Posterior-only 2-level vertebrectomy and fusion in a medically complex patient with lumbar metastasis: illustrative case

Abstract

BACKGROUND

Spinal metastases are commonly seen in patients with cancer and often indicate a poor prognosis. Treatment can include curative or palliative surgery, chemotherapy, and radiation therapy. The surgical approach varies widely on the basis of the affected region of the spine, the location of the tumor (anterior versus posterior), the goal of surgery, the health of the patient, and surgeon preference.

OBSERVATIONS

The authors present a case of a 68-year-old male with intractable lower-back pain and substantially diminished ambulation. Diagnostic imaging revealed a lumbar metastasis from a cholangiocarcinoma primary at L2–3 (4.5 cm anteroposterior × 5.7 cm transverse × 7.0 cm craniocaudal). The patient underwent a 2-level vertebrectomy with expandable cage placement and T10 to S2 fusion via a posterior-only approach. The patient regained much of his mobility and quality of life after the surgery.

LESSONS

Although this was a high-risk surgery, the authors show that a posterior-only approach can be used for lumbar vertebrectomies and fusion when necessary. Palliative surgeries carrying a high risk, especially in the setting of a limited prognosis, should include multidisciplinary deliberations and a thorough discussion of the risks and outcome expectations with the patient.

Spinal metastases are seen in 20%–40% of patients with cancer, with 20%–45% of cases involving the lumbar spine.^1,2 Prognosis is variable, depending on the tumor histology, systemic disease status, and comorbidities.^2–4 The operative approach depends on patient comorbidities, tumor location, and the goals of surgery. We describe a patient with a lumbar spine metastasis from cholangiocarcinoma and significant comorbidities who underwent a 2-level vertebrectomy with expandable cage placement and T10 to S2 fusion via a posterior-only approach.

Illustrative Case

A 68-year-old male initially presented with a 2-month history of progressively worsening lower-back pain refractory to opioid analgesics. His back pain was bilateral and radiating to both knees. His ambulatory status was significantly diminished from baseline, and he required a walker. His upper- and lower-extremity strength was intact with no focal deficits. There were no sensory abnormalities, and he was not complaining of radicular pain. The remainder of his neurological examination was unremarkable.

His medical history was significant for morbid obesity (body mass index 48.95 kg/m²), atrial fibrillation, arthritis, carpal tunnel syndrome, diabetes mellitus type 2, gastroesophageal reflux disease, hypertension, dyslipidemia, severe obstructive sleep apnea, prior tobacco abuse, restless leg syndrome, and spinal stenosis of the lumbar region. He also had a significant past surgical history that included a previous in situ thoracolumbar fusion for an unstable fracture 30 years prior, and he later underwent an L4–S1 decompressive laminectomy with bilateral L5–S1 foraminotomies.

Lumbar spine radiography and computed tomography (CT) showed a mass with bone destruction involving L2 and L3 (Fig. 1). Additionally, fractures of the L3 vertebral body, left L3 pedicle, and bilateral pars interarticularis were also seen. Lumbar magnetic resonance imaging with and without contrast showed an enhancing mass involving L2 and L3 and measuring 4.5 cm (anteroposterior) × 5.7 cm (transverse) × 7.0 cm (craniocaudal; Fig. 2) with epidural and paraspinal extension of this mass at the level of L3. Positron emission tomography revealed a hypermetabolic uptake of these lesions at L2 and L3. Further work-up identified a large hepatic mass, raising concern for a neoplastic process. Due to significant comorbidity and the likely necessity of a 2-level vertebrectomy requiring a combined anterior and posterior approach, the patient was referred to a tertiary care facility with capabilities and access surgeons to perform these approaches.

FIG. 1.

Preoperative CT of the lumbar spine showing a mass with bony destruction: sagittal sequence (A), axial sequence at the level of the L2 vertebral body (B), and axial sequence at the level of the L3 vertebral body (C).

FIG. 2.

Preoperative magnetic resonance imaging of the lumbar spine revealing the mass measuring 4.5 cm (anteroposterior) × 5.7 cm (transverse) × 7.0 cm (craniocaudal): sagittal short tau inversion recovery sequence (A) and axial T2-weighted sequence at the level of the L3 vertebral body (B).

Neurosurgical evaluation at the referral facility did not recommend surgical intervention and referred the patient back to our facility for oncological evaluation and treatment. The patient then represented to the emergency department with intractable back pain and weakness in his bilateral lower extremities. At this time, his lower-extremity deep tendon reflexes and strength were diminished (2/5 bilateral hip flexion, 4/5 bilateral knee extension). He could no longer ambulate. Significant discussion occurred among the patient, family, and multidisciplinary care team regarding the high morbidity and mortality associated with surgery and poor prognosis. The patient elected to undergo surgical intervention via a posterior-only approach to improve his mobility and enhance his quality of life.

Operative Approach

The patient was brought to the operating room and underwent an uneventful induction of general endotracheal anesthesia. Neuromonitoring was used throughout the entire procedure and was connected prior to positioning for preoperative baselines. The patient was positioned prone on a spinal Jackson table and was prepped and draped in a sterile fashion. We made a midline incision from T10 down to S2. Subcutaneous tissue dissection and subperiosteal technique were used to expose the thoracolumbosacral lamina in a typical fashion. We continued the exposure laterally over the facet joints to expose the transverse processes from T10 to S2. We obtained multiple intraoperative CT scans using O-arm technology (StealthStation, Medtronic) for the placement of pedicle screws from T10 to S1, skipping L2 and L3 bilaterally. Bilateral Medtronic Solera 5.5-mm titanium pedicle screws were inserted in the proper trajectory from T10 to S1, and we obtained a confirmatory O-arm scan to confirm proper placement of the surgical hardware. For pelvic fixation, we placed S2-alar-iliac Medtronic Ballast screws (8.5 mm × 80 mm) bilaterally by using an entry point between the S1 and S2 foramina. After screw placement, we then turned our attention to surgical decompression. We started by performing an L2 and L3 laminectomy using the combination of a high-speed microsurgical drill and Kerrison rongeurs. The L2 and L3 nerve roots were identified laterally, and wide foraminotomies were performed to expose the dorsal root ganglion on both sides for each nerve root. Using navigation as an aide, we then performed a bilateral L2 and L3 transpedicular decompression to remove the L2 and L3 pedicles on both sides, allowing access to the anterolateral lumbar spine. At this time, we could see fibrous tissue in the ventral epidural space infiltrating the vertebral body, which was biopsied for further pathological analysis. Using a 15-blade scalpel, the normal disc space at L1–2 was incised to create an annulotomy to facilitate the discectomy with pituitary rongeurs and curettes. This was repeated at the L3–4 disc space. Using the operative microscope, we then proceeded to perform the L2 and L3 vertebrectomy using a microsurgical drill, curettes, and pituitary rongeurs to remove fibrous soft tissue with the L2 and L3 vertebral bodies. During this process, significant bleeding was encountered from the tumor. Ultrasonic aspiration was used to further facilitate removal of the L2 and L3 vertebral body masses until there was circumferential decompression of the thecal sac at these segments and adequate space for placement of an expandable cage. The superior and inferior endplates were properly denuded of the fibrocartilaginous endplate. After hemostasis was maintained, surgical trialing for the expandable cage was performed. At this time, we realized that cage placement would be nearly impossible without sacrificing an L2 or L3 nerve root, which was not performed. Instead, the lateral half of the left L4 pedicle was removed with the drill to facilitate a more favorable angle for insertion. To help guide placement, a silk tie was placed in the inner aperture of the expandable cage to allow placement adjustments (Fig. 3). A Nuvasive X-Core expandable cage with a 41- to 67-mm expansion range and an 18-mm width was placed into the vertebrectomy defect. The Stealth navigation probe was placed on the posterior aspect of the cage, and a 41-mm positive projection was used to simulate the anterior extent of the cage. Under navigational guidance, the cage was inserted, and the silk tie was pulled posteriorly to allow the cage to swing into proper position. The cage was expanded until there was tight apposition of the vertebral endplates superiorly and inferiorly (Fig. 4). The cage position was confirmed and deemed appropriate based on anteroposterior and lateral fluoroscopic images. Placement of the titanium rods was then performed in a standard fashion, secured with set screws, and finally tightened appropriately (Fig. 3). The transverse processes and facets were decorticated with a high-speed air drill, and allograft was used for posterolateral arthrodesis. Two Jackson-Pratt drains were left in the epidural space. Hemostasis was maintained, and the wound was flushed with antibiotic irrigation. Vancomycin powder was placed in the wound. Surgical closure then proceeded in a standard fashion.

FIG. 3.

Illustration showing placement of the cage assisted by the silk tie (A, B) and the final hardware placement (C).

FIG. 4.

Placement of the vertebral cage visualized under intraoperative fluoroscopy: lateral view (A) and posterior view (B).

Postoperative Course

Postoperatively, the patient remained intubated after the procedure. He required hemodynamic support with vasopressors and packed red blood cell and platelet transfusions to stabilize his hemoglobin and correct a postoperative coagulopathy. Postoperative CT of the lumbar spine revealed adequate hardware placement (Fig. 5A and B). On postoperative day 5, the patient was extubated. Two weeks after surgery, he underwent palliative external beam radiation therapy to L1–4, 30 Gy in 10 fractions. During the admission, he had a gradual reduction in pain and improvement in his lower-extremity weakness. On postoperative day 44, he was discharged to a skilled nursing facility for rehabilitation and occupational therapy. His final pathology from surgery confirmed cholangiocarcinoma as the diagnosis.

FIG. 5.

Coronal (A) and sagittal (B) CT of the lumbar spine on postoperative day 1. Radiographs of the lumbar spine at approximately 1 year postoperatively with anteroposterior (C) and lateral (D) views. Both imaging modalities show stable hardware placement.

At 4 and 6 months postoperatively, he reported significant improvement in pain and quality of life. He was ambulating with a walker, could perform activities of daily living, and participated in physical therapy. Despite multimodal therapy, he continued to have metastatic spread to the left femur, left humerus, and left scapula. He underwent additional radiation to those areas and was being treated with durvalumab, cisplatin, and gemcitabine. At 1 year 6 days postoperatively, the patient was admitted for septic shock and ultimately died of cardiopulmonary arrest. Postoperative radiographs obtained shortly before the patient’s death continued to show stable hardware placement (Fig. 5C and D).

Patient Informed Consent

The necessary patient informed consent was obtained in this study.

Discussion

We describe a patient with significant comorbidities who underwent a 2-level vertebrectomy with expandable cage placement and T10 to S2 alar-iliac fusion via a posterior-only approach. Surgical fixation and multidisciplinary management postoperatively allowed the patient to have significant improvement in his mobility and his quality of life.

Observations

Surgical Approach

Historically, thoracolumbar vertebrectomies are performed via an anterior approach; however, posterior vertebrectomies are increasingly being performed. An anterior approach offers great visualization but carries risks of iatrogenic injury within the abdominal cavity and often requires an access surgeon.⁵ Additionally, an anterior approach may not be tolerated in patients with substantial comorbidities, such as a decline in respiratory function and morbid obesity.^6,7 Posterior approaches, such as transpedicular and costotransversectomy, do offer the potential for circumferential decompression of the spinal cord or lumbar thecal sac but carry the disadvantage of requiring extensive muscle dissection and bony decompression with associated higher blood loss.⁵ Although increased morbidity has been associated with combined anterior-posterior vertebrectomies, significant differences in outcomes and complications have not been observed for thoracolumbar anterior versus posterior vertebrectomies.^5,8

The literature regarding vertebrectomy approaches in cancer largely focuses on thoracic and cervical metastases; however, key differences in anatomy must be considered before the data can be translated to lumbar metastases. Lumbar vertebrae are significantly larger and must bear greater load.^2,9 Additionally, lumbar nerve roots innervate musculature vital for ambulation and cannot be sacrificed to increase accessibility.^2,9,10

In this case, a posterior-only approach for vertebrectomy was selected despite a largely anteriorly located tumor because of the patient’s high risk for an anterior or combined approach. Additionally, obesity has been associated with increased access times and postoperative complications.^6,7 With a more favorable prognosis and greater tolerance for surgery, a curative approach, such as an en bloc spondylectomy, could be considered in a patient with a solitary spinal metastasis.^11–13

Several surgical decisions allowed a successful operation. To increase working space for the cage placement, a lateral L4 pediculectomy was performed. During placement of the expandable cage, retroperitoneal structures in the lower abdomen were at risk, namely the abdominal aorta and its associated branches. The innovative use of navigation to simulate the length of the cage allowed the cage to be placed with precision with respect to the aorta anteriorly. Additionally, the innovative use of the silk tie threaded through the cage’s internal aperture allowed noninstrumented manipulation of the cage.

Spinal Metastases and Prognosis

Spinal metastases are seen on autopsy in up to 70% of patients with cancer, but only 14%–20% of patients will be symptomatic.^1,11,14 Unrelenting, progressive pain is the most common presenting symptom.¹⁴ Metastases are most common in the thoracic spine, followed by the lumbar and cervical spine, with most tumors being extradural.¹⁵ Breast, lung, and prostate primaries are most common.¹⁶ Prognosis is multifactorial; nonbreast primary cancer, the involvement of 2 or more vertebrae, 3 or more comorbidities, and a nonambulatory status are poor prognostic factors.^2,17 In patients with lumbar metastases treated surgically, the 6-month mortality rate was 25%.² Additionally, the presence of an unknown primary was associated with a mean survival of 5 months.¹⁴

The prognosis associated with spinal metastases can vary widely. Thus, deciding to operate versus using nonsurgical interventions, such as radiotherapy or chemotherapy, can be challenging. Indications for surgery include^4,15 intractable pain, a tumor resistant to chemotherapy or radiation, progressive neurological deficit, spinal instability, the need for diagnostic tissue, and tolerance to previous radiotherapy.

Tomita et al.⁴ and Tokuhashi et al.³ proposed systems to determine a prognosis score and treatment goal and recommended a surgical strategy based on the primary cancer, visceral metastases, and bone metastases. Meta-analysis has suggested prognosis predictions by Tokuhashi et al. and Tomita et al. may be of weak diagnostic value; thus, any scores must be considered in the context of individual patients.¹⁸ The nomogram put forth by the Skeletal Oncology Research Group may more accurately predict 3- and 12-month survival.¹⁹ Neither scoring system accounts for a patient’s level of biological and mechanical pain or spinal instability due to the neoplasm. Additionally, the scores’ impacts are limited by the lack of guidance for radiological therapy options and the flexibility to incorporate advances in therapy.

Advances in treatment have necessitated a more comprehensive decision-making framework. The NOMS (neurological, oncological, mechanical stability, and systemic disease) framework considers cord and nerve root compression, radiosensitivity, spinal instability, and patient tolerance to treatment.^19,20 Following the NOMS framework, a practitioner is guided to management with conventional external beam radiation therapy, stereotactic radiosurgery, separation surgery before radiation, and/or surgical stabilization.^19,20 Systemic variables, such as body mass index, ambulatory status, and chronic diseases, are also recommended to consider as part of the NOMS decision framework.²¹

The degree of spinal instability can be reliably classified by the Spinal Instability Neoplastic Score (SINS) with high sensitivity and specificity.²² As a component of the SINS, mechanical pain has shown significant correlations with patient-reported pain and physical function.^23,24 Patients with a high SINS, and thus greater instability, may be more likely to report benefits from instrumented surgical intervention.²³ Limited investigations have shown the SINS to be inconsistently correlated to postoperative survival and ambulatory function.^25–27

Surgical intervention for spinal metastases has been associated with improved ambulatory and neurological functions and reduced pain.^9,17 Of note, surgery is associated with increased complications and readmissions.² Due to surgical risks, a postoperative prognosis of at least 6 months is often a benchmark for surgery. However, arguments can be made for surgery for shorter prognoses. Fehlings et al.²⁸ found surgery with radiation and chemotherapy to provide immediate and sustained improvements in pain, neurological, functional, and quality of life outcomes in patients with a focal lesion and a prognosis of at least 3 months. On the basis of the case and evidence outlined previously, the decision to provide palliative surgery should include prognosis in the context of the patient’s pain, responsiveness to radiation, overall health, spinal instability, and the potential benefits of surgery.

Lessons

This case highlights a challenging surgical approach in a patient with a complex medical history. We show that a posterior-only approach can be used for lumbar vertebrectomies and fusion when necessary. The surgery was intended to be palliative and was accompanied by a significant morbidity risk. On the basis of prognosis, this surgery may be considered controversial. However, the palliative surgery did substantially reduce the patient’s pain and enhance his quality of life. Future similar cases should include substantial multidisciplinary discussions of the risks and realistic expectations with the patient.

Author Contributions

Conception and design: Arnold, Shaffer, Tang, Tsai, Guglielmi. Acquisition of data: Arnold, Shaffer, Guglielmi. Analysis and interpretation of data: Arnold, Shaffer, Guglielmi. Drafting the article: Johnson, Shaffer, Tang, Tsai. Critically revising the article: all authors. Reviewed submitted version of manuscript: all authors. Approved the final version of the manuscript on behalf of all authors: Arnold. Administrative/technical/material support: Arnold, Tang, Tsai, Guglielmi. Study supervision: Arnold, Guglielmi.