The Spine Journal 8 (2008) 1024–1029

Sacral intraspinal extradural primitive neuroectodermal tumor

Volker Musahl, MD, Jeffrey A. Rihn, MD, Frank E. Fumich, MD, James D. Kang, MD*Department of Orthopaedic Surgery, Division of Spinal Surgery, University of Pittsburgh Medical Center, Liliane Kaufmann Building,

3471 Fifth Avenue, Suite 1010, Pittsburgh, Pennsylvania 15213, USA

Received 21 October 2006; accepted 8 April 2007

Abstract BACKGROUND CONTEXT: Intraspinal pr

FDA device/drug

Nothing of value

manuscript.

* Corresponding a

Kaufmann Building,

15213. Tel.: (412) 60

E-mail address: k

1529-9430/08/$ – see

doi:10.1016/j.spinee.2

imitive neuroectodermal tumors (PNETs) are anexceedingly rare entity. A recent literature research revealed 28 cases reported. Only a few tumorsin the literature were extradural in location, in the cervical and thoracic spine. The average survivalafter combination treatment including chemotherapy, radiation, and surgical resection is 20 monthsfor the cases reported in the literature.PURPOSE: We report a case of a patient with sciatica and cauda equine–like symptoms.STUDY DESIGN: Case report.METHODS: Urgent sacral decompression and resection of the tumor was performed with rapidpain relief for the patient.RESULTS: Histology revealed a sacral extradural small blue-cell tumor, consistent with ES/PNETfamily tumors. An oncological workup revealed that the tumor presentation was metastatic withpulmonary and abdominal nodules. The patient underwent combination chemotherapy with vincris-tine, doxorubicin, and cyclophosphamide with mesna for 4 months.CONCLUSIONS: The patient was without disease after excision, two courses of 4-month chemo-therapy, and one course of 5-week radiation to the sacrum at 2 years. � 2008 Elsevier Inc. Allrights reserved.

Keywords: Primitive neuroectodermal tumors; Extradural; Sacral; Tumor

Introduction

Intraspinal primitive neuroectodermal tumors (PNETs)are an exceedingly rare entity. A recent literature researchrevealed less than 30 cases reported. Predilection of tumorlocation seemed to be the thoracic and lumbar spine [1]. Al-most all reported cases of PNETs were found in an intra-dural location. Only three tumors were extradural inlocation, in the cervical and thoracic spine [2–4]. Unfortu-nately, many PNETs were found to be metastatic at initialpresentation and therefore bear a devastating prognosis. Av-erage survival after combination treatment including che-motherapy, radiation, and surgical resection is 20 monthsfor the cases reported in the literature. We report a caseof a sacral intraspinal, extradural PNET in a young patientthat was metastatic at presentation.

status: not applicable.

received from a commercial entity related to this

uthor. Department of Orthopaedic Surgery, Liliane

3471 Fifth Avenue, Suite 1010, Pittsburgh, PA

5-3241; fax: (412) 687-3724.

angjd@upmc.edu (J.D. Kang)

front matter � 2008 Elsevier Inc. All rights reserved.

007.04.001

Case report

JW is a 27-year-old gentleman who was referred fromhis primary care physician. He complained of low backpain that suddenly occurred upon awakening 1 month ear-lier. He described the pain as sharp, severe (9/10), continu-ous, and only minimally relived by pain medications. Healso complained of pain radiating into the right buttock,the posterior aspect of the right thigh, and the posteriorleg. In addition, he experienced weakness in his right calfand a numb sensation over his right heel. Upon further ask-ing, the patient admitted to new onset of urinary inconti-nence and sexual dysfunction. He denied recent or pasttrauma to the low back or lower extremities. A review ofsystems revealed no history of headaches, dizziness, nau-sea, vomiting, fever, chills, night sweats, and no changein his weight. A review of the cardiovascular, respiratory,gastrointestinal, endocrine, integument, and psychiatricsystems was also negative. Oral steroids and physical ther-apy prescribed by his family physician did not provide suf-ficient relief of his symptoms. The patient used 60 mg ofoxycontin per day, which was prescribed by his primarycare physician for pain control. The patient’s past medical

1025V. Musahl et al. / The Spine Journal 8 (2008) 1024–1029

history was not significant for any medical problems. Hissocial history was not significant for alcohol or tobaccouse. The patient was unable to work as a telecommunica-tions worker for the past 2 weeks because of his leg painand weakness.

On physical examination, the patient was in no acutedistress and was alert and oriented. The patient had a nor-mal-based and well-balanced gait, which was coordinatedbut slow. He had positive heel walk, but no toe walkingwas possible on the right side. Strength testing of thelower-extremity muscle groups revealed 5/5 strength forbilateral iliopsoas, quadriceps, tibialis anterior, extensorhallucis longus, hamstrings, peroneals, and the left gastroc-nemius-soleus complex. The right gastrocnemius-soleuscomplex was graded slightly weaker, 4/5. The straight legraise was positive at 30� of flexion on the right and negativeon the left. The Achilles’ tendon reflex was absent on theright side and normal on the left. There was a well-described area of numbness to light touch on the right heel.A rectal examination was performed in the office andshowed a normal sphincter tone and intact perineal sensa-tion. The patient had a positive bulbocavernosus reflexand absent clonus, Hoffmann’s, or Babinski reflex.

Magnetic resonance imaging (MRI) obtained by his pri-mary care physician was available for review (Fig. 1). Onboth the sagittal and axial images, a mass within the spinalcanal at the S1–S2 level was detected. The mass had low-signal intensity on T1-weighted MRI and moderate signalintensity on T2-weighted MRI. There was a small area ofheterogenicity seen on both the T1- and T2-weightedMRI. The mass was seen in the S1–S2 area and emanatedfrom the S1 nerve root on the right side and crossing themidline halfway to the left. This mass appeared to be

Fig. 1. (Left) T2 sagittal MRI of the lumbosacral spine. Low-signal intensity intr

axial CT myelogram. Intraspinal extradural lesion at S1 (low-signal intensity) an

causing extradural compression of the thecal sac in the sa-crum pushing the remaining elements to the contralateralleft side. However, there was no surrounding edema orsoft-tissue destruction, and the mass appeared to be wellmarginated on MRI.

Hospital course

The patient was admitted to the hospital for furtherworkup, preoperative planning, and intravenous corticoste-roid therapy. He was started on 8 mg of decadron every 6hours. After 48 hours in the hospital, the patient explainedthat were was significant improvement of his pain. To fur-ther delineate the intradural extraspinal mass and evaluatepossible destruction of surrounding bony structures, a com-puted tomography (CT) myelogram was obtained. As pre-viously seen on MRI, the mass was right sided andextradural and did not show bony destruction. The obtainedlaboratory studies returned within normal limits and in-cluded an erythrocyte sedimentation rate, C-reactiveprotein, and complete blood count with differential.

The preliminarily impression was that the lesion in ques-tion was a benign peripheral nerve sheath tumor such asa schwannoma that compressed the sacral nerve roots andcaused the right leg pain, weakness, urinary, and sexualdysfunction. The differential diagnosis also included mas-sive disc herniation, neurofibroma, infection, and hemangi-oma. Rare tumors in the young adult population such aslymphoma, Ewing’s sarcoma, plasmocytoma, and even me-tastasis were included in the differential as well. The mostlikely diagnosis for this common presentation of sciatica ina young patient like JW with a documented intraspinalextradural mass was schwannoma. However, alerted by the

aspinal extradural lesion posterior to S1/2 disc. (Middle) Scout and (Right)

d spinal cord (high-signal intensity).

1026 V. Musahl et al. / The Spine Journal 8 (2008) 1024–1029

rather rapid onset and the cauda equine–like presentation, itwas disclosed to the patient that a malignancy could not beexcluded from the differential. Therefore, the patient was in-formed of and consented for a sacral laminectomy and an ex-cisional biopsy of the tumor. This would decompress thesacral nerve roots and reveal a definitive diagnosis.

Excisional biopsy (Fig. 2)

The patient was positioned in the prone position undergeneral anesthesia. The sacrum was approached througha posterior, midline, vertical incision. The S1 and S2 laminaof the sacrum was exposed subperiosteally with a Cobbelevator. An intraoperative X-ray film confirmed the levelof dissection. The spinous processes of both S1 and S2were resected, and access to the spinal canal was achievedthrough the interval of S1 and S2. A Kerrison rongeur wasintroduced in the S1–S2 interval, and a central laminec-tomy of S1 and S2 was performed [5,6]. On entering the ca-nal a soft, friable, gelatinous material in the epidural spacewas found to be pushing the cauda equina to the left side ofthe canal. After the laminectomy was completed, dissectionwas continued to identify the proximal and distal extent ofthe tumor. The tumor was found to be approximately2 cm�3 cm�5 cm in size, with small areas of hemorrhageand focal fibrosis. The tumor did not adhere to the underly-ing dura. Therefore, complete removal of the soft tumorwas possible without damaging the dura or sacrificing nerveroots. The origin of the tumor appeared to be in proximityof the exiting portion of the right S1 nerve root and themost cranial part of the S2 nerve root. Both the right S1and S2 neural foramen were explored, and no further tumorwas identified. The entire specimen was sent to pathology,and the wound was closed in standard fashion over a drain.

Postoperative course

Postoperatively, the patient experienced rapid relief ofhis right lower-extremity pain. His drain was discontinued

Fig. 2. An intraoperative image of the intraspinal extradural mass extend-

ing from S1 to S2.

on postoperative day 2. After a day of bed rest, the patientwas allowed to ambulate with weight bearing as tolerated.He was able to ambulate with a walker and showed slightimprovement in his right gastrocnemius strength at the timeof discharge from the hospital on postoperative day 4.

The final pathology report was consistent with an ES/PNET family tumor. Further workup was therefore man-dated. The diagnosis was disclosed to the patient and hisfamily. The oncology service was consulted as well as theradiation oncology service. His care was referred to the pe-diatric oncology specialist, and a metastatic workup wascompleted. A CT scan of the thorax, abdomen, and pelviswas obtained and showed several small pulmonary nodules,less than 0.5 cm�0.5 cm in size. There were also smallmesenteric lymph nodes seen that ranged from a few milli-meters up to a centimeter in diameter.

It was found that the patient had metastatic disease at thistime. Therefore, recommendations from the radiation oncol-ogy service included both multi-agent chemotherapy and ra-diation therapy. The patient was discharged from the hospitalon the third postoperative day with full ability to ambulateand ability to void without difficulty. Follow-up with radia-tion therapy and oncology for chemotherapy was scheduled.

His outpatient course with the pediatric oncology serviceincluded an initial chemotherapy regimen of vincristine,doxorubicin, and cyclophosphamide with mesna for 4months. He then underwent 5 weeks of radiation therapy tothe sacrum followed by an additional 4 months of chemother-apy. A repeat CT scan of the sacrum at one year follow-upshowed the patient to be free of disease. A repeat CT scan ofthe thorax, abdomen, and pelvis at one year follow-up showedthe pulmonary nodules to be stable and the mesenteric lymphnodes decreased in prominence. A Positron Emission To-mography (PET) scan was performed at 18 months followup and showed the patient to be disease free. The patient re-turned to work 15 months after the index procedure and iswithout symptoms or compromise at two year follow up.

Discussion

Soft-tissue PNETs in the extremities, paravertebral, andthoracoabdominal are well described in the literature; how-ever, PNETs of the spine are a rare entity. We identifiedthree reports in the literature about extradural, intraspinalPNETs (thoracic, lumbar, and lumbosacral) [2–4]. The av-erage age of these patients was 18 years, and the reportedsurvival ranged from 6 to 22 months. Two tumors wereisolated lesions, whereas one tumor was metastatic to thelungs.

Histopathologically, PNETs are undifferentiated small,blue, round-cell tumors. They have hyperchromatic nuclei,indistinct cytoplasm, and hair-like extensions that tend toform rosettes called Homer-Wright rosettes (Fig. 3). Theamount and quality of rosette formation can vary, but someauthors consider rosette formation pathogonomic for the

Fig. 3. (Top) Hematoxylin and eosin staining of the primitive neuroecto-

dermal tumor with sheets of small round cells with round nuclei and indis-

tinct cytoplasm. (Bottom) Hematoxylin and eosin staining of the primitive

neuroectodermal tumor (high-power magnification).

1027V. Musahl et al. / The Spine Journal 8 (2008) 1024–1029

diagnosis of PNET [7]. It is important to make the differen-tiation between central and peripheral PNET because thespine is reported to host both central and peripheral PNET.Peripheral PNET stain positive for multiple neural markers,including neuron-specific enolase, Leu-7, synaptophysin,and PGP 9.5 [7]. Therefore, the diagnosis is currently basedon both the presence of Homer-Wright rosettes and positivestaining for neural markers. Central PNET is a central ner-vous system tumor that occurs in the spine as so-called‘‘drop’’ metastasis.

Immunhistochemical staining is used to detect a specificprotein in a sample of tissue after the histopathological di-agnosis remains unresolved. Several different markers areused to establish the diagnosis of a PNET. Protein geneproduct 9.5 (PGP 9.5) and neuron-specific enolase are gly-colytic enzymes that are released into the CSF when neuraltissue is injured [8]. Leu-7 (also called CD57) is a myelin-associated glycoprotein present on Schwann cells. It isa nonspecific marker for neuroendocrine cells and neuralcells [9]. Synaptophysin is a membrane glycoprotein ofsynaptic vesicles that is expressed in neurons and endocrine

cells. It is therefore used as a marker for nerve terminalsand neuroendocrine tumors [10].

The occurrence of a PNET was first described by Stout in1918 [11]. Three years later, Ewing [12] described a diffuseround-cell endothelioma of the distal radius. However, it wasnot until 1975 that PNET and Ewing’s sarcoma (ES) wereconsidered related entities, when Angervall and Enzinger[13] described 39 cases of extraskeletal ES that occurredin the soft tissues of the lower extremities and the trunk. Acommon cytogenetic abnormality, t(11;22)(q24;q12), ofES and PNET, was later identified [14]. These tumors arenow referred to as ES/PNET family tumors.

Centrally occurring PNET was first described as an un-differentiated tumor in the cerebrum [15]. Later, it was ad-vocated that all central nervous system tumors arising fromprimitive neuroepithelial cells should be called PNET[16,17]. PNETs have since been classified by the WorldHealth Organization as embryonal tumors [18]. CentralPNETs most commonly arise in the cerebellum and werealso found in the pineal gland, brain stem, and peripheralnerves [19]. High-grade malignancy is secondary to metas-tasizing through CSF pathways into the spinal canal and thesubarachnoid space.

The clinical course of PNETs is reported to be rapid pro-gression and neurological compromise. As in this case, thepatient presented with rather acute neurological compro-mise and cauda equine–like symptoms. The evaluationand management of such symptoms should include protect-ing of the spinal cord (ie, dexamethasone treatment), ob-taining the appropriate imaging studies (ie, MRI or CTmyelogram) in a timely fashion, and urgent surgical decom-pression of the spinal cord with removal of the tumor [20].Most reports in the literature use MRI as their initial diag-nostic modality [1,21]. However, the final diagnosis ofa PNET can only be made histologically and usuallyrequires the addition of immunhistochemical staining.

Because MRI and CT myelogram are sensitive but notspecific in detecting PNETs, a broad differential diagnosismust be considered when working up an intraspinal, extra-dural, well-demarcated mass that does not involve the bone,ligaments, or disc. This differential must include (1) benigntumors, such as schwannoma, neurofibroma, lipoma, epidu-ral hemangioma, or external meningioma; (2) metastasis;(3) malignant tumors, such as lymphoma, plasmocytoma,myeloma, and rare sarcomas; and (4) infection. In the pa-tient presented in this report, there was no family historyof neuromuscular diseases, and a detailed history and phys-ical examination revealed no other lesions of the skin, softtissues, or bone. Preoperative serum parameters and hisclinical course were not suggestive of an infection. Thelocation of the tumor in the sacrum and the appearanceof the tumor on imaging were not consistent with certainlesions, including primary bone tumors, hemangioma, ormenigioma.

There are no guidelines in the literature as to how radicalthe excision of the tumor and how toxic the oncological

1028 V. Musahl et al. / The Spine Journal 8 (2008) 1024–1029

therapy should be. The presented case involved the sacrum,where radical resection would be associated with signifi-cantly increased risk for wound complications, loss ofbowel and bladder function, and increased length of hospi-tal stay [22]. In reviewing the literature, some reports de-scribe diffuse tumor infiltration and fixation to the spinalnerves, allowing for only partial tumor resection [1]. Theapproach taken for the reported patient included excisionalbiopsy and removal of the entire tumor without contaminat-ing surrounding tissues because the tumor did not adhere tothe dura, nerve roots, or lamina. This was followed by localradiation therapy and multiple cycles of chemotherapy.However, a cautionary statement must be made at thispoint. Whenever feasible, a tissue diagnosis and a screeningevaluation should be completed before attempting a mar-ginal or intralesional excision to avoid cutting through anunexpected or undiagnosed tumor.

Radiation therapy is a mainstay of treatment for PNET,and most reported cases in the literature were treated witha combination of surgery and radiation [4,23–26]. Theaccepted utilized radiation dose for local radiation in thespinal canal is 5,000 cGy [27]. Some authors recommendradiation therapy to the entire neuroaxis to account for mi-cro metastasis in the spinal canal [28]. This is especiallytrue for central intradural PNETs that tend to spread viathe CSF and subarachnoidally.

Literature on chemotherapeutic treatment regimens issparse, and current guidelines are according to the treatmentprotocols for medulloblastoma [29]. Multiple adjuvant orneoadjuvant chemotherapeutic agents are recommended forsystemic treatment. These include multiagent therapy withthe use of alkaloid (vincristine), DNA synthesis inhibitor cis-platin, and alkylating agents such as cyclophosphamide [30].However, progression of the tumor and systemic disease canoften not be stopped. Recommendations for treating theserare tumors are therefore combination of surgical excision,radiation therapy, and chemotherapy according to the me-dulloblastoma guidelines.

Future directions in the treatment of PNETs includeadjuvant therapy such as adoptive immunotherapy withintrathecal administration of interleukin-2 and lympho-kine-activated killer cells [31]. Peripheral blood stem celltransfusion is a possible treatment for tumors in whichhigh-dose chemotherapy is required [32].

Summary

PNETs originating around the spinal cord are very rarebut present an important differential diagnosis in young pa-tients presenting with a history of sciatica with neurologicalcompromise. Algorithms for the management of patientswith sciatica should include careful initial empiric treat-ment, appropriate imaging modalities, and time- andfashion-appropriate surgical interventions. The prognosisof these tumors is very poor. Tumor size, extent of the

disease, presence of metastasis, and surgical margin influ-ence overall survival. Multiagent chemotherapy combinedwith surgical excision and radiation therapy is the preferredtreatment for patients with PNET around the spinal cord.

References

[1] Kim YW, et al. Primary intraspinal primitive neuroectodermal tumor

at conus medullaris. Yonsei Med J 2004;45:533–8.

[2] Liu HM, et al. Intraspinal primitive neuroectodermal tumor arising

from the sacral spinal nerve root. J Comput Tomogr 1987;11:

350–4.

[3] Harimaya K, et al. Primitive neuroectodermal tumor and extraskeletal

Ewing sarcoma arising primarily around the spinal column: report of

four cases and a review of the literature. Spine 2003;28:E408–12.

[4] Dorfmuller G, et al. Intraspinal primitive neuroectodermal tumor: re-

port of two cases and review of the literature. Acta Neurochir (Wien)

1999;141:1169–75.

[5] White AH, et al. Lumbar laminectomy for herniated disc: a prospective

controlled comparison with internal fixation fusion. Spine 1987;12:

305–7.

[6] Postacchini F, et al. The surgical treatment of central lumbar stenosis.

Multiple laminotomy compared with total laminectomy. J Bone Joint

Surg Br 1993;75:386–92.

[7] Enzinger FM, Weiss S, Goldblum J. Soft tissue tumors 2001;

1265–321.

[8] Virji MA, MD W, Herberman RB. Tumor markers in cancer diagno-

sis and prognosis. Cancer J Clin 1988;38:104–26.

[9] Kraus MD, Haley J. Lymphocyte predominance Hodgkin’s disease:

the use of bcl-6 and CD57 in diagnosis and differential diagnosis.

Am J Surg Pathol 2000;24:1068–78.

[10] Reisinger C, et al. The synaptophysin/synaptobrevin complex disso-

ciates independently of neuroexocytosis. J Neurochem 2004;90:1–8.

[11] Stout A. Tumor of the ulnar nerve. Proc NY Pathol Soc 1918;18(2).

[12] Ewing J. Diffuse endothelioma of bone. Proc NY Pathol Soc

1921;21(17).

[13] Angervall L, Enzinger FM. Extraskeletal neoplasm resembling

Ewing’s sarcoma. Cancer 1975;36:240–51.

[14] Aurias A, et al. Translocation of chromosome 22 in Ewing’s sarcoma.

C R Seances Acad Sci III 1983;296:1105–7.

[15] Hart M, Earle K. Primitive neuroectodermal tumors of the brain in

children. Cancer 1973;32:890–7.

[16] Backer A, Mount S, Zarka M. Desmoplastic round cell primary

tumor of unknown origin with lymph nodes and lung metastasis:

Histological, cytological, ultrastructural, cytogenetic and molecular

findings. Virchows Arch 1998;432:135–41.

[17] Rorke L. The cerebellar medulloblastoma and relationship to primi-

tive neuroectodermal tumors. J Neuropath Exp Neorol 1983;42:1–15.

[18] Kielhuer P, Bueger PC, Scheithauer BW. The new WHO classifica-

tion of brain tumors. Brain Pathol 1993;3:225–68.

[19] Srikanth D, Lee C, Ghassan S. Primary intramedullary primitive neu-

roectodermal tumor of the spinal cord: case report and review of the

literature. Neurosurgery 1997;41:1417–20.

[20] Boden SD, Wiesel SW. Lumbar spine imaging: role in clinical deci-

sion making. J Am Acad Orthop Surg 1996;4:238–48.

[21] Virani MJ, Jain S. Primary intraspinal primitive neuroectodermal

tumor (PNET): a rare occurence. Neurol India 2002;50:75–80.

[22] Guo Y, et al. Bowel and bladder continence, wound healing, and

functional outcomes in patients who underwent sacrectomy. J Neuro-

surg Spine 2005;3:106–10.

[23] Albrecht CF, et al. Primary intraspinal primitive neuroectodermal tu-

mor: report of two cases and review of the literature. J Neurooncol

2003;61:113–20.

[24] Konsick EJ, et al. Primitive neuroectodermal tumors of the central

nervous system in children. J Neurosurg 1978;48:741–6.

1029V. Musahl et al. / The Spine Journal 8 (2008) 1024–1029

[25] Kepes JJ, et al. Primitive neuroectodermal tumors of the cauda equina

in adults with no detectable primary intracranial neoplasmdthree

case studies. Clin Neuropathol 1985;4:1–11.

[26] Yavuz AA, et al. Primary intraspinal primitive neuroectodermal

tumor: case report of a tumor arising from the sacral spinal nerve

root and review of the literature. Am J Clin Oncol 2002;25:

135–9.

[27] Emami B, et al. Tolerance of normal tissue to therapeutic irradiation.

Int J Radiat Oncol Biol Phys 1991;21:109–22.

[28] Reddy AT, et al. Outcome for children with spratentorial primitive

neuroectodermal tumors treated with surgery, radiation, and chemo-

therapy. Cancer 2000;88:2189–93.

[29] Tornesello A, et al. Progressive disease in children with medulloblasto-

ma/PNET during preradiation chemotherapy. J Neurooncol 1999;45:

135–40.

[30] Hisaoka M, Hashimoto H, Muraro T. Peripheral primitive neuroecto-

dermal tumor with ganglioneuroma-like areas arising from the sacral

nerve root. Virchows Arch 1997;431:365–9.

[31] George RE, et al. In vitro cytolysis of primitive neuroectodermal

tumors of the posterior fossa (medulloblastoma) by lymphokine-

activated killer cells. J Neurosurg 1988;69:403–9.

[32] Burchill SA, et al. Minimal residual disease at the time of peripheral

blood stem cell harvest in patients with advanced neuroblastoma.

Med Pediatr Oncol 2001;36:213–9.