNS tumors in adults; every year in A the United States alone, up to 1700 new cases are diagnosed.3,8,11 These tumors can be classified according to anatomical location as intramedullary, intradural extramedullary, dumbbell (intradural, both intraspinal and extraspinal), and extradural. Intramedullary tumors are primarily glial tumors (ependymomas and astrocytomas), and extradural lesions consist primarily of metastases, hematopoietic tumors, and primary bony tumors. Intradural extramedullary spinal cord tumors consist primarily of peripheral nerve sheath tumors and meningiomas16 and less commonly of metastases, lipomas, spinal nerve sheath myxomas, paragangliomas, sarcomas, and vascular tumors.2

Spinal neurofibromas are found in up to 38% of patients with neurofibromatosis Type 1 (NF1).10,18,24,25 However, they are estimated to cause symptoms or become clinically significant in only approximately 5% of patients, although this estimate is probably low.13,25,27 An alternate form of NF, known as familial spinal NF, is categorized by multiple neurofibromas symmetrically affecting the entire axial spine.6,19,26 Histologically, neurofibromas are composed of Schwann cells with abnormal nuclei and scant cytoplasm and of fibroblasts in a matrix of collagen fibers and myxoid material.17 They grow along the course of nerve fibers and are generally surrounded by a thick epineurium. Schwannomas are composed of neoplastic Schwann cells forming 2 basic patterns: compact, elongated cells with occasional nuclear palisading (Antoni A pattern) or less cellular, loosely textured cells with indistinct processes (Antoni B).17 Approximately 90% of schwannomas are solitary and sporadic; 4% arise in patients with NF Type 2 (NF2), and another 5% are multiple but unassociated with NF2.5

Spinal neurofibromas and schwannomas are rare, and published studies are limited to single-institution case series.4,7,9,12,15,20–22 These contributions are valuable but limited by sample size and age of data sets. We sought to characterize our institutional experience with the management of these lesions, focusing on patient demographics, history of NF, clinical presentation, and factors associated with extent of resection and tumor recurrence.

Methods

Data Collection

Patients were identified through a search of our institutional neuropathology database and a separate review of current procedural terminology codes by two of the authors (N.M.B. and C.P.A). All research activities were approved by the Committee on Human Research, University of California, San Francisco, institutional review board. Pathology records were reviewed, and the following peripheral nerve sheath tumors were included in the study: neurofibromas, schwannomas, and malignant peripheral nerve sheath tumors. Medical records were reviewed to establish patient age, sex, presence of NF1 or NF2 according to clinical criteria, presenting symptoms and duration, use of radiotherapy, tumor location, extent of resection (defined by postoperative MR images or operative report), time to last follow-up visit, and date of tumor recurrence. Tumor location was classified by the level of the involved nerve root (cervical, thoracic, or lumbosacral) and relationship with adjacent dura as follows: intradural extramedullary, extradural intraspinal (tumor within the spinal canal), or extradural paraspinal (tumor at the neural foramen or extending outward into adjacent tissues). For the purposes of this article, they are described as intradural, extradural, and paraspinal. Extent of resection was defined as gross-total resection (GTR) if there was no evidence of residual disease on postoperative MR images and as subtotal resection (STR) if residual tumor was present. Use of radiotherapy was limited to a small number of patients in this study. All patients who received radiation received stereotactic radiosurgery; however, for some, it was received at an outside institution and treatment regimens were unavailable for review. Patients with malignant peripheral nerve sheath tumors almost exclusively received treatment with fractionated radiotherapy at outside facilities. One patient received intensity-modulated radiation therapy in addition to proton-beam therapy.

Statistical Analysis

Univariate analysis of continuous variables was performed by using the Student t-test, and categorical variables were compared by using the chi-square test or Fisher exact test. Multivariate analysis was performed by using binary logistic regression. Variables were included in the multivariate model if univariate analysis demonstrated a statistically significant relationship, defined as p < 0.05. To assess recurrence-free survival times (the likelihood of a recurrence event over time), we generated Kaplan-Meier plots and used the log-rank test. All analyses were performed by using SPSS version 22 (IBM).

Results

Patient Demographics

A total of 221 tumors in 199 patients were identified. Patient demographics are summarized in Table 1. The mean patient age was 45 years (range 1–88 years). Tumor type (pathology) was classified as neurofibroma in 53 cases, schwannoma in 163 cases, and malignant peripheral nerve sheath tumor in 5 cases. The mean age was 36 years among patients with neurofibroma and 48 years among those with schwannoma (p < 0.001). Of the 221 tumors, 98 (44%) were in female and 123 (56%) in male patients. Tumors were slightly more common among males: 28 (53%) of 53 neurofibromas and 92 (56%) of 163 schwannomas were in males; however, the difference was not significant (p = 0.646). Tumors were associated with NF in 53 patients (24%); 42 tumors (19%) occurred in patients with NF1 and 11 (5%) in patients with NF2. NF was strongly associated with neurofibromas; NF was present in 40 of 53 neurofibroma and 12 of 163 schwannoma (p < 0.001) patients. Among neurofibromas, 39 of 40 occurred in patients with NF1 and the other in a patient with NF2; among the 12 schwannomas with NF, 10 were in patients with NF2 and 2 were in patients with NF1.

TABLE 1.

Characteristics, treatments, and outcomes for 221 tumors in 199 patients

CharacteristicNo. (%)*
Patient age (yrs)
 Mean45
 Median46
 Range1–87
Patient sex
 M123 (56)
 F98 (44)
Neurofibromatosis
 Type 142 (19)
 Type 211 (5)
 None168 (76)
Symptom duration (mos)
 Mean16
 Median6
 Range0–120
Clinical presentation
 Spinal pain168 (76)
 Weakness79 (36)
 Numbness/paresthesias76 (34)
 Gait disturbance12 (5)
 Bowel/bladder incontinence12 (5)
 Incidental finding4 (2)
Tumor type
 Neurofibroma53 (24)
 Schwannoma163 (74)
 Malignant peripheral nerve sheath5 (2)
Tumor spinal level
 Cervical85 (38)
 Thoracic62 (28)
 Lumbosacral74 (34)
Location
 Intradural160 (73)
 Extradural47 (21)
 Paraspinal14 (6)
Extent of resection
 GTR166 (75)
 STR54 (24)
 Biopsy1 (0.5)
Postop radiotherapy14 (6)
Recurrence20 (9)
Time to recurrence (mos)
 Mean49
 Median39
 Range4.5–119
Follow-up duration (mos)
 Mean32
 Median16
 Range0–162

Unless otherwise indicated, data reflect the number (%) of tumors, not the number (%) of patients.

Clinical Presentation

Among all patients, the most common presenting symptom was spinal pain (76%), followed by weakness (36%), sensory symptoms (34%), gait disturbances (5%), and bowel or bladder incontinence (5%). Only 2% of tumors were asymptomatic and found incidentally. Demographics of patients with neurofibroma and schwannoma are compared in Table 2. Pain at presentation was described by 58% of patients with neurofibroma and by 81% of those with schwannoma (p = 0.003). Weakness was present for 43% with neurofibroma and 34% with schwannoma (p = 0.141). Sensory symptoms were present in 26% of patients with neurofibromas and 38% with schwannomas (p = 0.168). Gait disturbances were less common and occurred only in 4% of patients with neurofibromas and 6% with schwannomas (p = 0.549); bowel and bladder incontinence were present in 9% of patients with neurofibromas and 4% with schwannomas (p = 0.146). No neurofibromas and 2% of schwannomas were found incidentally (p = 0.250). The mean symptom duration was 16 months and did not differ significantly between patients with neurofibromas (19 months) and schwannomas (15 months) (p = 0.560).

TABLE 2.

Comparison of tumor types

CharacteristicNeurofibroma (n = 53), No. (%)*Schwannoma (n = 163), No. (%)*p Value
Patient age (yrs)<0.001
 Mean3648
 Median3549
 Range1–7614–88
Patient sex0.646
 M28 (53)92 (56)
 F25 (47)71 (44)
Neurofibromatosis
 Typei39 (74)2 (1)<0.001
 Type 21 (1)10 (6)0.222
 None13 (25)151 (93)<0.001
Symptom duration (mos)0.560
 Mean1915
 Median66
 Range0.5–1200–120
Clinical presentation
 Spinal pain31 (58)132 (81)0.003
 Weakness23 (43)55 (34)0.141
 Numbness/paresthesias14 (26)62 (38)0.168
 Gait disturbance2 (4)10 (6)0.549
 Bowel/bladder incontinence5 (9)7 (4)0.146
 Incidental finding0 (0)4 (2)0.250
Location
 Cervical39 (74)44 (27)<0.001
 Thoracic7 (13)55 (34)0.004
 Lumbosacral7 (13)64 (39)<0.001
Location
 Intradural35 (66)123 (75)0.179
 Extradural13 (25)31 (19)0.387
 Paraspinal5 (9)9 (6)0.315
Extent of resection
 GTR27 (51)135 (83)<0.001
 STR26 (49)27 (17)<0.001
 Biopsy0 (0)1 (0.1)0.568
Postop radiotherapy4 (8)6 (4)0.245
Time to recurrence (mos)0.640
 Mean4553
 Median3753
 Range10–1185–119
Follow-up duration (mos)0.300
 Mean3831
 Median1716
 Range0–1630–157

Unless otherwise indicated, data reflect the number (%) of tumors, not the number (%) of patients.

Tumor Locations

Among all tumors, 38% were localized to the cervical spine, 28% to the thoracic spine, and 34% to the lumbosacral spine. Overall classification was intradural (73%), intraspinal extradural (21%), and paraspinal (6%). Specifically, neurofibromas were 66% intradural, 25% extradural, and 9% paraspinal. Schwannomas were 76% intradural, 19% extradural, and 5% paraspinal. The proportion of intradural, extradural, and paraspinal lesions did not differ between tumor types. Among neurofibromas, an intradural location was associated with an increased rate of STR.

Tumor types were compared by location. Overall, 74% of neurofibromas and 27% of schwannomas were localized to the cervical spine (p < 0.0001). However, in this study, schwannomas were more prevalent, and of the 85 cervical tumors included, 39 (46%) were neurofibromas, 44 (52%) were schwannomas, and 2 (2%) were malignant peripheral nerve sheath tumors. The proportion of thoracic tumors was 13% for neurofibromas and 34% for schwannomas (p = 0.004); the proportion of lumbosacral tumors was 13% for neurofibromas and 39% for schwannomas (p < 0.001). Of 62 thoracic tumors included in this study, 55 were schwannomas; of 74 lumbosacral tumors, 64 were schwannomas.

Extent of Resection

All patients received treatment for their tumors at the time of diagnosis. Among 216 neurofibromas and schwannomas, the extent of resection was GTR in 162 lesions, STR in 53, and biopsy in 1. Of the patients for whom tumor resection was subtotal, 10 received postoperative radiotherapy; 8 had previously undergone STR, 1 had undergone biopsy only, and 1 had 2 adjacent lesions (1 of which underwent GTR, and the other was asymptomatic and treated by radiotherapy alone). Of the 5 malignant peripheral nerve sheath tumors included in this study, 4 received GTR, 1 STR, and all postoperative radiotherapy.

Rates of GTR differed according to tumor pathology and location. The rate of GTR was 51% among neurofibromas and 83% among schwannomas (p < 0.001). Rates of GTR based on location were 54% for cervical, 90% for thoracic, and 86% for lumbosacral tumors (p < 0.001). NF status was also associated with extent of resection. Rates of GTR were 44% among patients with NF1 or NF2 and 85% among those without NF (p < 0.001). A multivariate model was generated to identify factors associated with GTR and used variables found to be significant on univariate analysis. A complete comparison of factors associated with GTR and STR for neurofibromas and schwannomas is presented in Tables 3 and 4, respectively. Only cervical location (OR 0.231, 95% CI 0.109–0.488) and presence of NF mutation (OR 0.199, 95% CI 0.0.073–0.541) were associated with an effect on GTR; tumor pathology and intradural location were not (Table 5).

TABLE 3.

Neurofibroma analysis

CharacteristicGTR (n=27), No. (%)*STR (n = 26), No. (%)*P Value
Patient age (yrs)0.174
 Mean3933
 Median3733
 Range5–641–76
Patient sex0.884
 M14 (52)14 (54)
 F13 (48)12 (46)
Neurofibromatosis
 Type 115 (56)24 (92)0.002
 Type 20 (0)1 (4)0.491
 None12 (44)1 (4)0.001
Symptom duration (mos)0.349
 Mean2413
 Median116
 Range0.5–1201–108
Clinical presentation
 Spinal pain19 (70)12 (46)0.067
 Weakness10 (37)13 (50)0.331
 Numbness/paresthesias7 (26)7 (27)0.931
 Gait disturbance2 (7)0 (0)0.490
 Bowel/bladder incontinence1 (4)4 (15)0.183
 Incidental finding0 (0)0 (0)
Location
 Cervical16 (59)23 (88)0.016
 Thoracic6 (22)1 (4)0.100
 Lumbosacral5 (19)2 (8)0.420
Durai location
 Intradural14 (52)21 (81)0.026
 Extradural9 (33)4 (15)0.202
 Paraspinal4 (15)1 (4)0.351
Postop radiotherapy04 (15)0.051
Recurrence2 (7)7 (27)0.076
Time to recurrence (mos)0.233
 Mean1753
 Median1740
 Range10–2417–117
Follow-up duration (mos)0.022
 Mean2452
 Median955
 Range0–1600–162

Unless otherwise indicated, data reflect the number (%) of tumors, not the number (%) of patients.

TABLE 4.

Schwannoma analysis

CharacteristicGTR (n = 135), No. (%)*STR (n = 27), No. (%)*p Value
Patient age (yrs)0.960
 Mean4848
 Median4950
 Range16–7914–88
Patient sex0.321
 M79 (59)13 (48)
 F56 (41)14 (52)
Neurofibromatosis
 Typei2 (1)0 (0)1.000
 Type 26 (4)4 (15)0.064
 None127 (95)23 (85)0.117
Symptom duration (mos)0.240
 Mean1419
 Median612
 Range0.1–1201–72
Clinical presentation
 Spinal pain108 (80)23 (85)0.789
 Weakness42 (31)12 (44)0.180
 Numbness/paresthesias48 (36)13 (48)0.218
 Gait disturbance8 (6)2 (7)0.673
 Bowel/bladder incontinence7 (5)0 (0)0.602
 Incidental finding4 (3)0 (0)1.000
Location
 Cervical29 (21)15 (56)<0.001
 Thoracic50 (37)5 (18)0.064
 Lumbosacral56 (42)7 (26)0.130
Durai location
 Intradural105 (78)18 (67)0.218
 Extradural24 (18)7 (26)0.326
 Paraspinal6 (4)2 (7)0.621
Postop radiotherapy1 (1)4 (15)0.003
Recurrence5 (4)5 (19)0.004
Time to recurrence (mos)0.262
 Mean6434
 Median5836
 Range5–1194–58
Follow-up duration (mos)0.136
 Mean2842
 Median1524
 Range0–1560–139

Unless otherwise indicated, data reflect the number (%) of tumors, not the number (%) of patients.

TABLE 5.

Multivariate analysis of factors associated with GTR

VariableOdds Ratio95% CIp Value
Cervical location0.2310.109–0.488<0.001
NF mutation0.1990.073–0.5410.002
Tumor pathology0.8800.310–2.5000.811
Intradural location1.2590.572–2.7680.567

Tumor Cntrol

The mean time to last follow-up visit was 32 months; 60% of patients were followed up for at least 1 year. The mean follow-up time was slightly longer for patients with neurofibroma (38 months) than for patients with schwannoma (31 months), but the difference was not significant (p = 0.300). Among neurofibromas and schwannomas, 20 recurred (total recurrence rate 9%). Among patients with at least 1 year of follow-up, the rate was 15% (20 of 130), and for those with at least 2 years of follow-up the rate was 20% (18 of 91). For those with at least 3 years of follow-up, the rate was 23% (17 of 73). Recurrence was more common for neurofibromas (17%) than for schwannomas (7%) (p = 0.026). The recurrence-free period was significantly shorter for patients with neurofibromas than for those with schwannomas (Fig. 1) (p = 0.018). Recurrences were more common among patients with NF (17%) than without NF (7%) (p =0.022). Time to recurrence was also shorter among patients with NF mutations (Fig. 2). The rate of recurrence was highest for cervical tumors (16%), compared with 2% for thoracic tumors and 8% for lumbosacral tumors (p = 0.015). In addition to differences in recurrence rates, the times to recurrence also differed among tumors at different locations and were highest for thoracic tumors (Fig. 3). Recurrence rates for intradural, extradural, and paraspinal lesions did not differ.

FIG. 1.
FIG. 1.

Tumor pathology was associated with recurrence-free survival (time to tumor recurrence). Time to recurrence was longer for schwannomas than neurofibromas (log-rank, p = 0.018).

FIG. 2.
FIG. 2.

Neurofibromatosis mutation status was associated with recurrence-free survival. Time to recurrence was significantly shorter among patients with NF1 or NF2 than among those without known NF mutations (log-rank, p = 0.012).

FIG. 3.
FIG. 3.

Tumor location was associated with recurrence-free survival. Time to recurrence was shorter among patients with tumors in the cervical spine than among those with tumors in the lumbosacral or thoracic spine. Recurrence-free survival time was longest among patients with tumors in the thoracic spine (log-rank, p = 0.038).

Extent of resection was significantly associated with tumor recurrence. The rate of recurrence was 4% for patients who underwent GTR and 24% for those who did not (p < 0.001). When tumor recurrence over time was assessed by use of Kaplan-Meier plots, this difference remained significant (Fig. 4). The number of patients receiving radiotherapy was limited; however, rates of recurrence were 27% (12 of 45) among patients who underwent STR and 0 (0 of 8) among those who underwent STR plus radiotherapy (p = 0.097). According to a multivariate model, only extent of resection was associated with tumor recurrence (OR = 0.160, 95% CI 0.051–0.503) (Table 6).

FIG. 4.
FIG. 4.

Extent of resection was associated with recurrence-free survival. Time to recurrence was significantly longer among tumors that underwent GTR than among those that underwent STR (log-rank, p = 0.011).

TABLE 6.

Multivariate analysis of factors associated with tumor recurrence

VariableOdds Ratio95% CIp Value
GTR0.1600.051–0.5030.002
Tumor pathology0.7160.149–4.3870.677
Cervical location1.6460.517–5.2410.399
NF mutation0.8900.181–4.3870.886
Radiotherapy0.3260.037–2.9100.316

Malignant Peripheral Nerve Sheath Tumors

Small sample size precluded a detailed analysis of this cohort. Among 5 patients with malignant peripheral nerve sheath tumors, the mean age at presentation was 42 years; 2 patients were female and 3 male. Only 1 patient had a history of NF1, and none had a history of NF2. Tumor location was cervical spine for 2 and lumbosacral region for 3. Follow-up in this cohort was limited to 0, 4, 6, 10, and 52 months. All but 1 patient underwent GTR of the tumor. Only 1 patient received radiotherapy before surgery (20 months before STR); unfortunately, disease recurred and the patient died 10 months after surgery. The remaining 3 patients all received postoperative radiotherapy, all at outside facilities. For 2 of these patients, therapy consisted of fractionated radiotherapy; these patients had no evidence of recurrent disease at 4 and 6 months, respectively. The third patient received intensity-modulated radiotherapy with proton-beam therapy; he remained alive without evidence of recurrent tumor at 52 months.

Discussion

This study characterized the clinical presentation of spinal nerve sheath tumors and identified factors associated with extent of resection and tumor recurrence. Among the 199 patients with 221 tumors, the mean age at diagnosis was 45 years, and the mean age differed significantly between patients with neurofibromas and those with schwannomas (36 vs 48 years, respectively). Given the high number of NF1 patients with neurofibromas included in this study (40 of 53) the younger age at presentation is not unexpected on the basis of the underlying tumor biology and is consistent with other reports in the literature.9,21 Among all NF patients included in this study, the majority (39 of 40) of neurofibromas were in patients with NF1 and the majority (10 of 12) of schwannomas were in patients with NF2. The associations between NF1 and neurofibromas and NF2 and schwannomas, as well as the underlying molecular mechanisms driving tumorigenesis, have been well described.1,14 There was no significant association between these tumors and patient sex, although neurofibromas (53%) and schwannomas (56%) were both slightly more commonly found in men.

Neurofibromas demonstrated a predilection for the cervical spine, which is consistent with other reports.7,21 However, among all cervical tumors in our cohort, approximately half (47%) were neurofibromas and the rest were schwannomas. Unlike neurofibromas, schwannomas are more evenly distributed throughout the spine and their prevalence is higher. Because these tumors share similar radiographic features, the presence of a cervical lesion does not make the diagnosis of neurofibroma more likely, and it is difficult to make any inference about nerve sheath tumor pathology based on clinical presentation or radio-graphic features unless there is a known history of NF1.

Although neurofibromas are often more challenging lesions for obtaining GTR and their biology is distinct from that of sporadic schwannomas, our multivariate analysis found that only cervical location and the presence of NF, and not tumor type, were associated with lower rates of GTR. The association with cervical location is not unexpected, given the anatomical constraints of the spinal cord and nerve roots compared with those of lesions in the thoracic (where the nerve roots can be sacrificed) or lumbosacral spine (where the nerve roots can be moved). The association with NF mutations was surprising but not completely unexpected. We believe that this association was identified because of the effect of the NF mutations on the invasive potential of the tumor as well as the presence of nerve fascicles within tumor tissue, making these lesions more challenging for complete resection because the planes are less distinct. Indeed, there is a biological rationale based on pathology and laboratory data that supports this notion.23 Among schwannomas included in this study, the rate of GTR was 66% (8 of 12) among patients with NF and 84% (127 of 151) among those without NF, although the difference was not statistically significant (p = 0.261), probably because of small numbers.

The rate of tumor recurrence among all nerve sheath tumors was 9% and significantly higher among neurofibromas (17%) compared with schwannomas (7%). The mean follow-up times between these groups was slightly higher among those with neurofibroma (38 months) than with schwannoma (31 months); however, this difference was not significant (p = 0.300). The recurrence rate was higher among patients with at least 1 year or at least 3 years of follow-up (20% and 23%, respectively). For previous surgical series, recurrence rates of 1.9%–5.8% and GTR rates of 85%-97% have been reported.4,7,9,12,15,20–22 However, many of these studies are characterized by higher numbers of schwannomas than neurofibromas. The likely explanation for the higher recurrence rates in our study is the higher number of spinal neurofibromas (53) and higher percentage of cervical nerve sheath tumors (38%) than are typically reported in the literature. Our multivariate analysis found that cervical location and the presence of NF mutations were associated with lower rates of GTR, factors that certainly contributed to the results.

The only variable found to correlate with tumor recurrence was extent of resection; not correlated were presence of NF, tumor pathology, tumor location, and use of postoperative radiotherapy. Although the presence of NF and cervical location can contribute to extent of resection, they are not independently associated with an increased risk for tumor recurrence, making surgical resection the most effective factor associated with disease control in these patients.

Limitations of this study include its retrospective nature and limited follow-up times, which hindered our ability to assess tumor control in this cohort. The rarity of these lesions makes prospective studies challenging. Limited follow-up time is an inevitable consequence of our clinical practice, in which complex patients from a large catchment area are referred to our center for surgical treatment and then choose to follow up with local providers. Another limitation is the lack of pathologic markers, for example MIB-1, to stratify tumors by their mitotic index because these markers would probably help predict which patients are more likely to experience tumor recurrence. Postoperative MR images were obtained for 67% of patients; however, future prospective studies should rely solely on postoperative imaging to determine extent of resection.

Conclusions

Resection is an effective treatment for spinal nerve sheath tumors. Neurofibromas were found more commonly in the cervical spine and are associated with higher rates of recurrence and lower rates of GTR. Multivariate analysis found that cervical location and NF were associated with lower rates of GTR. A second multivariate model to assess factors associated with tumor recurrence found that only extent of resection was significant. Maximal safe resection remains the ideal therapy for these lesions; however, patients with cervical tumors or NF should be counseled on their increased risk for recurrence.

Author Contributions

Conception and design: Ames, Safaee. Acquisition of data: Ames, Safaee, Tihan. Analysis and interpretation of data: Ames, Safaee. Drafting the article: Ames, Safaee. Critically revising the article: all authors. Reviewed submitted version of manuscript: all authors. Statistical analysis: Ames, Safaee. Administrative/technical/material support: Ames. Study supervision: Ames.

References

  • 1

    Abramowicz AGos M: Neurofibromin in neurofibromatosis type 1 - mutations in NF1gene as a cause of disease. Dev Period Med 18:2973062014

    • Search Google Scholar
    • Export Citation
  • 2

    Abul-Kasim KThurnher MMMcKeever PSundgren PC: Intradural spinal tumors: current classification and MRI features. Neuroradiology 50:3013142008

    • Search Google Scholar
    • Export Citation
  • 3

    Aghayev KVrionis FChamberlain MC: Adult intradural primary spinal cord tumors. J Natl Compr Canc Netw 9:4344472011

  • 4

    Albanese VPlatania N: Spinal intradural extramedullary tumors. Personal experience. J Neurosurg Sci 46:18242002

  • 5

    Antinheimo JSankila RCarpén OPukkala ESainio MJääskeläinen J: Population-based analysis of sporadic and type 2 neurofibromatosis-associated meningiomas and schwannomas. Neurology 54:71762000

    • Search Google Scholar
    • Export Citation
  • 6

    Carey JCViskochil DH: Neurofibromatosis type 1: A model condition for the study of the molecular basis of variable expressivity in human disorders. Am J Med Genet 89:7131999

    • Search Google Scholar
    • Export Citation
  • 7

    Celli P: Treatment of relevant nerve roots involved in nerve sheath tumors: removal or preservation?. Neurosurgery 51:6846922002

  • 8

    Chamberlain MCTredway TL: Adult primary intradural spinal cord tumors: a review. Curr Neurol Neurosci Rep 11:3203282011

  • 9

    Conti PPansini GMouchaty HCapuano CConti R: Spinal neurinomas: retrospective analysis and long-term outcome of 179 consecutively operated cases and review of the literature. Surg Neurol 61:34442004

    • Search Google Scholar
    • Export Citation
  • 10

    Egelhoff JCBates DJRoss JSRothner ADCohen BH: Spinal MR findings in neurofibromatosis types 1 and 2. AJNR Am J Neuroradiol 13:107110771992

    • Search Google Scholar
    • Export Citation
  • 11

    Engelhard HHVillano JLPorter KRStewart AKBarua MBarker FG: Clinical presentation, histology, and treatment in 430 patients with primary tumors of the spinal cord, spinal meninges, or cauda equina. J Neurosurg Spine 13:67772010

    • Search Google Scholar
    • Export Citation
  • 12

    Fernandes RLLynch JCWelling LGonçalves MTragante RTemponi V: Complete removal of the spinal nerve sheath tumors. Surgical technics and results from a series of 30 patients. Arq Neuropsiquiatr 72:3123172014

    • Search Google Scholar
    • Export Citation
  • 13

    Huson SMHarper PSCompston DA: Von Recklinghausen neurofibromatosis. A clinical and population study in southeast Wales. Brain 111:135513811988

    • Search Google Scholar
    • Export Citation
  • 14

    Irving RMMoffat DAHardy DGBarton DEXuereb JHMaher ER: Somatic NF2 gene mutations in familial and nonfamilial vestibular schwannoma. Hum Mol Genet 3:3473501994

    • Search Google Scholar
    • Export Citation
  • 15

    Jeon JHHwang HSJeong JHPark SHMoon JGKim CH: Spinal schwannoma; analysis of 40 cases. J Korean Neurosurg Soc 43:1351382008

    • Search Google Scholar
    • Export Citation
  • 16

    Klekamo JSamii M: Surgery of Spinal Tumors BerlinSpringer2007

  • 17

    Louis DNOhgaki HWiestler ODCavenee WKBurger PCJouvet A: The 2007 WHO classification of tumours of the central nervous system. Acta Neuropathol 114:971092007

    • Search Google Scholar
    • Export Citation
  • 18

    Poyhonen MLeisti ELKytölä SLeisti J: Hereditary spinal neurofibromatosis: a rare form of NF1?. J Med Genet 34:1841871997

  • 19

    Riccardi V: Neurofibromatosis: Phenotype Natural History and Pathogenesis 2BaltimoreJohns Hopkins University Press1992

  • 20

    Safavi-Abbasi SSenoglu MTheodore NWorkman RKGharabaghi AFeiz-Erfan I: Microsurgical management of spinal schwannomas: evaluation of 128 cases. J Neurosurg Spine 9:40472008

    • Search Google Scholar
    • Export Citation
  • 21

    Seppälä MTHaltia MJSankila RJJääskeläinen JEHeiskanen O: Long-term outcome after removal of spinal neurofibroma. J Neurosurg 82:5725771995

    • Search Google Scholar
    • Export Citation
  • 22

    Seppälä MTHaltia MJSankila RJJääskeläinen JEHeiskanen O: Long-term outcome after removal of spinal schwannoma: a clinicopathological study of 187 cases. J Neurosurg 83:6216261995

    • Search Google Scholar
    • Export Citation
  • 23

    Su WXing RGuha AGutmann DHSherman LS: Mice with GFAP-targeted loss of neurofibromin demonstrate increased axonal MET expression with aging. Glia 55:7237332007

    • Search Google Scholar
    • Export Citation
  • 24

    Thakkar SDFeigen UMautner VF: Spinal tumours in neurofibromatosis type 1: an MRI study of frequency, multiplicity and variety. Neuroradiology 41:6256291999

    • Search Google Scholar
    • Export Citation
  • 25

    Tonsgard JHKwak SMShort MPDachman AH: CT imaging in adults with neurofibromatosis-1: frequent asymptomatic plexiform lesions. Neurology 50:175517601998

    • Search Google Scholar
    • Export Citation
  • 26

    Visckochil DCarey JCAlternate and related forms of the neurofibromatoses. Huson SMHughes RAC: The Neuroflbromatoses: A Pathogenetic and Clinical Overview LondonChapman and Hall1994

    • Search Google Scholar
    • Export Citation
  • 27

    von Deimling AKrone WMenon AG: Neurofibromatosis type 1: pathology, clinical features and molecular genetics. Brain Pathol 5:1531621995

    • Search Google Scholar
    • Export Citation

If the inline PDF is not rendering correctly, you can download the PDF file here.

Article Information

Contributor Notes

Correspondence Christopher P. Ames, Department of Neurological Surgery and Orthopedic Surgery, 505 Parnassus Ave., Rm. M779, San Francisco, CA 94143. email: amesc@neurosurg.ucsf.edu.INCLUDE WHEN CITING DOI: 10.3171/2015.5.FOCUS15183.DISCLOSURE Dr. Ames has received grants from DePuy Synthes Spine and AOSpine North America; is a consultant for DePuy, Medtronic, and Stryker; is a patent holder for Fish & Richardson; and has direct stock ownership in Doctor’s Research Group and Visulase. Dr. Chou is a consultant for Globus, Medtronic, and Orthofix. Dr. Mummaneni receives royalties from DePuy Spine, Taylor and Francis/QMP Publishing, Springer Publishing Company, and Thieme Publishers; receives honoraria from Globus, DuPuy, and AOSpine; has overseen support of non-study-related clinical or research effort for AANS/CNS Joint Section on Disorders of the Spine and Peripheral Nerves; and has direct stock ownership in Spinicity/ISD.
Headings
Figures
  • View in gallery

    Tumor pathology was associated with recurrence-free survival (time to tumor recurrence). Time to recurrence was longer for schwannomas than neurofibromas (log-rank, p = 0.018).

  • View in gallery

    Neurofibromatosis mutation status was associated with recurrence-free survival. Time to recurrence was significantly shorter among patients with NF1 or NF2 than among those without known NF mutations (log-rank, p = 0.012).

  • View in gallery

    Tumor location was associated with recurrence-free survival. Time to recurrence was shorter among patients with tumors in the cervical spine than among those with tumors in the lumbosacral or thoracic spine. Recurrence-free survival time was longest among patients with tumors in the thoracic spine (log-rank, p = 0.038).

  • View in gallery

    Extent of resection was associated with recurrence-free survival. Time to recurrence was significantly longer among tumors that underwent GTR than among those that underwent STR (log-rank, p = 0.011).

References
  • 1

    Abramowicz AGos M: Neurofibromin in neurofibromatosis type 1 - mutations in NF1gene as a cause of disease. Dev Period Med 18:2973062014

    • Search Google Scholar
    • Export Citation
  • 2

    Abul-Kasim KThurnher MMMcKeever PSundgren PC: Intradural spinal tumors: current classification and MRI features. Neuroradiology 50:3013142008

    • Search Google Scholar
    • Export Citation
  • 3

    Aghayev KVrionis FChamberlain MC: Adult intradural primary spinal cord tumors. J Natl Compr Canc Netw 9:4344472011

  • 4

    Albanese VPlatania N: Spinal intradural extramedullary tumors. Personal experience. J Neurosurg Sci 46:18242002

  • 5

    Antinheimo JSankila RCarpén OPukkala ESainio MJääskeläinen J: Population-based analysis of sporadic and type 2 neurofibromatosis-associated meningiomas and schwannomas. Neurology 54:71762000

    • Search Google Scholar
    • Export Citation
  • 6

    Carey JCViskochil DH: Neurofibromatosis type 1: A model condition for the study of the molecular basis of variable expressivity in human disorders. Am J Med Genet 89:7131999

    • Search Google Scholar
    • Export Citation
  • 7

    Celli P: Treatment of relevant nerve roots involved in nerve sheath tumors: removal or preservation?. Neurosurgery 51:6846922002

  • 8

    Chamberlain MCTredway TL: Adult primary intradural spinal cord tumors: a review. Curr Neurol Neurosci Rep 11:3203282011

  • 9

    Conti PPansini GMouchaty HCapuano CConti R: Spinal neurinomas: retrospective analysis and long-term outcome of 179 consecutively operated cases and review of the literature. Surg Neurol 61:34442004

    • Search Google Scholar
    • Export Citation
  • 10

    Egelhoff JCBates DJRoss JSRothner ADCohen BH: Spinal MR findings in neurofibromatosis types 1 and 2. AJNR Am J Neuroradiol 13:107110771992

    • Search Google Scholar
    • Export Citation
  • 11

    Engelhard HHVillano JLPorter KRStewart AKBarua MBarker FG: Clinical presentation, histology, and treatment in 430 patients with primary tumors of the spinal cord, spinal meninges, or cauda equina. J Neurosurg Spine 13:67772010

    • Search Google Scholar
    • Export Citation
  • 12

    Fernandes RLLynch JCWelling LGonçalves MTragante RTemponi V: Complete removal of the spinal nerve sheath tumors. Surgical technics and results from a series of 30 patients. Arq Neuropsiquiatr 72:3123172014

    • Search Google Scholar
    • Export Citation
  • 13

    Huson SMHarper PSCompston DA: Von Recklinghausen neurofibromatosis. A clinical and population study in southeast Wales. Brain 111:135513811988

    • Search Google Scholar
    • Export Citation
  • 14

    Irving RMMoffat DAHardy DGBarton DEXuereb JHMaher ER: Somatic NF2 gene mutations in familial and nonfamilial vestibular schwannoma. Hum Mol Genet 3:3473501994

    • Search Google Scholar
    • Export Citation
  • 15

    Jeon JHHwang HSJeong JHPark SHMoon JGKim CH: Spinal schwannoma; analysis of 40 cases. J Korean Neurosurg Soc 43:1351382008

    • Search Google Scholar
    • Export Citation
  • 16

    Klekamo JSamii M: Surgery of Spinal Tumors BerlinSpringer2007

  • 17

    Louis DNOhgaki HWiestler ODCavenee WKBurger PCJouvet A: The 2007 WHO classification of tumours of the central nervous system. Acta Neuropathol 114:971092007

    • Search Google Scholar
    • Export Citation
  • 18

    Poyhonen MLeisti ELKytölä SLeisti J: Hereditary spinal neurofibromatosis: a rare form of NF1?. J Med Genet 34:1841871997

  • 19

    Riccardi V: Neurofibromatosis: Phenotype Natural History and Pathogenesis 2BaltimoreJohns Hopkins University Press1992

  • 20

    Safavi-Abbasi SSenoglu MTheodore NWorkman RKGharabaghi AFeiz-Erfan I: Microsurgical management of spinal schwannomas: evaluation of 128 cases. J Neurosurg Spine 9:40472008

    • Search Google Scholar
    • Export Citation
  • 21

    Seppälä MTHaltia MJSankila RJJääskeläinen JEHeiskanen O: Long-term outcome after removal of spinal neurofibroma. J Neurosurg 82:5725771995

    • Search Google Scholar
    • Export Citation
  • 22

    Seppälä MTHaltia MJSankila RJJääskeläinen JEHeiskanen O: Long-term outcome after removal of spinal schwannoma: a clinicopathological study of 187 cases. J Neurosurg 83:6216261995

    • Search Google Scholar
    • Export Citation
  • 23

    Su WXing RGuha AGutmann DHSherman LS: Mice with GFAP-targeted loss of neurofibromin demonstrate increased axonal MET expression with aging. Glia 55:7237332007

    • Search Google Scholar
    • Export Citation
  • 24

    Thakkar SDFeigen UMautner VF: Spinal tumours in neurofibromatosis type 1: an MRI study of frequency, multiplicity and variety. Neuroradiology 41:6256291999

    • Search Google Scholar
    • Export Citation
  • 25

    Tonsgard JHKwak SMShort MPDachman AH: CT imaging in adults with neurofibromatosis-1: frequent asymptomatic plexiform lesions. Neurology 50:175517601998

    • Search Google Scholar
    • Export Citation
  • 26

    Visckochil DCarey JCAlternate and related forms of the neurofibromatoses. Huson SMHughes RAC: The Neuroflbromatoses: A Pathogenetic and Clinical Overview LondonChapman and Hall1994

    • Search Google Scholar
    • Export Citation
  • 27

    von Deimling AKrone WMenon AG: Neurofibromatosis type 1: pathology, clinical features and molecular genetics. Brain Pathol 5:1531621995

    • Search Google Scholar
    • Export Citation
TrendMD
Cited By
Metrics

Metrics

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 649 555 29
PDF Downloads 335 235 21
EPUB Downloads 0 0 0
PubMed
Google Scholar