Background Extensive lytic lesions of the vertebral body (VB) increase risk of fracture and instability and require stabilization of the anterior column. Vertebral augmentation is an accepted treatment option, but when osteolysis has extensively destroyed the VB cortical boundaries (a condition herein defined as ‘extreme osteolysis’), the risk of cement leakage and/or insufficient filling is high. Vertebral body stents (VBSs) might allow partial restoration of VB height, cement containment, and reinforcement, but their use in extreme osteolysis has not been investigated.
Objective To assess retrospectively the feasibility and safety of VBS augmentation in patients with ‘extreme osteolysis’ of the VB.
Methods We retrospectively analyzed 41 treated vertebrae (from T1 to L5). VB reconstruction was assessed on postprocedure CT images and rated on a qualitative 4-point scale (poor-fair-good-excellent). Clinical and radiological follow-up was performed at 1 month and thereafter at intervals in accordance with oncological protocols.
Results VBS augmentation was performed at 12 lumbar and 29 thoracic levels, with bilateral VBS in 23/41. VB reconstruction was judged satisfactory (good or excellent) in 37/41 (90%) of levels. Bilateral VBS received higher scores than unilateral (p=0.057, Pearson’s X2). We observed no periprocedural complications. Cement leaks (epidural or foraminal) occurred at 5/41 levels (12.2%) without clinical consequences. Follow-up data were available for 27/29 patients, extending beyond 6 months for 20 patients (7–28 months, mean 15.3 months). VBS implant stability was observed in 40/41 cases (97.5%).
Conclusions Our results support the use of VBS as a minimally invasive, safe and effective option for reconstructing the anterior column in prominent VB osteolysis.
Statistics from Altmetric.com
Spinal osteolysis may cause instability, leading to fractures and neural compression.1 Stability restoration is therefore of paramount importance in the treatment of spinal lytic tumors.
Radiation and systemic therapies are used to achieve tumor control and pain palliation, but invasive treatments are often required to prevent or arrest vertebral collapse in patients with lesions affecting the weight-bearing portions of the vertebrae, including the vertebral body (VB).
Posterior surgical fixation is widely used in such cases, but should be accompanied by anterior column stabilization, either with corpectomy and grafting, or with cement vertebral augmentation (VA).2 Posterior fixation, however, may not be feasible in patients with advanced disease, multilevel lesions, and poor bone quality. Corpectomy and grafting is an effective treatment but is an invasive procedure that has significant morbidity risk, especially in fragile patients.3 4
Standalone VA is considered a viable option to achieve pain palliation and reinforce the anterior column,5–7 but when the osteolysis causes extensive destruction of the cortical boundaries of the VB, a condition here defined as ‘extreme osteolysis’ (EO), the injection of cement may be challenging or impossible.8 9
The vertebral body stent (VBS; DePuySynthes-Johnson&Johnson) is a balloon-expandable, barrel-shaped, metallic device, which is inserted via monopedicular or bipedicular access. On expansion, the VBS keeps the created cavity open after balloon deflation until cement is injected. The VBS was introduced for treatment of vertebral compression fractures but its use in extensive neoplastic osteolysis has not been investigated and recommended.10–15 Nevertheless, in EO of the VB, the VBS might serve as a stabilizing implant owing to its large support surface, its potential to achieve VB height restoration and its help in containing the injection of cement (online supplementary figure 1S). The purpose of this study was to assess the feasibility and safety of VB reconstruction using VBS in EO. The durability of the results was assessed and the rate of subsequent spine surgery at target levels recorded.
Supplementary file 1
Materials and methods
This is a retrospective study on a prospectively maintained database of a consecutive series of patients with neoplastic EO of one or more VBs, treated with VBS (March 2013 to November 2016). EO was defined as an extensive lytic lesion of the VB, with wide cortical destruction, combining, to variable extent, involvement of the posterior wall, the anterolateral boundaries, and the disc endplates (types 4–6 according to the scoring system of Tomita et al).16 The study was approved by the local ethics committee. Patients were informed of the investigative use of VBS to treat their condition and provided informed consent.
All patients underwent preprocedural spinal CT and gadolinium-enhanced MRI at the target level to define the extent of osteolysis, degree of vertebral collapse (<50% or >50%), and presence of epidural mass (EM) (see online supplementary figure 2S).
All target lesions were deemed to be unstable or potentially unstable according to their Spinal Instability Neoplastic Score (SINS).17
Therapeutic decisions for each patient were reached by a multidisciplinary spine-tumor board, defining indications and timing of invasive, radiation, and chemotherapy treatments.
All procedures were performed under biplane fluoroscopic guidance. The VBS was implanted through a transpedicular 7G trocar; a unilateral stent was inserted when a lateralized lytic lesion or a small VB was present.
When deemed necessary, before VBS deployment, a cavity was created in the VB using a coaxial osteotomic curette (Medtronic, Minneapolis, Minnesota, USA), followed by vacuum suction.
VBSs were expanded and implanted in the VB by hydraulic balloon inflation. After balloon deflation and removal, high-viscosity polymethylmethacrylate (PMMA; Vertaplex HV, Stryker, Kalamazoo, Michigan, USA) was injected under fluoroscopic monitoring, to obtain VBS filling and, when possible, interdigitation into adjacent trabecular bone. In multilevel osteolysis, when indicated, further vertebrae were subjected to VA during the same procedure, at adjacent or distant levels. Patients were allowed to stand and walk as early as 3 hours after the procedure, and most commonly discharged on the same day.
Assessment of VB reconstruction and follow-up
VB reconstruction was assessed on postprocedure plain films and CT scans. CT datasets were reconstructed with a bone algorithm with 3 mm and 10 mm thick maximum intensity projection images in three orthogonal planes, and reviewed by a neuroradiologist (AC) and a neurosurgeon (PS). Extravertebral cement leaks were recorded. Based on the restoration of VB height, the position of the stents in the VB, and the cement filling of the lytic cavities and adjacent trabecular spaces, an overall score of VB reconstruction success was assigned, under consensus, on the basis of an a priori-defined, qualitative, four-grade scale, rating the VB reconstruction as poor, fair, good, or excellent. Poor indicated failure to achieve sufficient augmentation of the anterior column, whereas excellent indicated appropriate stent expansion to fill the lytic lesion and reconstruct the destroyed portion of the VB, satisfactory height restoration, and cement filling (figure 1). An excellent result would appear as an internal prostheses of the affected VB. Good and excellent ratings were considered satisfactory results.
Patients were followed up at 1 month clinically and with upright spine plain films. Thereafter the patients underwent routine oncological clinical and imaging follow-up. From these records, spine images could be derived and assessed at intervals in accordance with oncological follow-up protocols, or when clinical conditions prompted referral for spinal imaging. Imaging follow-up was evaluated to assess significant new findings at the treated and adjacent levels.
Descriptive statistics for clinical and demographic data were expressed as mean, or as median±range.
Implant outcomes were stratified according to the VB reconstruction score, into the following categories: poor, fair, good, or excellent. Differences between categories in the degree of height reduction and tumor histotype (metastases vs multiple myeloma-plasmacytoma) were tested using Pearson’s χ2 . The same test was employed to assess differences in the VB reconstruction scores and cement leak occurrence when treated levels were classified on the basis of bilateral or unilateral VBS. The existence of a relationship between patients’ SINS and VB reconstruction scores was investigated using Spearman’s rs test. A p value of <0.05 was considered statistically significant. Analyses were conducted using SPSS version 20.0.0 (IBM, Armonk, New York, USA).
The study group included 29 patients and procedures to treat 41 levels with EO between T1 and L5. SINS ranged between 7 and 18 (mean 10.7; median 10).
A summary of patients’ characteristics and features of the lytic lesions is provided in table 1.
One patient had a neurological deficit due to spinal cord compression, and another patient had radicular sciatic pain before the procedure.
Conscious sedation was used in 17/29 patients, and general anesthesia in 12/29. VBS procedures were performed as a standalone intervention in 26/29 cases (36 levels), with a percutaneous posterior surgical fixation in 1/29 cases (one level), and after laminectomy and posterior surgical fixation in 2/29 cases (four levels). VBS was bilateral at 23/41 levels, and unilateral at 18/41 levels. Cavity creation was performed at 35/41 levels.
During the same procedure, additional VA with cement-only was performed at adjacent or distant vertebral levels (affected by lytic lesions, but not defined as EO) in 20/29 cases, at a total of 63 levels.
Cement leakage was detected in 14/41 cases (34%), without clinical consequences. No other clinical intraprocedural complications were seen. No patients showed new or worsening neurological deficit.
VB reconstruction by VBS was judged excellent at 31/41 (75.6%), good at 6/41 (14.6%), fair at 4/41 (9.8%), and poor at 0/41 of the treated levels, leading to a satisfactory result (excellent or good rating) in 37/41 (90%) of cases. Online supplementary Table 1S summarizes the technical results.
VB reconstruction scores did not correlate with degree of height reduction, SINS, or tumor histotype. The occurrence of cement leaks did not correlate with unilateral or bilateral implants. Only the difference in the VB reconstruction score between bilateral and unilateral implants approached statistical significance (p=0.057, Pearson’s X2 as shown in online supplementary Table 2S.
Clinical and imaging follow-up at 1 month after the procedure was available for 27/29 patients (39/41 treated levels). Owing to deaths from unrelated causes, follow-up data at ≥6 months (7–28 months, mean 15.3 months) were available for 20 patients (28/41 levels).
Spine stability at the target-levels was observed until the last available follow-up in 40/41 cases (97%). In one patient, ventral mobilization of the VBS implants at T1 was noted at 1-month follow-up, causing transient dysphagia. Only one patient in this series required subsequent spinal surgery at the target level, 6 months after VBS, owing to an adjacent-level fracture.
Four patients showed mild adjacent-level impaction fracture, without clinical consequences. One patient developed radicular pain 3 months postprocedure, with a new disc herniation adjacent to a target level, and was treated conservatively.
Local progression of disease was observed in two patients (2/41 levels, 4.9%), at 3 and 23 months postprocedure, respectively, with increased EM, but without neurological sequelae.
In our series of patients with EO in whom VBS was used to reconstruct and augment the VB, technical success was achieved in 90% of cases, regardless of tumor histology, with no significant clinical complications, and with stable results at follow-up. Subsequent target-level spine surgery occurred in 1/29 patients.
All patients had a SINS level indicating unstable or potentially unstable vertebral lesions. The main aim of the procedure was reconstruction and augmentation of the anterior column in VBs that had fractured or were at risk of collapse.
In patients with EO of the VB, surgery is considered the standard treatment to restore stability. Posterior fixation should be combined with corpectomy and grafting, with placement of different cages, cement or autologous bone.3 This approach, however, is associated with significant morbidity and is mostly indicated in patients with solitary spinal metastasis, in good general health, and with a long life expectancy.18 Moreover, surgical fixation might not be the ideal solution in patients with multilevel metastatic involvement or poor bone quality.
Augmentation procedures, such as vertebroplasty and kyphoplasty, either as standalone procedures,19 following radiofrequency ablation,20 or in combination with posterior fixation,21 might be contraindicated or unfeasible in the presence of EO. Extensive loss of the integrity of cortical boundaries may favor extravertebral cement leakage, potentially resulting in compression of neural structures and/or insufficient filling of the vertebral lytic lesion. This is likely to lead to unsatisfactory augmentation and reinforcement of the anterior column.8 22 23 A VBS, introduced for the treatment of vertebral compression fractures,10–15 has not been investigated or recommended in patients with EO, and cranial to T6.
In this study, VBS was chosen as a standalone procedure when deemed clinically appropriate and when surgery was contraindicated, or in combination with a posterior surgical approach, as an alternative to a surgical reconstruction of the anterior column with a cage. In our series a combined VBS and posterior surgical fixation was chosen in three patients presenting spinal lesions with high SINS values(13–18). In patients with multilevel involvement, the decision about which levels to treat to prevent or to arrest a fracture was based on the extent and location of the lytic lesions suggesting biomechanical risk of collapse.24 Owing to its minimal invasiveness, and minimal recovery time, VBS could be more deliberately offered as a palliative treatment, for patients having a poor prognosis and/or a low Tokuhashi score.25
When the lytic lesion was felt to have a solid soft-tissue consistency during insertion of the trocar, we performed curettage with a coaxial osteotome, followed by vacuum suction through an 8G cannula (online supplementary figure 2S). In our opinion, the creation of a cavity in the VB reduces the risks of displacement of solid tumorous tissue through the dehiscent cortical boundaries—namely the posterior wall, and of epidural PMMA leak.
In a unilateral VBS implant, cement-only VA was performed through the contralateral pedicle, to ensure bilateral and homogeneous augmentation.
The injected volume of PMMA varied according to the size of the lytic lesions, trabecular compliance, stent expansion, and distribution of injected cement.
Efficacy of the procedure
Vertebral body reconstruction
The main goal of the VBS procedure and the primary endpoint of our analysis was reconstruction of the VB, which is important for restoration of axial load-bearing capability of the anterior column.12 This was attempted by creating a construct similar to ‘armed-concrete’ in the VB with metallic stents and PMMA. With their large support-surface, filled with PMMA, VBS could provide primary reinforcement of the anterior column, and their tight mesh may help to achieve cement containment (figure 2 and online supplementary figure 3S). Where necessary and possible, we tried to restore the VB height, to favor a more physiological biomechanical condition.
In the absence of a validated system, a neuroradiologist and a neurosurgeon assigned, under consensus, an arbitrary VB reconstruction score. This was based on a qualitative overall assessment of the postprocedure plain films and CT images, taking into account stent placement, expansion, cement filling, and VB height restoration. The reconstruction achieved was judged satisfactory in 90% of levels. The VB height as seen on postprocedure CT scans was substantially maintained at follow-up in 40/41 levels. More subtle phenomena of bone remodeling/subsidence around the solid implants of VBS and PMMA were observed, but had no significant effect on stability, spinal alignment, or patients’ symptoms. Reconstruction at four treated levels was judged unsatisfactory (fair), based on assessment of postprocedure CT images; all four levels had received a unilateral VBS implant, and 3/4 were lumbar levels. Of these patients, only one required subsequent posterior surgical stabilization, one died before the 6-month follow-up but had no instability complications, and the remaining two had an unremarkable follow-up. Although low scores for VB reconstruction seemed not to correlate with a poor outcome, as indicated by stability at follow-up, this might be because the number of patients was small. Bilateral implant, when possible, seems to offer more satisfactory VB reconstruction results, especially in the larger lumbar vertebrae. Successful VB reconstruction might relate to the restoration of the load-bearing capability of the anterior column, as suggested by the observed stability of the target levels at follow-up. Moreover, in the available late follow-up CT examination, formation of a new cortical bone shell was noted around the VBS, and in some cases intervertebral osseous fusion was observed (figure 3), implying achievement of stability.
Other studies have assessed the use of VBS to restore VB height and alignment in osteoporotic fractures. In contrast to balloon kyphoplasty, the VBS system maintains the restored VB height because the stent remains expanded in the VB after balloon deflation.12 In our opinion VBS has other potential advantages in patients with EO. The metallic mesh guarantees a reasonably uniform and predictable barrel-shaped balloon expansion, while a ‘non-armed’ compliant balloon might expand following the path of least resistance in a severely altered anatomy, where residual bone, sclerosis, and lytic soft-tissue lesions coexist. The barrel shape of the VBS, with its large support surface, provides mechanical support and recreates VB walls, such as in lateral or endplate dehiscent cortical boundaries. Furthermore, the metallic mesh helps to contain the viscous PMMA cement (see Figure 1 and Figure 2 online supplementary figure 3S). When bilateral VBSs are fully deployed in the VB they resemble a solid and efficient VB prosthesis, offering a scaffold for the subsequent PMMA filling, and respecting the intervertebral disc spaces, which are usually untouched by the neoplastic lesions. In the follow-up period we observed five patients with a new compression fracture adjacent to the VBS-treated target level. Only one required treatment with VA and surgical stabilization due to worsening focal scoliosis; the others were asymptomatic and did not require interventions. Adjacent new fractures are a known phenomenon, and might be attributable to multilevel metastatic involvement, or primary or secondary osteoporosis. There might also be a concomitant effect of altered biomechanics, caused primarily by the target-level fracture and, to a lesser degree by the fixation of the fracture, which is likely to increase the stiffness of the treated level.26–29
In several cases we performed VA with cement-only at adjacent or distant levels, either owing to multilevel osteolysis, or for prophylaxis (Figure 3).
Pain palliation was a concurrent indication for treatment in most of our patients, but did not represent an endpoint of this study. Precise assessment of pain palliation is usually difficult while evaluating different treatment modalities, and is even more challenging in a retrospective study. Moreover, many patients had multilevel spinal and extraspinal metastatic involvement, and therefore might have had multifactorial causes of their pain23 . Also, some patients might not have had any significant pain from the target vertebral lesion and might have undergone treatment solely for stabilization purposes, to avoid or arrest fracture. Finally, patients also received different treatment regimens of radiotherapy, chemotherapy, supportive care (including steroids and analgesic drugs), and concomitant VA at adjacent or distant levels. Nevertheless, clinical charts recording oncology follow-up reported amelioration of clinically significant pain solely attributed to the VBS–VA procedure in 17/29 patients (with no new or changed therapeutic regimen). In 4/29 cases, amelioration of pain was attributable to the procedure in combination with a new or changed regimen of chemotherapy or radiotherapy (it was impossible to determine whether one of these therapeutic measures or a combination was responsible for the pain palliation), while in 4/29 patients no significant pain amelioration was noted after the procedure. The remaining 4/29 patients had no definite preprocedure pain clearly attributable to the vertebral target lesion.
The pain palliation offered by a VA procedure, in this case performed with VBS, might have some advantages over standard radiation therapy alone. Beside dose constraints in patients who have already been irradiated, beneficial effects of radiotherapy on pain might be delayed by weeks or months,30 and about 20–30% of patients are non-responders.31 32 In addition, radiation cannot ensure immediate stabilization of the affected vertebra, since treatment might be followed by a phase of increased vertebral fragility and fracture risk.33 Nevertheless, we are not proposing an exclusive role for VBS but rather a complementary role to multimodal treatment with radiation and/or chemotherapy to achieve rapid pain relief, immediate reinforcement of the anterior column, and local disease control. Indeed, the multimodal approach might have been the key to the low rate (4.9%) of local disease progression at the treated levels, observed in this series.
No clinically significant intraprocedural or periprocedural complications were observed. Postprocedure CT showed PMMA leaks in 34% of cases (epidural or foraminal leak in 12.2% levels), all without clinical consequences.
Previous studies reported a lower rate of extravertebral PMMA leakage for VBS than for vertebroplasty10 11 17 but the patients mostly did not have neoplastic lesions, which carry higher leakage risk. As far as we know, there are no data about the rate of PMMA leakage in patients with EO of the VB treated with VBS. A previous study34 reported a low complication rate (1.4%) during VA of neoplastic lytic lesions with a dehiscent posterior wall, and our results seem to support the use of VBS as a safe procedure even in patients with EO.
In our series, 21/41 levels had an EM, visible on preprocedure MRI, and we observed no worsening of neurological status after the procedure. We believe that cavity creation in the VB, and fracture reduction, with re-expansion of the collapsed VB in a craniocaudal direction, might have had a role in preventing clinically significant soft-tissue migration in the central canal.
At 1-month follow-up, one patient exhibited ventral mobilization of the two stents implanted at the T1 level, causing mild dysphagia. The posterior half of the stents remained between C7 and T2 vertebral bodies, still ensuring support, and flexion–extension plain films showed no mobility of the stents, and maintenance of spinal alignment. A conservative approach was chosen. At 3-months' follow-up the dysphagia had resolved and there was no further mobilization of the stents, which also remained stable at the 6- and 8-month follow-ups. It is certainly conceivable that the stents might mobilize in the absence of an intact VB cortical shell, but our results so far show this to be an unusual occurrence. Nevertheless, to obviate this potential problem, a system to anchor the VBS to the pedicle(s) is under evaluation at our center.
The limitations of our study relate mainly to its retrospective design, rather arbitrary inclusion criteria, and heterogeneous follow-up. It also lacked a control group, either of patients treated conservatively or with a standard surgical technique. However, the patients included in this study had all had a fracture or were at significant risk of developing a vertebral fracture, and were deemed unsuitable for standard surgical intervention—namely, anterior stabilization surgery. The scale adopted to assess the efficacy of VB reconstruction in the absence of alternatives in the literature was qualitative, arbitrary, and not validated, but the rating reflected the consensus opinion of a neuroradiologist and a neurosurgeon. Nevertheless, this study is the result of a prospectively established database; all treated cases were included and management and follow-up reflected a multidisciplinary established clinical practice. This study is a first step, and a larger study might provide more robust information on which to base vertebral augmentation procedures in such challenging cases.
Extreme VB neoplastic osteolysis poses a treatment challenge in a group of fragile patients. In our study, the use of percutaneous VBS proved to be a minimally invasive, feasible, safe, and effective technique to augment and, ultimately, provide stability to the anterior spinal column, with durable results. It might therefore be considered as a valuable option, as a standalone intervention or in combination with a posterior surgical approach of decompression and stabilization.
Contributors All authors contributed to the presented work by substantial contributions to the conception or design of the work; or the acquisition, analysis, or interpretation of data for the work and drafting the work or revising it critically for important intellectual content and final approval of the version to be published and agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Patient consent Not required.
Ethics approval Ethics committee of Canton Ticino.
Provenance and peer review Not commissioned; externally peer reviewed.
Correction notice Since this paper was first published figures 2 and 3 have been switched. The legend of figure 3 has also been updated.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.