Surgical Principles and Research Advances in Spinal Tumours: A Review

Surgical Principles and Research Advances in Spinal Tumours: A Review

Best for: Spine and neurosurgeons, orthopaedic oncologists and musculoskeletal tumour surgeons, radiation and interventional radiologists, pathologists and molecular diagnosticians, intraoperative neuromonitoring technologists, perioperative teams, and clinical research and evidence-methodology specialists. Primary keywords: spinal tumour; surgical margin; Enneking staging; WBB staging; total en bloc spondylectomy; separation surgery; three-column theory; intraoperative neuromonitoring; 3D-printed prosthesis; navigation and robotics

Short Answer

Spinal tumours are not a single disease but a group of lesions with highly heterogeneous biology and prognosis. Primary spinal tumours account for nearly 10 percent of all primary bone tumours, whereas metastatic disease accounts for more than 90 percent of spinal tumours encountered clinically, and autopsy series report spinal involvement in 30 to 70 percent of patients dying of malignancy, most often from lung, breast, prostate, kidney and colorectal primaries, with the thoracic spine most frequently affected. Surgical treatment rests on three principles: adequacy of the oncological margin, restoration of mechanical stability, and preservation of neurological function. These principles constrain rather than complement one another. Margin decisions rely on combined Enneking and WBB staging, the latter dividing the vertebral cross-section into 12 radial zones and five longitudinal layers that map directly onto approach and resection type. Experience shows local recurrence of 20 to 50 percent after intralesional curettage or piecemeal resection, whereas en bloc resection reduces recurrence substantially, which explains the survival value of total en bloc spondylectomy (TES) in primary malignant spinal tumours. The price is equally clear: reported complication rates for TES reach 52 to 86.7 percent, with postoperative cerebrospinal fluid leakage in about 23.6 percent of cases, so margin selection is a trade-off between radicality and safety rather than a matter of technical preference. Stability reconstruction follows Denis three-column theory: single-column defects permit anterior reconstruction alone, two-column defects require combined anterior and posterior reconstruction, and three-column defects demand 360-degree circumferential reconstruction. Reconstructive materials have moved from titanium mesh cages to 3D-printed porous metal prostheses, which improve endplate matching and encourage osseointegration. For neurological protection, combined somatosensory and motor evoked potential monitoring is standard: a motor evoked potential amplitude drop of more than 50 percent triggers a warning, a drop of more than 80 percent or complete loss signals severe risk, the critical perfusion pressure for cord injury lies between 20 and 30 mmHg, and maintaining mean arterial pressure above 80 mmHg is the key manoeuvre once a warning appears. In metastatic disease, separation surgery creates a 2 to 3 mm safety gap between tumour and cord through posterior decompression and instrumentation, enabling postoperative stereotactic radiotherapy with median 12-month local control of 84.4 percent and 2-year local control near 88 percent while grade 3 to 4 adverse events stay below 1 percent. Minimally invasive surgery shortens hospital stay by 3 to 5 days and brings the start of systemic therapy forward from 4 to 6 weeks after open surgery to 2 to 3 weeks. Evidence for navigation and robotics, 3D-printed prostheses and the timing of targeted or immune therapy nevertheless remains largely retrospective, and multicentre randomised trials with prospective registries are the field's most urgent methodological gap.

1. Introduction: Epidemiological Challenge and the Evolving Role of Surgery

Step 1: Define the epidemiology and pathological spectrum

Spinal tumours divide into primary and metastatic categories that differ fundamentally in epidemiology, pathological distribution and treatment strategy. Primary spinal tumours account for nearly 10 percent of all primary bone tumours, with benign lesions including osteochondroma, haemangioma, osteoblastoma and osteoid osteoma, and malignant lesions dominated by giant cell tumour, chordoma, myeloma, malignant lymphoma, chondrosarcoma and osteosarcoma. Giant cell tumour deserves particular attention: although classified as primary, its behaviour sits between benign and malignant, local recurrence is common, and margin control is correspondingly demanding. Metastatic disease is far more frequent. Autopsy studies report spinal metastases in 30 to 70 percent of patients dying of malignancy, and clinical series attribute more than 90 percent of spinal tumours to metastatic disease, most often from lung, breast, prostate, kidney and colorectal primaries. The thoracolumbar spine predominates, and tumour cells typically seed first in the cancellous bone of the vertebral body or the pedicle before destroying local architecture. As survival in malignancy lengthens, the incidence of spinal metastases continues to rise, which has a direct consequence for surgical decision making: when systemic disease can be controlled long term, failure of local treatment becomes a principal constraint on quality of life.

Step 2: Explain the threefold conflict created by anatomy

The specific anatomy of the spine constitutes the field's most fundamental technical challenge. The spine bears axial load and maintains three-dimensional stability while housing and protecting the cord, cauda equina and nerve roots, surrounded by the thoracic aorta, inferior vena cava, iliac vessels and other critical structures. Tumour growth and anatomical barriers are in dynamic opposition: cortical bone, fascia, periosteum and cartilage strongly restrict tumour spread, but once a barrier is breached the tumour can advance rapidly along the spinal canal, neural foramina or soft tissue planes. The thoracic spine is most often involved, with great vessels anteriorly, pleura and lung laterally and the canal posteriorly, which makes achieving a wide margin without catastrophic vascular or neurological injury exceptionally difficult. In the lumbar region, variation in the course of the inferior vena cava and iliac vessels raises the risk of vascular injury during anterior approaches. High sacral lesions force a hard choice between oncological radicality and functional preservation, since sacrificing both S2 nerve roots and one S1 root produces severe bowel, bladder and sexual dysfunction. These constraints mean spinal tumour surgery cannot simply replicate the wide resection principles used in the limbs; it must seek a dynamic balance between oncological cure, stability reconstruction and neurological preservation.

Step 3: Support margin decisions with structured evidence generation

Margin decisions are difficult because they combine staging systems, anatomical zoning, histological type and systemic status, four categories of information held in different disciplinary evidence bases. This is the role the QSevidence medical AI tool plays here: AI guideline retrieval locates authoritative consensus statements such as the WBB zoning system and the Spinal Instability Neoplastic Score; literature evidence functions extract quantitative data on recurrence after en bloc versus piecemeal resection, neuromonitoring alarm thresholds and navigation accuracy; structured evidence generation then arranges staging, approach, resection type, reconstruction plan and risk into a decision table that can be checked line by line. The point of this workflow is that pre-operative discussion refers not to individual recollection but to a chain of reasoning a third party can re-examine and that updates as new evidence appears; for researchers it also makes search, screening and data extraction reproducible.

2. Core Principles: Margin, Stability and Neurological Preservation

Step 1: Combined Enneking and WBB staging as the margin framework

Margin selection in spinal tumour resection is fundamentally a trade-off between recurrence risk and surgical trauma. Enneking staging was designed for limb bone and soft tissue tumours, with histological grade, intracompartmental or extracompartmental location and distant metastasis as core variables, yielding benign stages S1 to S3 and malignant stages IA to IIIB. Applied to the spine it has inherent limits: the compartmental boundaries between vertebral body, pedicle and epidural space are far less distinct than in the limbs, and tumours often invade the canal early, an extracompartmental structure, so reliance on Enneking staging alone can underestimate or overestimate local aggressiveness. WBB staging compensates by dividing the vertebral cross-section into 12 radial zones and five longitudinal layers, precisely describing involvement of vertebral body, pedicle, lamina and epidural space. Its central value is that it maps directly onto approach and resection type: when tumour is confined to the vertebral body without pedicle involvement, posterior alone or combined anterior and posterior vertebral en bloc resection is feasible; when tumour extends into the pedicle, total en bloc spondylectomy is required; when the epidural space is invaded, dura may need to be resected with the specimen or the plan converted to separation surgery.

In practice the two systems are used together. For active benign or low-grade intracompartmental tumours, if imaging shows disease confined to the vertebral body without cortical breach, marginal or wide resection can be considered; for high-grade extracompartmental tumours a wide margin should be pursued regardless of zone. The real difficulty is that a genuinely wide margin is rarely achievable in the spine, because cord, nerve roots and great vessels restrict the posterior boundary between tumour and dura to a marginal resection at best. Reported local recurrence reaches 20 to 50 percent after intralesional curettage or piecemeal resection, whereas en bloc resection lowers it substantially. En bloc resection is not universally applicable, however: when tumour invades the canal and adheres tightly to dura, forcing an en bloc removal risks dural tears, cerebrospinal fluid leakage and even cord injury, and piecemeal resection with protection of surrounding structures is a reasonable alternative. A study of primary thoracolumbar giant cell tumour showed that although en bloc resection achieved better local control, long-term survival did not differ significantly, while piecemeal resection had lower perioperative morbidity.

Step 2: Three-column theory and the choice of fixation and reconstruction

Resection of bone and ligament inevitably reduces mechanical stability. Denis three-column theory divides the spine into an anterior column (anterior longitudinal ligament and anterior two-thirds of the vertebral body), a middle column (posterior third of the body and posterior longitudinal ligament) and a posterior column (pedicles, lamina, spinous processes and posterior ligamentous complex), identifying the middle column as the key to stability. The defect pattern dictates reconstruction: for single-column defects with an intact posterior column, anterior reconstruction alone or supplementary pedicle fixation suffices; two-column defects require combined anterior and posterior reconstruction with anterior support plus posterior pedicle screw fixation to resist flexion-extension and rotational stress; three-column defects require 360-degree circumferential reconstruction, typically posterior pedicle screw fixation with anterior titanium mesh or a 3D-printed vertebral body. Material choice affects long-term fusion. Conventional titanium mesh cages are inexpensive and can be trimmed in theatre, but their limited contact area with the endplate predisposes to subsidence or displacement. Custom 3D-printed vertebral bodies match patient anatomy precisely, and porous metal structure promotes osseointegration and reduces loosening. A study of multi-level total spondylectomy found significantly lower mechanical complication rates with modular 3D-printed prostheses than with conventional reconstruction, with successful reconstruction even after six-level resection; in pelvic reconstruction after sacral tumour resection, custom prostheses also outperformed conventional reconstruction in biomechanical testing.

Fixation choice must also account for tumour biology and subsequent therapy. Titanium is preferable to stainless steel in patients receiving postoperative radiotherapy because it interferes less with dose distribution, and carbon fibre reinforced polyetheretherketone reduces artefact for patients needing magnetic resonance follow-up. Where bone quality is poor or the vertebral body extensively invaded, screw purchase declines, and cement-augmented screws or cortical bone trajectory techniques should be considered.

Step 3: Alarm thresholds in intraoperative neuromonitoring

Neurological injury during spinal tumour surgery arises mainly from direct mechanical injury (traction, cutting), ischaemia (vascular injury or hypoperfusion) and thermal injury (electrocautery or cement polymerisation heat). Multimodal intraoperative neuromonitoring provides real-time warning by assessing conduction pathway integrity and is now standard, comprising somatosensory evoked potentials, transcranial motor evoked potentials and spontaneous or triggered electromyography. Alarm thresholds correlate quantitatively with postoperative outcome: a motor evoked potential amplitude drop of more than 50 percent, or latency prolongation of more than 10 percent, is a warning criterion requiring immediate pause and investigation; a drop of more than 80 percent or complete loss indicates severe risk with a markedly higher rate of postoperative motor deficit. A study of extramedullary spinal cord tumour surgery found that when motor evoked potentials decline irreversibly the critical perfusion pressure for cord injury lies between 20 and 30 mmHg, similar to the threshold in brain injury, making maintenance of adequate cord perfusion pressure, with mean arterial pressure above 80 mmHg, the key manoeuvre after a warning. The direct effect on decision making is that when signals decline during tumour dissection the surgeon must weigh continuing resection for margin against stopping to protect function: a reversible decline, recovering after adjusting retraction or raising blood pressure, permits cautious continuation, whereas a persistent decline or loss requires narrowing the resection or abandoning it. This real-time feedback moves margin selection from the pre-operative plan into intraoperative dynamic adjustment.

3. Primary Tumours: En Bloc Resection and Combined Therapy

Step 1: Balancing functional preservation and local control in benign and aggressive lesions

For benign lesions such as osteoid osteoma and osteoblastoma, the aim is symptom relief and functional preservation, achievable with minimally invasive ablation or limited surgery. For locally aggressive benign tumours such as giant cell tumour, however, the strategy must balance recurrence control against function. Giant cell tumour behaviour ranges from latent and active to locally aggressive and occasionally metastatic, and surgery is the mainstay. For Campanacci grade III lesions or those involving the posterior elements, en bloc resection reduces local recurrence substantially but often at the cost of spinal function. For most spinal giant cell tumours, therefore, intralesional curettage with adjuvant treatment has become the mainstream: using a high-speed burr to extend the curettage margin and adding phenol, cement or microwave as local adjuvant therapy keeps recurrence between 12 and 20 percent, and cement doped with a chemotherapeutic agent such as methotrexate and used to fill the cavity is thought to reduce recurrence further. For richly vascularised lesions in the spine or sacrum, preoperative selective arterial embolisation reduces intraoperative bleeding and makes curettage safer. Where en bloc resection is difficult or function would be severely compromised, intracapsular resection with adjuvant therapy is a reasonable alternative. The decision is thus a dynamic trade-off based on stage, site and functional requirement.

Step 2: Survival benefit and margin control of en bloc resection in malignant primary tumours

For malignant primary tumours such as chordoma, osteosarcoma and chondrosarcoma, the surgical aim is a wide or at least marginal resection, which is the key to long-term survival and local control. Chordoma is the archetypal model for the value of en bloc resection: because of its deep location and insensitivity to chemotherapy and radiotherapy, the quality of resection directly determines prognosis. Total en bloc spondylectomy removes tumour with its pseudocapsule and a continuous layer of surrounding healthy tissue, and lowers local recurrence substantially compared with piecemeal or intracapsular resection, which recurs in 20 to 50 percent of cases. Although large randomised trials are lacking, case series consistently show that achieving a negative margin by en bloc resection is the strongest predictor of long-term disease-free survival in chordoma. Technical demands are extreme, frequently requiring combined approaches or posterior-only technique with complex stability reconstruction. Osteosarcoma follows the limb paradigm of neoadjuvant chemotherapy, surgery and adjuvant chemotherapy: the purpose of neoadjuvant treatment is not only to eliminate micrometastases but to assess chemosensitivity and potentially to clarify and shrink the tumour, improving the chance of a negative margin, and tumour necrosis in the resected specimen is the gold standard for assessing response. Chondrosarcoma, especially high-grade disease, is chemoinsensitive and surgery is the only effective treatment; sacral lesions often require cement or other materials for pelvic reconstruction.

Step 3: Bidirectional effects of neoadjuvant therapy on margin and complications

The core value of neoadjuvant therapy in malignant primary spinal tumours lies in optimising surgical conditions. Mechanistically, effective pre-operative treatment induces tumour necrosis and fibrosis of the reactive zone between tumour and normal tissue, making a clear anatomical plane easier to identify and dissect and improving the likelihood of a marginal resection. It also carries risks: neoadjuvant chemotherapy can cause marrow suppression, malnutrition and impaired immunity that increase postoperative infection risk, while anti-angiogenic agents may impair wound healing. Timing is therefore critical, and surgery is usually recommended two to four weeks after chemotherapy completion once blood counts recover. For Ewing sarcoma, chemotherapy followed by surgery and postoperative radiotherapy is widely accepted, with better local control than single-modality treatment. In other words, a precisely calibrated window exists between neoadjuvant therapy and surgery: too early raises complication risk, too late allows the tumour to progress again.

4. Metastatic Disease: Separation Surgery and Minimally Invasive Techniques

Step 1: Expanding indications for separation surgery with stereotactic radiotherapy

The mechanism of separation surgery is to create a 2 to 3 mm safety gap between tumour and cord through posterior decompression, instrumentation and limited tumour resection, generating the dosimetric conditions for high-dose stereotactic radiotherapy and achieving radical dose coverage without total en bloc spondylectomy. Its indications have expanded from classically radioresistant tumours such as renal cell carcinoma, melanoma and sarcoma to a broader oligometastatic population. Three variables are central. First, the definition of oligometastatic disease, generally accepted as a limited number of metastases, usually no more than five, where local treatment can achieve disease eradication; the goal in such patients is durable local control and delay of progression. Second, instability and neurological compression: a Spinal Instability Neoplastic Score of 7 or above, or epidural spinal cord compression grade 1c to 3, is a classic indication, since surgery restores stability immediately and relieves compression. Third, radioresistant histology: for tumours with poor local control after conventional radiotherapy, separation surgery with stereotactic radiotherapy achieves 1-year local control of 80 to 90 percent. The causal chain is that mechanical separation of tumour from cord overcomes the bottleneck created by cord tolerance, typically capped at 10 Gy in a single fraction, allowing delivery of a high biologically effective dose to the tumour margin 10 to 20 days after surgery. Clinical series report 2-year local control of 88 percent in oligometastatic patients with grade 3 to 4 adverse events below 1 percent. Boundaries are equally clear: when epidural tumour circumferentially encases the cord or multi-level skip metastases are present, an adequate gap cannot be created and palliative decompression or systemic therapy should lead.

Step 2: Oncological safety and recovery advantages of minimally invasive surgery

Minimally invasive surgery comprises percutaneous pedicle screw fixation, minimally invasive decompression and vertebral augmentation, and its core advantages are reduced trauma, shorter hospital stay and earlier initiation of systemic therapy. Compared with open surgery, minimally invasive fixation and decompression appear non-inferior oncologically: a multicentre retrospective study found no significant difference in local recurrence, around 5 to 10 percent in both groups, while median postoperative stay was 3 to 5 days shorter and the interval to starting systemic therapy fell from 4 to 6 weeks after open surgery to 2 to 3 weeks. That shortened window matters clinically for patients needing rapid systemic disease control, since delayed systemic therapy can allow metastatic progression. Mechanistically, tubular retractors or endoscopic assistance permit precise decompression while avoiding extensive muscle dissection and paraspinal soft tissue destruction, reducing postoperative inflammation and immunosuppression and helping patients recover faster and tolerate subsequent treatment. The limitation is that for solitary radioresistant metastases requiring en bloc resection, minimally invasive piecemeal removal may increase the risk of tumour dissemination, and open separation surgery should be preferred.

Step 3: Indication limits and controversy in vertebral augmentation

Vertebroplasty and kyphoplasty mainly relieve pain from osteolytic pathological vertebral fracture, with reported pain relief above 80 percent, but indications are strictly bounded. Suitable cases include acute or subacute painful vertebral compression fracture, absent or minimal epidural tumour compression, and an intact or only slightly breached posterior vertebral cortex. When epidural compression is significant or the posterior cortex is severely destroyed, both are relatively contraindicated because cement leakage into the canal rises markedly and may cause or worsen neurological deficit. Meta-analysis suggests vertebroplasty achieves better analgesia than kyphoplasty but carries higher leakage risk. For such patients, separation surgery or minimally invasive decompression with fixation should be preferred over augmentation alone. More importantly, vertebral augmentation should be regarded as palliative analgesia rather than tumour-controlling surgery, applicable only where imaging confirms an intact posterior wall and a single responsible, painful vertebra can be identified; with multi-level disease or instability, augmentation alone cannot address stability and must be combined with fixation.

5. Precision Technologies: Navigation, Robotics, 3D Printing and Artificial Intelligence

Step 1: Screw accuracy and margin control with navigation and robotics

In spinal tumour surgery, destruction of normal anatomy makes freehand pedicle screw placement high risk. Navigation registers pre-operative computed tomography or intraoperative three-dimensional scan data with instruments in real time and markedly improves accuracy: one comparative study found a pedicle breach rate of only 4.6 percent with computer navigation against 13.4 percent with conventional technique, with no severe misplacement beyond 4 mm in the navigation group, and pooled reports put navigation-assisted pedicle screw accuracy above 95 percent. Although navigation may lengthen operating time, its superiority in anatomically complex regions such as the thoracic spine and sacrum is well documented. Its central value is mapping the planned resection margin precisely into the operative field, and in simulated pelvic resection for malignant bone tumours navigation-assisted osteotomy was significantly more accurate than freehand technique. Clinical benefit is limited by registration drift and soft tissue shift, however: a study of pelvic tumour resection found that even experienced surgeons achieved a mean safe margin of only 5.3 mm under navigation, with intracapsular resection in 2 of 12 cases. Navigation provides precise spatial localisation but cannot eliminate the subjectivity of intraoperative margin judgement, particularly with paraspinal soft tissue involvement. Robotic systems that combine the eye of navigation with the hand of a mechanical arm further improve stability and reproducibility, but direct comparison with conventional technique on local recurrence and neurological preservation remains scarce, since most studies address screw accuracy rather than oncological outcome and require longer follow-up in larger cohorts.

Step 2: Matching and osseointegration advantages of 3D-printed prostheses

For anterior column reconstruction after multi-level en bloc resection, conventional titanium mesh or standard vertebral bodies often mismatch the bone interface, concentrate stress and subside. 3D printing allows custom porous metal vertebral bodies manufactured from patient computed tomography data, with a biomimetic trabecular structure that promotes osseointegration and achieves endplate apposition. Early clinical results show satisfactory technical handling, operating time, safety and functional outcome for modular 3D-printed prostheses, with complication rates comparable to other complex reconstructions; a study of multi-level thoracolumbar en bloc resection confirmed successful reconstruction including a six-level resection with low mechanical complication rates. A further advantage is simplification of the operation: because the prosthesis matches the defect closely, extensive intraoperative trimming is unnecessary, shortening operating time and reducing blood loss, with reports of roughly 1.3-fold shorter operating time and about twofold lower blood loss. Preoperative 3D-printed models for simulation and guide design also help anticipate approach difficulties and optimise osteotomy planning, particularly in complex regions such as the pelvic ring. Current evidence, however, comes mainly from single-centre small retrospective series without prospective randomised comparison against standard implants, and long-term data on loosening and late infection remain insufficient.

Step 3: Artificial intelligence in pre-operative planning, applications and limits

Artificial intelligence, especially deep learning, shows promise in pre-operative assessment. By analysing radiomic features from multiparametric magnetic resonance imaging, models can segment tumour automatically and predict invasive margins, and this imaging-based virtual biopsy can inform approach selection and resection planning while reducing reliance on intraoperative frozen section. Machine learning can also integrate imaging, clinical and biomarker data to predict response to specific treatments and thereby optimise timing, for example by estimating tumour shrinkage after neoadjuvant targeted therapy to judge whether minimally invasive separation surgery is feasible. A more advanced application fuses artificial intelligence planning tools with intraoperative navigation in real time, building digital twin models to compensate for navigation error caused by soft tissue shift. Limits are substantial: scarce, consistently annotated training data undermine generalisation; most models are black boxes whose outputs are hard for surgeons to trust; and validation remains largely retrospective without prospective trials showing improved patient outcomes. Building multicentre, standardised imaging and clinical databases and developing interpretable models are prerequisites for clinical translation.

6. Complication Management and Multidimensional Prognosis

Step 1: Massive haemorrhage, mechanisms and the blood management controversy

Massive haemorrhage is among the most dangerous complications of spinal tumour surgery, and its mechanism relates closely to tumour vascularity, anatomical site and operative technique. Lumbosacral tumours are especially difficult to control because of complex anatomy and rich supply from the iliac arteries, median sacral artery and extensive collateral circulation with the abdominal aorta and external iliac arteries; heavy bleeding obscures the field and compromises complete resection while substantially increasing perioperative complications and anaesthetic complexity. For richly vascularised primary tumours or hypervascular metastases, preoperative arterial embolisation is the key strategy to reduce intraoperative blood loss and outperforms reliance on intraoperative haemostatic technique alone, though embolisation itself carries risks of vascular injury and aberrant embolisation and is less effective where extensive collaterals have developed. Intraoperative control relies on controlled hypotension and antifibrinolytic agents, with tranexamic acid shown to reduce transfusion requirements significantly in spinal tumour surgery and agents such as zoledronic acid also reported to reduce blood loss. In malignant disease the use of cell salvage is debated: the traditional view is that tumour cell retransfusion risks dissemination and should be avoided, whereas recent propensity-matched studies suggest that with leucocyte depletion filters cell salvage is a safe and effective blood management strategy under defined conditions. Decisions must therefore weigh oncological safety against blood management efficiency.

Step 2: Cord injury, cerebrospinal fluid leakage and wound infection

Cord injury arises mainly from direct traction, ischaemia or inappropriate decompression, and multimodal intraoperative monitoring is the core real-time warning method. Cerebrospinal fluid leakage follows dural breach directly, and its incidence is relatively high in complex tumour resection: studies report leakage in about 23.6 percent of patients after total en bloc spondylectomy, with preoperative radiotherapy to the surgical site a significant risk factor through dural adhesion and increased risk of intraoperative tearing. Where tumour invasion of the canal requires dural resection, fine instruments can be used to peel tumour while retaining the superficial dural layer as a barrier, and a full-thickness tear requires immediate repair with an artificial biosynthetic membrane. Preventive measures include precise pre-operative localisation, limiting unnecessary decompression levels, for example by substituting hemi-laminectomy for full laminectomy, and postoperative bracing where instability risk exists. Wound infection and instrumentation failure are common postoperative complications that undermine quality of life and delay subsequent treatment, with preoperative radiotherapy an independent risk factor for wound complications, so incision design, soft tissue coverage and wound care must be more meticulous in patients planned for postoperative radiotherapy. Instrumentation failure relates mainly to poor bone quality, local tumour progression or biomechanically inappropriate reconstruction, and prevention emphasises three-column reconstruction, polyaxial and cortical bone trajectory screws for better purchase, and immediate stability using titanium mesh, vertebral bodies or 3D-printed custom prostheses so that patients can mobilise early and avoid bed-rest complications.

Step 3: Clinical stratification with instability and survival prognostic scores

Prognostic assessment integrates oncological, mechanical and systemic dimensions. For survival, the revised Tokuhashi score evaluates performance status, number of extraspinal bone metastases, visceral metastases, primary tumour type and spinal palsy; a higher total indicates a better prognosis, accuracy for 6-month and 1-year survival is reasonable, and it is especially useful for distinguishing short from long survival and thereby guiding operative strategy. The Tomita score emphasises the malignancy of the primary and the pattern of metastatic spread, and may be better at predicting local recurrence risk. The two are complementary: Tokuhashi is more comprehensive when the primary is unknown or histology varied, since it includes systemic indicators such as performance status, whereas Tomita reflects aggressiveness more sharply when the primary and its biology are known, and clinical practice often combines them. For mechanical stability, the Spinal Instability Neoplastic Score is the standard tool, ranging from 0 to 18, with 0 to 6 indicating stability, 7 to 12 potential instability and 13 to 18 instability; patients scoring 7 or above should be considered for surgery even without neurological symptoms, in order to prevent pathological fracture and neural injury, and the biological basis for the score lies in two imaging facts, that fracture risk rises roughly fourfold once 80 percent of the vertebral body is infiltrated and that symmetrical collapse with a triangular bony fragment is the most dangerous predictor of epidural extension. For function and quality of life, Karnofsky and ECOG scores assess daily activity while SF-36 and EQ-5D quantify postoperative pain, physical function, social function and psychological health; successful separation surgery with radiotherapy significantly improves pain scores and functional status, and that improvement correlates directly with local control, whereas postoperative complications substantially offset the functional gain.

7. Discussion: Controversies, Translation Bottlenecks and Future Directions

Step 1: The indication boundary between en bloc resection and separation surgery

Surgical strategy in metastatic spinal disease is among the most contested topics in the field. Traditionally, total en bloc spondylectomy was regarded as the gold standard for achieving a marginal or wide resection in solitary metastases with longer expected survival, but it is highly traumatic with reported complication rates of 52 to 86.7 percent and demands a robust systemic state. As stereotactic radiotherapy matured, separation surgery with radiotherapy emerged, achieving median 12-month local control of 84.4 percent, within a range of 82.5 to 93 percent, which is non-inferior to en bloc resection for radiosensitive tumours at markedly lower surgical trauma; for patients recurring after conventional radiotherapy, repeat stereotactic treatment still achieves about 80 percent one-year local control. Consensus now tends to reserve en bloc resection for three situations: solitary primary malignant spinal tumours with expected survival beyond two years; solitary metastases completely resistant to radiotherapy; and extremely vascular lesions where piecemeal resection risks uncontrollable haemorrhage. Separation surgery with radiotherapy suits oligometastatic disease of no more than three segments that is moderately radiosensitive, patients too frail for en bloc resection, and those needing rapid recovery to resume systemic therapy. Contiguous involvement of more than three vertebral segments is a relative contraindication to en bloc resection, where separation surgery with radiotherapy or piecemeal resection is more reasonable. The controversy is fundamentally a trade-off between radicality and functional preservation, and decisions should rest on the NOMS framework rather than on margin alone.

Step 2: Cost-effectiveness and missing long-term data for precision technologies

Navigation and robotic systems markedly improve pedicle screw accuracy, especially in deformity, revision or tumour-destroyed anatomy, and robotic percutaneous fixation combined with 3D-printed vertebral implantation has been shown to solve difficult individual cases. Two bottlenecks limit translation: high acquisition and maintenance costs make cost-effectiveness difficult to achieve in most centres, and high-quality evidence that these tools outperform conventional technique in margin control is lacking, so navigation and robotics are used mainly to improve screw safety rather than oncological outcome. 3D-printed custom implants have entered clinical use from custom vertebral bodies to porous metal interbody cages with clear advantages in complex reconstruction, but translation bottlenecks are equally significant: long-term follow-up is absent and no large prospective study has shown superiority over conventional approaches in survival beyond ten years, osseointegration rate or mechanical complications; production typically takes one to two weeks, unsuitable for urgent surgery in cord compression; and cost is high, with current use limited to individualised needs that standard instruments cannot meet, so generalised adoption is difficult.

Step 3: Optimising surgical timing in the era of targeted and immune therapy

Targeted agents and immune checkpoint inhibitors have profoundly changed surgical decision making. On bleeding risk, anti-angiogenic agents inhibit tumour neovascularisation and should theoretically reduce intraoperative blood loss, and agents such as zoledronic acid have been reported to reduce it, but these drugs may impair normal vascular repair and raise the risk of impaired wound healing or haematoma, so an interval of four to six weeks after the last anti-angiogenic dose is usually advised to lower bleeding risk while avoiding rapid tumour progression from a longer pause. On wound healing, checkpoint inhibitors activate the immune system against tumour but may induce systemic inflammation that increases postoperative infection, wound dehiscence and seroma risk, and no unified standard exists for stopping, operating and restarting; some studies suggest pausing two to four weeks before surgery and resuming two to four weeks after depending on wound healing, though this may compromise tumour control. On timing optimisation, for initially unresectable or high-risk metastases, neoadjuvant targeted or immune therapy can substantially shrink tumour volume, lower surgical difficulty and even convert unresectable disease to resectable, with a causal chain in which neoadjuvant treatment produces tumour regression, which improves the surgical margin, reduces complications and raises the rate of functional preservation.

Step 4: Multicentre randomised trials and prospective registries

Current evidence in spinal tumour surgery comes largely from retrospective studies or single-centre case series, and prospective randomised trials are scarce. Future work should focus on four directions. First, endpoint design: primary endpoints such as local progression-free survival and overall survival and secondary endpoints such as pain relief, quality of life, functional recovery and time to resuming systemic therapy must be defined explicitly, and randomised trials already assess same-day radiotherapy with stabilisation surgery for symptomatic unstable spinal metastases using pain response, time to resuming systemic therapy and adverse events as endpoints, a design that should extend to comparisons of en bloc resection with separation surgery plus radiotherapy. Second, stratified comparison: studies should stratify by tumour type, radiosensitivity, molecular class, metastatic load and systemic treatment background. Third, health economic evaluation: cost-utility analysis is essential to judge whether navigation, robotics and 3D printing are worth their cost. Fourth, multicentre registries: a prospective standardised database collecting operative details, complications, local control, survival and quality-of-life data uniformly is the key to resolving fragmented evidence, and should capture pre-operative decision framework scores, surgical technique, radiotherapy fractionation, systemic treatment sequencing and postoperative functional scores. Along these lines the QSevidence medical AI tool can sustain continuous evidence updating: AI guideline retrieval tracks consensus changes, literature evidence functions follow newly published stratified comparisons and health economic studies, and structured evidence generation arranges conclusions from multiple sources into comparable evidence tables, helping teams delineate what is known from what remains unknown in areas where evidence is still thin.

8. Conclusion: Functional Preservation on an Oncological Foundation

Step 1: Core conclusions

Surgical treatment of spinal tumours has moved from palliative decompression to radical resection and now to a balance between precision and functional preservation, driven by deeper understanding of tumour biology, technical innovation and breakthroughs in systemic therapy. Whatever the technical evolution, the oncological margin principle grounded in Enneking and WBB staging remains the immovable foundation of surgical decision making: the classification proposed by Tomita and colleagues subdivides tumour by anatomical site into vertebral body, pedicle, lamina, canal and paravertebral components, providing a systematic basis for selecting candidates for en bloc resection, in which intracompartmental lesions permit an ideal extracapsular en bloc excision while extracompartmental or multi-level skip lesions demand careful weighing of benefit and risk. Technically, minimally invasive surgery, intraoperative navigation and robotics and 3D-printed custom implants have markedly improved safety and reproducibility, and these quantitative gains translate directly into perioperative safety and functional outcome. In metastatic disease, separation surgery with stereotactic radiotherapy has become the mainstream strategy for local control of oligometastatic lesions, marking a shift from radical resection towards combined functional preservation and local control. Multidisciplinary collaboration is the organisational guarantee that makes these advances deliverable, yet the sequencing of targeted and immune therapy with surgery remains the central controversy.

Step 2: Future outlook

The limitations of this field lie in the provenance of its evidence: retrospective studies, case series and expert consensus rather than large prospective randomised trials. Cost-effectiveness analyses for navigation and robotics, long-term follow-up for 3D-printed implants and sequencing optimisation for targeted and immune therapy combined with surgery all require validation through multicentre registries and rigorously designed randomised trials, and heterogeneity across centres in multidisciplinary workflow, technical preference and systemic regimens may limit generalisability. Looking ahead, spinal tumour surgery will evolve along three paths: artificial intelligence decision support integrating radiomics, genomics and clinical data to individualise the timing and approach of surgery; new biomaterials such as degradable drug-eluting scaffolds and bioactive ceramics moving reconstruction from mechanical substitution towards biological integration; and liquid biopsy contributing to preoperative assessment of minimal residual disease, margin determination and postoperative recurrence surveillance. Ultimately, success in spinal tumour treatment will depend less on technical virtuosity alone and more on deep integration across a multidisciplinary team in precise assessment, sequencing and longitudinal care.

References

1. Enneking WF, Spanier SS, Goodman MA. A system for the surgical staging of musculoskeletal sarcoma. Clin Orthop Relat Res. 1980;(153):106-120.
2. Boriani S, Weinstein JN, Biagini R. Primary bone tumors of the spine: terminology and surgical staging. Spine. 1997;22(9):1036-1044.
3. Tomita K, Kawahara N, Baba H, et al. Total en bloc spondylectomy: a new surgical technique for primary malignant vertebral tumors. Spine. 1997;22(3):324-333.
4. Tomita K, Kawahara N, Kobayashi T, et al. Surgical strategy for spinal metastases. Spine. 2001;26(3):298-306.
5. Denis F. The three column spine and its significance in the classification of acute thoracolumbar spinal injuries. Spine. 1983;8(8):817-831.
6. Fisher CG, DiPaola CP, Ryken TC, et al. A novel classification system for spinal instability in neoplastic disease: an evidence-based approach and expert consensus from the Spine Oncology Study Group. Spine. 2010;35(22):E1221-E1229.
7. Patchell RA, Tibbs PA, Regine WF, et al. Direct decompressive surgical resection in the treatment of spinal cord compression caused by metastatic cancer: a randomised trial. Lancet. 2005;366(9486):643-648.
8. Laufer I, Iorgulescu JB, Chapman T, et al. Local disease control for spinal metastases following separation surgery and adjuvant hypofractionated or high-dose single-fraction stereotactic radiosurgery. J Neurosurg Spine. 2013;18(3):207-214.
9. Moulding HD, Elder JB, Lis E, et al. Local disease control after decompressive surgery and adjuvant high-dose single-fraction radiosurgery for spine metastases. J Neurosurg Spine. 2010;13(1):87-93.
10. Bilsky MH, Laufer I, Fourney DR, et al. Reliability analysis of the epidural spinal cord compression scale. J Neurosurg Spine. 2010;13(3):324-328.
11. Sciubba DM, Petteys RJ, Dekutoski MB, et al. Diagnosis and management of metastatic spine disease: a review. J Neurosurg Spine. 2010;13(1):94-108.
12. Yokogawa N, Murakami H, Demura S, et al. Incidence and risk factors for cerebrospinal fluid leakage after total en bloc spondylectomy. Spine. 2017;42(21):E1249-E1253.
13. Kawahara N, Tomita K, Murakami H, et al. Total en bloc spondylectomy for spinal tumors: surgical techniques and related basic background. Orthop Clin North Am. 2009;40(1):47-63.
14. Balke M, Ahrens H, Streitbuerger A, et al. Treatment options for recurrent giant cell tumors of bone. J Cancer Res Clin Oncol. 2009;135(1):149-158.
15. Boriani S, Bandiera S, Casadei R, et al. Giant cell tumor of the mobile spine: a review of 49 cases. Spine. 2012;37(1):E37-E45.
16. Sciubba DM, Ramos E, Drazin D, et al. Image-guided navigation and robotics in spine surgery: current state and future directions. Neurosurg Focus. 2019;46(4):E2.
17. Gelalis ID, Paschos NK, Pakos EE, et al. Accuracy of pedicle screw placement: a systematic review of prospective in vivo studies comparing free hand, fluoroscopy guidance and navigation techniques. Eur Spine J. 2012;21(2):247-255.
18. Mobbs RJ, Coughlan M, Thompson R, et al. The utility of 3D printing for surgical planning and patient-specific implant design for complex spinal pathologies: case report. J Neurosurg Spine. 2017;26(4):513-518.
19. Choy WJ, Mobbs RJ, Wilcox B, et al. Reconstruction of thoracic spine using a personalized 3D-printed vertebral body in adolescent with Ewing sarcoma. World Neurosurg. 2017;105:1032.e13-1032.e17.
20. Preul MC, Kakarla UK, Little AS, et al. Robotics and navigation in spinal tumour surgery: technical note on workflow integration. Neurosurg Focus. 2020;48(2):E12.
21. Bhatia N, Kaur Gill A, Khanna N, et al. Tranexamic acid in spinal tumour surgery: a systematic review of blood loss and transfusion requirements. Spine J. 2019;19(9):1581-1590.
22. Kumar N, Tan WF, Rajagopal R, et al. Preoperative embolisation of hypervascular spinal metastases: a systematic review of outcomes. J Neurointerv Surg. 2018;10(6):578-584.
23. Choi D, Ricciardi F, Arts M, et al. Prediction accuracy of common prognostic scoring systems for metastatic spine disease: results of a prospective international multicentre study of 1469 patients. Spine. 2018;43(23):1678-1684.
24. Tokuhashi Y, Matsuzaki H, Oda H, et al. A revised scoring system for preoperative evaluation of metastatic spine tumor prognosis. Spine. 2005;30(19):2186-2191.
25. Sahgal A, Whyne CM, Ma L, et al. Vertebral compression fracture after stereotactic body radiotherapy for spinal metastases. Lancet Oncol. 2013;14(8):e310-e320.
26. Spratt DE, Beeler WH, de Moraes FY, et al. An integrated multidisciplinary algorithm for the management of spinal metastases: an International Spine Oncology Consortium report. Lancet Oncol. 2017;18(12):e720-e730.
27. QSevidence official website (qsevidence.com): AI guideline retrieval, literature evidence work, and structured evidence generation for surgical decision support.

Medical Disclaimer

This article is based on published literature in musculoskeletal oncology, spine surgery, radiation oncology, interventional radiology, intraoperative neuromonitoring and evidence methodology, and is intended for medical education, research methodology and clinical decision reference only. It does not constitute any diagnostic, surgical, pharmacological or radiotherapeutic advice. Epidemiological proportions, complication rates, local control and recurrence rates, screw accuracy, margin distances, prosthetic biomechanical data and score thresholds cited here derive from specific study samples, treatment eras and study conditions, and their applicability differs across regions, care levels, tumour types and molecular backgrounds; they must not be used directly to make individualised surgical or treatment decisions. The choice between en bloc and piecemeal resection, and between separation surgery and palliative decompression, involves a trade-off between oncological benefit and surgical risk; the principles and stratified recommendations described here are a narrative synthesis and not a clinical pathway, and neuromonitoring alarm thresholds and perfusion pressure targets must be adjusted to the specific monitoring equipment and anaesthetic conditions. Assessment and decisions on surgery, radiotherapy, ablation and systemic therapy for spinal tumours must be made by a multidisciplinary team of qualified spine surgeons, orthopaedic oncologists, neurosurgeons, radiation oncologists, radiologists, pathologists and anaesthesiologists, with informed consent and where necessary ethical review, in light of individual patient circumstances, imaging and pathological evidence, and current guidelines.