Home | About JCVJS | Editorial board | Ahead of print | Current Issue | Archives | Instructions | Subscribe | Advertise | Contact us |   Login 
Journal of Craniovertebral Junction and Spine
Search Articles   
Advanced search   

   Table of Contents  
Year : 2021  |  Volume : 12  |  Issue : 4  |  Page : 336-360  

Primary extradural tumors of the spinal column: A comprehensive treatment guide for the spine surgeon based on the 5th Edition of the World Health Organization bone and soft-tissue tumor classification

1 Department of Orthopedic Surgery, Icahn School of Medicine – The Mount Sinai Hospital, New York, USA
2 Department of Neurosurgery, Cantonal Hospital Lucerne, Lucerne, Switzerland
3 Department of Radiation Oncology, North Shore University Hospital, Manhasset, USA
4 Department of Pathology, North Shore University Hospital, Manhasset, USA
5 Department of Orthopedics, Balgrist University Hospital, Zurich, Switzerland
6 Department of Radiology, University Hospital Zurich, Zurich, Switzerland
7 Department of Neurosurgery, North Shore University Hospital, Manhasset, USA

Date of Submission16-Aug-2021
Date of Acceptance06-Nov-2021
Date of Web Publication11-Dec-2021

Correspondence Address:
Alexander Spiessberger
Department of Neurosurgery, North Shore University Hospital, 300 Community Drive, Manhasset, NY 11030
Login to access the Email id

Source of Support: None, Conflict of Interest: None

DOI: 10.4103/jcvjs.jcvjs_115_21

Rights and Permissions

Background: In 2020, the World Health Organization (WHO) published the 5th version of the soft tissue and bone tumor classification. Based on this novel classification system, we reviewed the current knowledge on all tumor entities with spinal manifestations, their biologic behavior, and most importantly the appropriate treatment options as well as surgical approaches.
Methods: All tumor entities were extracted from the WHO Soft-Tissue and Bone Tumor Classification (5th Edition). PubMed and Google Scholar were searched for the published cases of spinal tumor manifestations for each entity, and the following characteristics were extracted: Growth pattern, ability to metastasize, peak age, incidence, treatment, type of surgical resection indicated, recurrence rate, risk factors, 5-year survival rate, key molecular or genetic alterations, and possible associated tumor syndromes. Surgical treatment strategies as well as nonsurgical treatment recommendations are presented based on the biologic behavior of each lesion.
Results: Out of 163 primary tumor entities of bone and soft tissue, 92 lesions have been reported along the spinal axis. Of these 92 entities, 54 have the potential to metastasize. The peak age ranges from conatal lesions to 72 years. For each tumor entity, we present recommended surgical treatment strategies based on the ability to locally destruct tissue, to grow, recur after resection, undergo malignant transformation as well as survival rates. In addition, potential systemic treatment recommendations for each tumor entity are outlined.
Conclusion: Based on the 5th Edition of the WHO bone and soft tumor classification, we identified 92 out of 163 tumor entities, which potentially can have spinal manifestations. Exact preoperative tissue diagnosis and interdisciplinary case discussions are crucial. Surgical resection is indicated in a significant subset of patients and has to be tailored to the specific biologic behavior of the targeted tumor entity based on the considerations outlined in detail in this article.

Keywords: Chordoma, primary spinal tumors, sarcoma

How to cite this article:
Arvind V, Nevzati E, Ghaly M, Nasim M, Farshad M, Guggenberger R, Sciubba D, Spiessberger A. Primary extradural tumors of the spinal column: A comprehensive treatment guide for the spine surgeon based on the 5th Edition of the World Health Organization bone and soft-tissue tumor classification. J Craniovert Jun Spine 2021;12:336-60

How to cite this URL:
Arvind V, Nevzati E, Ghaly M, Nasim M, Farshad M, Guggenberger R, Sciubba D, Spiessberger A. Primary extradural tumors of the spinal column: A comprehensive treatment guide for the spine surgeon based on the 5th Edition of the World Health Organization bone and soft-tissue tumor classification. J Craniovert Jun Spine [serial online] 2021 [cited 2022 Dec 1];12:336-60. Available from: https://www.jcvjs.com/text.asp?2021/12/4/336/332210

   Introduction Top

The core principles guiding surgical treatment for primary bone and soft-tissue tumors have been introduced by Enneking et al. more than 40 years ago and comprise three different types of surgical tumor resection: Intralesional, marginal en bloc, and wide en bloc resection.[1] It has been suggested that tumor location (intracompartmental versus extracompartmental) and histologic grade should be used to determine the mode of resection. Since the introduction of Enneking's system additional research regarding primary bone and soft-tissue tumors, new nonsurgical treatment modalities such as stereotactic radiosurgery or targeted molecular therapies and novel radiographic techniques together have significantly improved demarcating tumor extent and curbing tumor invasion.

This article is based on the 5th Edition of the World Health Organization (WHO) tumor classification of bone and soft-tissue tumors, published in 2020. We compiled the most recent knowledge of all tumor entities, which have been described to occur along the spinal axis and surrounding soft tissues.[2] This comprehensive overview summarizes clinical knowledge as well as imaging findings of all primary, extradural spinal tumors described in the literature.

We describe our treatment algorithms, which is individualized for each tumor entity and loosely based on Enneking's classification system, and modified by contemporary imaging protocols.

   Methods Top

The 5th Edition of the WHO soft tissue and bone tumors classification, published in 2020 was reviewed and individual tumor entities extracted into a spreadsheet. Medical databases (PubMed and Google Scholar) were searched for publications reporting occurrences of each entity listed in the WHO classification along the spinal axis (spinal bones or paraspinal soft tissues). If an entity has been reported to occur along the spinal axis, a case report with exemplary imaging findings was obtained. For each tumor entity, the following data were extracted from the WHO classification or other key references: Relevant differential diagnoses, growth pattern (infiltrative/destructive), potential for malignant transformation, potential to metastasize, peak age, incidence, recommended type of surgical resection (A, B, C), recurrence rate, treatment, risk factors, 5-year overall survival rate, key molecular or genetic alterations, and possible associated tumor syndromes. All primary bone and soft tissue tumor entities listed in the 5th Edition of the WHO tumor classification were listed in a spreadsheet and a note was made on entities reported to occur along the spinal axis. In a second spreadsheet, exemplary imaging findings of each entity have been listed or say: “Exemplary imaging findings of each entity are listed in a second spreadsheet.” Moreover, finally, in a third spreadsheet, the above-mentioned key characteristics for each entity have been listed.

   Results Top

A comprehensive list of all primary bone and soft-tissue tumors, as listed in the most recent WHO classification is given in [Appendix 1] and comprises a total of 163 entities. Of note, the following tumors can arise in either bone or soft tissue: Hemangioma, epitheloid hemangioma, epitheloid hemangioendothelioma, angiosarcoma, desomplastic fibroma, fibrosarcoma, chondroma, and osteosarcoma.

Tumor entities are classified by the cell of tumor origin [Appendix 1]. For soft-tissue neoplasms, the following cells of origin are as follows: Adipocytic, fibroblastic and myofibroblastic, fibrohistiocytic, vascular, pericytic (perivascular), smooth muscle, skeletal muscle, gastrointestinal stromal, chondro-osseous, and peripheral nerve sheath. Two further categories exist for all soft-tissue tumors that do not fall into the above mentioned: Tumors of uncertain differentiation and undifferentiated small round cell sarcomas. In the case of bone tumors, the following subclassification based on the cell population of origin exists: Chondrogenic, osteogenic, fibrogenic, vascular, osteoclastic giant cell-rich, or notochordal. Two further subcategories are listed in the WHO classification: Other mesenchymal bone tumors and hematopoietic neoplasms of the bone.

The results of our literature search are outlined in [Appendix 2] and [Appendixe 3] and show that 92 out of 163 entities were reported to occur either in spinal bones or paraspinal soft tissue. We categorized 92 entities with imaging [Appendix 2] and clinical/molecular findings [Appendix 3], as well as recommended surgical and nonsurgical treatment options.

Appendix 3 shows a comprehensive characterization of each tumor by: Growth pattern (infiltrative/locally destructive or not), ability to metastasize, ability to undergo malignant transformation, mean age at diagnosis, incidence, suggested mode of resection (intralesional resection A, marginal en bloc resection B, wide, or compartmental en bloc excision C), recurrence rate, treatment strategy, tumor risk factors, 5-year overall survival (OS) rate, genetic/molecular tumor characteristics, possible associated tumor syndromes, and corresponding cross -sectional imaging findings are presented in Appendix 2.

As shown in Appendix 3, the incidence rates for primary extradural spinal bone or soft-tissue tumors range from 2% (hemangioma) to a low of only two published cases for spinal nodular fasciitis. The survival rates of malignant lesions range from 94% for 5 year OS for ossifying fibromyxoid tumor to 7% for dedifferentiated osteosarcoma. A total of 54 entities are capable of forming metastases, 1 additional entity can form so called benign pulmonary metastases (chondroblastoma). The peak age ranges from conatal lesions (lymphangioma) to 72 years (pleomorphic rhabdomyosarcoma).

   Discussion Top

The most recent edition of the WHO classification of bone and soft-tissue tumors lists a total of 163 tumor entities, out of which 92 have been previously reported in the literature to potentially occur in the spine. Surgical resection is the integral part of treatment for most of these lesions and follows the overriding principles outlined by Enneking et al. in 1980,[1] as shown in [Figure 1]. Type B and C resections are more complex than type A resections with higher rates of complications; however, type B/C resections are associated with superior oncologic outcome as compared to type A resections for malignant lesions.[3] It must be noted that given to the unique anatomy of the spine, when compared to long bones, in many cases, a type B resection might be indicated. While type B resections may not be technically feasible, spine surgeons may opt for type C resections with a wider excision. [Figure 2] provides an overview of important growth characteristics of malignant bone and soft-tissue tumors. As indicated, the growth pattern of sarcomas is infiltrative. Even with a rim of reactive tissue, the pseudocapsule may act only as a weak barrier to prevent tumor spread. While the pseudocapsule has been shown to restrict tumor permeation after radio- or chemotherapy it is not a true barrier for tumor spread.[4] Cortical bone as well as major fascial planes, such as pleura or peritoneum are considered bone fide barriers. It is known from radiologic studies that infiltrating tumor nests, known as skip lesions, outside the primary tumor can be depicted on magnetic resonance imaging (MRI) in up to 16.5% of patients.[5] As shown in [Figure 2], once the cortical bone of the vertebra is breached, the tumor cells can freely spread until they reach the next level of solid barrier [routes A-D in [Figure 2]]. As has been shown in previous correlating studies between preoperative imaging and intraoperative histologic analysis, the mean discrepancy between tumor margin on preoperative MRI and intraoperative histology for osteosarcomas is 5 mm.[6],[7] Since short-tau inversion recovery and postcontrast T1 imaging overestimates the tumor extend by 1.68 cm, tumor outline is best depicted on noncontrast-enhanced T1 images.[8] Therefore, in our own experience if a malignant tumor is confined to one compartment, we perform either a type B resection with a margin of 5 mm on top of the tumor outline in the preoperative noncontrast T1 images, or we perform a type C resection, which will remove the whole tumor bearing compartment. If a malignant tumor extends into more than one compartment (e.g., cortical bone erosion in the case of vertebral osteosarcomas), we prefect to discuss either neo-adjuvant treatment to “downsize” the tumor (the more compartments the tumor extends into, the less likely a true wide en bloc resection can be achieved) or surgery to encompass an en bloc resection of the primary tumor bearing compartment plus the extension into a neighboring compartment with a safety margin of at least 5 mm.
Figure 1: Overview of the three different surgical types of resection in the treatment of spinal tumors

Click here to view
Figure 2: Illustration of potential routes of and barriers to spread of spinal sarcomas. Lesions, detached from the primary tumor are termed skip lesions. Barriers to skip lesions are: (A) Cortical bone, (B) pleura in cases when the lateral vertebral cortex has been breached, (C) muscle fascia in case of posterior cortical tumor breach, (D) dura in case of cortical breach of the spinal canal

Click here to view

How to incorporate these principles into surgical practice depends on the index level. In the case of C1 and C2, oncologic resections type B and C in most cases require a transmandibular approach [Figure 3]. When compared to the rest of the cervical spine negative margins are less likely to be obtained due to the anatomical complexity of the region.[9] For the rest of the mobile spine the WBB system has been proposed to choose the appropriate approach or combination of approach to perform a type B or C resection [Figure 4].[10] The choice of approach for oncologic resections of the sacrum is mainly determined by the anatomic level of the lesion as well as the presence of visceral tumor infiltration. [Figure 5] outlines our institutional algorithm to such lesions. Only lesions located below the inferior margin of the sacroiliac joint (SIJ) without visceral invasion are resected using a posterior-only approach. All other lesions are resected using an anterior/posterior approach. Reconstruction of the pelvic ring is necessary if more than 50% of the SIJs are resected. In instances where the tumor extends by more than 3 cm beyond the SIJ, we consider them as primarily inoperable (due to the large tumor volume and complexity of reconstruction).
Figure 3: Oncologic resection (type B and C) of primary tumors of C1 and C2 are carried out in most cases utilizing a transmandibular approach

Click here to view
Figure 4: Choice of approach for oncologic tumor resections of the subaxial spine (excluding sacrum), based on the Weinstein-Boriani-Biagini system. The vertebra is divided in 12 zones and based on the tumor location either an anterior approach (purple), posterior approach (green) are chosen. For each scenario the osteotomy sites are indicated and in cases necessitating combined approach the suggested order is indicated (I, II)

Click here to view
Figure 5: Sacral resections can be performed in a posterior only approach (green) or combined anterior/posterior approach (purple and green). It is our practice to liberally perform anterior approaches to separate the tumor from the visceral structures, to ligate the bilateral internal iliac arteries, to harvest abdominal wall flaps for reconstruction

Click here to view

Reconstruction of large resection cavities in many cases requires the involvement of plastic surgery and is beyond the scope of this article.

En bloc resections are technically demanding and have been shown to have higher complication rates when compared to type A resections, particularly when more than 1 level is being resected (Spiessberger A, PubMed ID pending), even though lesion etiology seems to have less impact on complication rates.

Given the profile of potential complications in the case of type B and C resections, rigorous preoperative planning is of paramount importance. Neurologic deficits are particularly devastating to patients and should be avoided at all costs. Other than direct mechanical injury, ischemic spinal cord injury has been reported to occur on rare occasions.[11],[12] Even though spinal cord blood supply is highly collateralized, postoperative infarcts can be a complication due to segmental vessel ligation.[11],[13] Spinal cord blood supply is established through the anterior spinal artery, a branching vessel of the vertebral arteries, as well as from as posterior spinal arteries through branching vessels of either vertebral or posterior inferior cerebellar arteries. Collateral flow is provided through variable radiculomedullary vessels, typically 2-3 cervical (bilaterally equal), 2-3 thoracic (left more than right), and 0-1 lumbar (left more than right).[12] Three major radiculomedullary vessels are described: The artery of cervical enlargement (usually a branching vessel from the ascending cervical artery at C6), the artery “von Haller” (usually the T5 segmental vessel) as well as the artery of Adamkiewicz (usually the T10 segmental vessel).[14] Watershed areas, susceptible to ischemic infarction in cases of hypotension or hypoxia have been suggested in the mid thoracic spine as well as the posterior aspect of the conus medullaris.[15] Type B and C resections require segmental artery ligation; however, recent studies have suggested that up to three adjacent segmental vessel can be sacrificed safely.[16],[17] We believe, that caution should be taken when ligating one of the three major radiculomedullary vessels, as described above. Preoperative high-resolution CT angiography can help localize the level of these three vessels. Intraoperative temporary nerve root/segmental vessel clamping with cautious observation of motor evoked potential/somatosensory evoked potential is important as well. In addition, intraoperative and postoperative hypotension should be avoided at all costs when a major radiculomedullary vessel has been sacrificed. It is also worth noting that the choice of vasopressor might make a difference as well. Animal studies comparing norepinephrine and phenylephrine in their properties to increase spinal cord perfusion in the setting of hypotension have shown, that norephinephrine provides better restoration of blood flow and oxygenation.[18] One should also recognize that radiculomedullary vessel ligation may not only render the patient more susceptible to ischemic cord injury, but also surgical trauma to segmental vessels or vertebral arteries can lead to embolic cord infarcts caused by vessel dissections.[19] In the case of cervical type B and C resections, preoperative endovascular sacrifice of one vertebral artery in case high degree tumor encasement (>180°) can be safely performed following careful study of a CT angiogram of both cervical vessels and posterior circulation. Side dominance, potential stenoses, size or absence of the posterior communicating arteries (in the case of fetal posterior cerebral artery variants) must be determined. Moreover, temporary endovascular balloon occlusion can be considered to determine the safety of vessel occlusion.

   Conclusion Top

Based on the 5th Edition of the WHO bone and soft tumor classification, we identified 92 out of 163 tumor entities, which potentially can have spinal manifestations. Exact preoperative tissue diagnosis and interdisciplinary case discussions are crucial. Surgical planning has to be tailored to the specific biologic behavior of the targeted tumor entity based on the considerations outlined in detail in this article.

Financial support and sponsorship


Conflicts of interest

There are no conflicts of interest.

   Appendixes Top

[Additional file 1]

[Additional file 2]

[Additional file 3]

[Additional file 4]

[Additional file 5]

   References Top

Enneking WF, Spanier SS, Goodman MA. A system for the surgical staging of musculoskeletal sarcoma. Clin Orthop Relat Res 1980;153:106-20.  Back to cited text no. 1
WHO Classification of Tumours Editorial Board. Soft Tissue and Bone Tumours: WHO Classification of Tumours (Medicine). 5th ed. Lyon: World Health Organization; 2020.  Back to cited text no. 2
Yamazaki T, McLoughlin GS, Patel S, Rhines LD, Fourney DR. Feasibility and safety of en bloc resection for primary spine tumors: A systematic review by the Spine Oncology Study Group. Spine (Phila Pa 1976) 2009;34:S31-8.  Back to cited text no. 3
Gitelis S, Thomas R, Templeton A, Schajowicz F. Characterization of the pseudocapsule of soft-tissue sarcomas. An experimental study in rats. Clin Orthop Relat Res 1989;246:285-92.  Back to cited text no. 4
Saifuddin A, Sharif B, Oliveira I, Kalus S, Barnett J, Pressney I. The incidence of skip metastases on whole bone MRI in high-grade bone sarcomas. Skeletal Radiol 2020;49:945-54.  Back to cited text no. 5
Jin T, Deng ZP, Liu WF, Xu HR, Li Y, Niu XH. Magnetic resonance imaging for the assessment of long bone tumors. Chin Med J (Engl) 2017;130:2547-50.  Back to cited text no. 6
O'Flanagan SJ, Stack JP, McGee HM, Dervan P, Hurson B. Imaging of intramedullary tumour spread in osteosarcoma. A comparison of techniques. J Bone Joint Surg Br 1991;73:998-1001.  Back to cited text no. 7
Putta T, Gibikote S, Madhuri V, Walter N. Accuracy of various MRI sequences in determining the tumour margin in musculoskeletal tumours. Pol J Radiol 2016;81:540-8.  Back to cited text no. 8
Molina CA, Ames CP, Chou D, Rhines LD, Hsieh PC, Zadnik PL, et al. Outcomes following attempted en bloc resection of cervical chordomas in the C-1 and C-2 region versus the subaxial region: A multiinstitutional experience. J Neurosurg Spine 2014;21:348-56.  Back to cited text no. 9
Boriani S, Weinstein JN, Biagini R. Primary bone tumors of the spine. Terminology and surgical staging. Spine (Phila Pa 1976) 1997;22:1036-44.  Back to cited text no. 10
Ng YH, Kato S, Demura S, Shinmura K, Yokogawa N, Nakade Y, et al. Delayed ischemic spinal cord injury after total en bloc spondylectomy in the thoracic spine. J Orthop Sci 2021 Jan 8;S0949-2658(20)30373-0.  Back to cited text no. 11
Melissano G, Civilini E, Bertoglio L, Calliari F, Campos Moraes Amato A, Chiesa R. Angio-CT imaging of the spinal cord vascularisation: A pictorial essay. Eur J Vasc Endovasc Surg 2010;39:436-40.  Back to cited text no. 12
Du C, Liu S, Jia F, Liu X, Wei F. Delayed incomplete paraplegia after en bloc spondylectomy of thoracic metastasis, a case report. 2020.  Back to cited text no. 13
Gailloud P. The artery of von Haller: A constant anterior radiculomedullary artery at the upper thoracic level. Neurosurgery 2013;73:1034-43.  Back to cited text no. 14
Gailloud P, Gregg L, Galan P, Becker D, Pardo C. Periconal arterial anastomotic circle and posterior lumbosacral watershed zone of the spinal cord. J Neurointerv Surg 2015;7:848-53.  Back to cited text no. 15
Kato S, Kawahara N, Tomita K, Murakami H, Demura S, Fujimaki Y. Effects on spinal cord blood flow and neurologic function secondary to interruption of bilateral segmental arteries which supply the artery of Adamkiewicz: An experimental study using a dog model. Spine (Phila Pa 1976) 2008;33:1533-41.  Back to cited text no. 16
Murakami H, Kawahara N, Tomita K, Demura S, Kato S, Yoshioka K. Does interruption of the artery of Adamkiewicz during total en bloc spondylectomy affect neurologic function? Spine (Phila Pa 1976) 2010;35:E1187-92.  Back to cited text no. 17
Streijger F, So K, Manouchehri N, Gheorghe A, Okon EB, Chan RM, et al. A direct comparison between norepinephrine and phenylephrine for augmenting spinal cord perfusion in a porcine model of spinal cord injury. J Neurotrauma 2018;35:1345-57.  Back to cited text no. 18
Montalvo M, Bayer A, Azher I, Knopf L, Yaghi S. Spinal cord infarction because of spontaneous vertebral artery dissection. Stroke 2018;49:e314-7.  Back to cited text no. 19


  [Figure 1], [Figure 2], [Figure 3], [Figure 4], [Figure 5]


    Similar in PUBMED
   Search Pubmed for
   Search in Google Scholar for
 Related articles
    Access Statistics
    Email Alert *
    Add to My List *
* Registration required (free)  

  In this article
    Article Figures

 Article Access Statistics
    PDF Downloaded289    
    Comments [Add]    

Recommend this journal