Recently, there has been rising interest in the use of porous tantalum for spinal fusion because of its high porosity, high compressive strength, and elasticity resembling that of cancellous bone. All of these factors result in rapid osteointegration.1–3 Although porous tantalum trabecular metal has been successfully used as an anterior cervical and lumbar interbody fusion device,4–6 it is still associated with a certain rate of pseudoarthrosis and delayed union.1,7
To obtain optimal clinical outcome, it is essential to achieve fusion.8,9 The imaging techniques currently employed to assess spinal fusion include plain radiographs and computed tomography (CT) scans, including thin-section CT and multiplanar reconstruction CT. However, assessing bony fusion in the spine can be difficult, particularly when the material under investigation (i.e., porous tantalum) does not permit visualization of the formation of bridging trabeculae due to its dense radiopaque nature. To our knowledge, there is no consensus on the optimal technique for evaluating spinal fusion using porous tantalum implants. The radiographic technique is helpful only to the extent of determining graft location and the absence of extrusion; however, it is impossible to determine bony ingrowth or fusion with radiographs, due to the high density of tantalum. Because of the amount of artifact, CT scan is not the imaging modality of choice for tantalum spinal implants compared with titanium implants.10
Magnetic resonance imaging (MRI), which can aid in the identification of tiny lesions undetected by radiographs, has been increasingly used for assessment of spinal fusion.11–14 Although MRI is appealing, its utility in assessing spinal fusion remains unproven,15 especially when porous tantalum implants have been employed. The purpose of this study was to determine the performance of MRI in assessing spinal fusion in a porcine model, in which lumbar interbody fusion was carried out using porous tantalum implants.
MATERIALS AND METHODS
Animal Model and Study Design
A total of twelve 12-week-old female Danish Landrace pigs from different litters, weighing about 50 kg, were chosen for this investigation. Each animal underwent a discontinuous 3-level anterior intervertebral lumbar arthrodesis at L2/3, L4/5, and L6/7 using porous tantalum trabecular metal interbody fusion devices (23 mm in width × 15 mm in depth × 11 mm in height; Trabecular Metal™, Zimmer) under general anesthesia. To facilitate the evaluation and comparison in the performance of different imaging techniques assessing spinal fusions, we tried to induce different fusion status at different levels by using 2 different fusion techniques and 2 different cages (PT-ring and PT-cage) on the basis of our previous studies.16,17 A detailed description of surgical procedures had been reported in our previous study.16 Briefly, bone graft was harvested from the right iliac crest with the pig placed in a prone position. The L2–3 intervertebral space was identified under fluoroscopic control before surgical intervention. The facet joints of the neighboring vertebrae at this level were exposed through a posterior midline incision and paraspinal intermuscular approach. Pedicle screws (5 mm in diameter × 30 mm in length, Diapason; Stryker® Spine, Cestas, France) were inserted into the vertebrae of L2 and L3 for fixation, followed by careful suturing of the incision in the back.
The pig was subsequently turned over in a supine position. A paramedian abdominal incision was made and a retroperitoneal anterior approach was carried out. After preparation of L2–3 disc space, a solid porous tantalum cage (PT-cage) was implanted. The L4–5 and L6–7 disc levels were similarly prepared, and then hollow porous tantalum rings (PT-rings) packed with iliac crest autograft were implanted, respectively. Both of the 2 levels were then secured anteriorly with 2 staples (22 mm in length × 16 mm in depth; Howmedica GmbH, Schönkirchen, Germany). Prophylactic ampicillin (1.0 g intravenously, Anhypen; Gist-Brocades, Delft, the Netherlands) was given before and immediately after surgery, and twice a day during the following 3 days.
All pigs were kept in individual pens and were fed with a normal diet containing 1.4% calcium and 0.7% phosphorus (percent food weight). The pigs were observed for 6 months before sacrifice. All animal surgeries and experiments complied with the Danish Law on Animal Experimentation and were approved by the Danish Ministry of Justice Ethical Committee (J.nr. 2004–561–898).
MRI was carried out on all the pigs under general anesthesia with a 1.5-T MR system (General Electric Medical Systems, Milwaukee, WI) using a spine array coil (5 × 11 in.). T1-weighted spin-echo, T2-weighted spin-echo, Gd-DTPA-enhanced T1-weighted spin-echo, and Gd-DTPA-enhanced T2-weighted spin-echo magnetic resonance (MR) images were then obtained from the lumbar spine in sagittal and coronal planes. The acquisition parameters were similar to those used in clinical practice as follows: (1) Axial localizer (spoiled gradient); (2) Sagittal T1 (echo time [TE] minimum full/repetition time [TR] 400) (thickness 4 mm/spacing 0.3 mm; flip angle 90°; image matrix 512 × 256; field of view [FOV] 30 cm2); (3) Sagittal T2 (TE/TR = 100/6675 ms) (thickness 4 mm/spacing 0.3 mm; flip angle 90°; matrix 512 × 256; FOV 30 cm2); (4) Coronal T1 (TE minimum full/TR 500) (thickness 4 mm/spacing 0.3 mm; flip angle 90°; matrix 512 × 256; FOV 24 cm2); (5) Coronal T2 (TE 100/TR 6675) (thickness 4 mm/spacing 0.3 mm; flip angle 90°; matrix 512 × 256; FOV 24 cm2).
All the MRI analysis was carried out by 2 experienced radiologists who were blinded to the purpose of our study. All the analyses were reviewed 6 months after the first examination to determine the intraobserver reliability. Region of interests (ROIs) at the fusion segments of L2–3, L4–5, and L6–7 were listed as follows: (1) the cranial vertebral bone, (2) the cranial interface between vertebrae and implant, (3) the anterior part of porous tantalum implant, (4) the central part of porous tantalum implant, (5) the posterior part of implant, (6) the caudal interface between vertebrae and implant, and (7) the caudal vertebral bone. An intermediate signal intensity (SI) band between the vertebral body and implant on the basis of MR imaging was defined as osseous healing at the vertebrae-implant interface, whereas high SI band was considered to be fibrous tissue layer at the vertebrae-implant interface and represented nonunion. ImageJ (National Institute of Mental Health, Bethesda, MD) was used for image analysis and processing.
After sacrifice, the spinal column from L1 to L7 with the sacrum of each pig was removed en bloc, stripped of soft tissue and stored at −200°C until examination. Plain radiographs in anterior–posterior and lateral views were obtained. Assessment of all radiographs was carried out by the same 2 experienced radiologists who were blinded to the purpose of our study. After all the analyses were collected, the radiologists together delineated the inconsistent results. The presence or absence of radiolucent lines at the vertebrae-implant interface were recorded.
After radiographical examination, the implants at L2–L3, L4–L5, and L6–L7 were harvested together with the neighboring vertebral body. Each specimen was cut in half longitudinally in the sagittal plane with a water-cooled diamond saw (Exact Apparatebau, Nordenstedt, Germany). These were then dehydrated in graded ethanol (70%–99%) containing 0.4% basic fuchsin and embedded in methylmethacrylate for histomorphometric analyses. The sections were cut to a thickness of 40 μm with 500-μm intervals between each section to obtain the maximal range of sampling, using the sawing microtome KGD 95 (Meprotech, Galileistraat 24, NL-17045E, Heerhugowaard, The Netherlands). The surface was counterstained with 2% light green for 2 minutes. Four sections were produced from each vertebrae-implant specimen. For qualitative histological assessment, all sections were blindly evaluated under light microscopy. Histological sections were used to define the status of tissue in contact with the device at the cranial and caudal vertebrae-implant interfaces as being bony or fibrous tissue surrounding the implant.
As receiver operating characteristic curve (ROC) is particularly useful when comparing 2 or more diagnostic tests, the ROC plot was obtained by calculating the sensitivity and specificity for each observed data value and plotting sensitivity against 1 minus specificity. The smallest cutoff value of sensitivity and specificity for fibrous tissue at the vertebrae-implant interface was the minimum observed test value − 1, and the largest cutoff value was the maximum observed test value +1. All the other cutoff values from the ROC curves were the averages of 2 consecutive observed test values. The area under the ROC curve (AUC) was calculated as an indicator of diagnostic power.
Statistical analyses were carried out using SPSS software (version 16.0, SPSS, Inc., Chicago, IL). All data of MR signal intensity were expressed as mean ± standard deviation (SD). To assess intraobserver and interobserver reliability, the Cohen κ value was calculated as described by Altman for interobserver reliability data and the intraclass correlation coefficient was calculated for intraobserver variation.18,19 The statistical analyses of MR signal intensity were carried out using multiple analyses of variance (MANOVA). When significant main effects were found, specific comparisons were made by paired t tests. The χ2 test was used to evaluate correlation between the variables. All statistical tests were 2-tailed with P < 0.05 considered statistically significant.
All the pigs survived surgery, and no intraoperative complications were observed. One pig had chronic infection at the site of pedicle screw fixation postoperatively, and was terminated at 1 month before scheduled sacrifice. Therefore, 11 pigs were finally included in the analysis of the results after the 6-month follow-up period. Radiographs revealed no obvious kyphosis nor migration of any of the interbody implants. According to the results of the histological sections, the solid porous tantalum cages (PT-cage) had 54.5% fusion rate (bone ingrowth into both side of the vertebrae-implant interface (6:5). The hollow porous tantalum rings (PT-rings) had 68.2% fusion rate (15: 7). In total, 21 levels were successfully fused and 12 levels went on to nonunion out of 33 levels.
On the basis of T1- and T2-weighted spin-echo images (Figures 1 and 2), the porous tantalum implant produced a minor metallic artifact that consisted of signal voids and geometric image distortions. The SI of T1-weighted spin-echo, T2-weighted spin-echo, Gd-DTPA-enhanced T1- weighted spin-echo (T1-Gd-DTPA), and Gd-DTPA-enhanced T2-weighted spin-echo (T2-Gd-DTPA) MR images at the vertebral bone and the bone-implant interface were measured. Results are shown in Figure 3. On the imaging of all the MR sequences, there was a significant higher SI band at the vertebrae-implant interface of nonfused segment, compared with that of fused segments, and that of the vertebral bone and implant themselves (P < 0.001, Figure 3). There were also significant differences in the SI of the T1-weighted spin-echo, T1-Gd-DTPA, and T2-Gd-DTPA MR images between the bone-implant interface of fused segments and vertebral bone (Figure 3).
The SI of MR imaging from all the MR sequences at the porous tantalum implant are demonstrated in Table 1. The SI at the central part of the hollow porous tantalum ring (PT-ring) was higher than that at the anterior or posterior parts (P < 0.05, Table 1), which predicted that the porous tantalum implant produced limited metallic artifact on the MR images, which could not affect the SI analysis. The intraobserver reliability of the SI of MR imaging was 0.90 with a 95% confidence interval (CI) ranging between 0.85 and 0.93. The interobserver reliability was 0.88.
The performance of T1-weighted, T2-weighted MR images and conventional radiograph films (Figures 1, 2, and 4) on the basis of histological findings at the vertebrae-implant interface are shown in Table 2. Receiver operating characteristic (ROC) plot (Figure 5) was drawn on the basis of the observed data set in Table 3. The estimated values of ROC curve parameters for T1-weighted, T2-weighted MR images, and conventional radiograph films with respect to the identification of fibrous tissue at the vertebrae-implant interfaces are shown in Table 3. The high SI band of T1- and T2-weighted MR images compared with histological findings resulted in 0.801 ± 0.063 AUC (95% CI: [0.677, 0.926]) and 0.759 ± 0.069 AUC (95% CI: [0.623, 0.895]), respectively. The radiolucent line of the conventional radiograph films at the vertebrae-implant interface compared with histological findings resulted in 0.749 ± 0.071 AUC (95% CI: [0.611, 0.888]) (Table 3). The κ value concerning the interobserver reliability was 0.81.
Because the model with a ROC curve that lies entirely above the ROC curve of another should perform better, there is a tendency that T1-weighted MR images are superior to both T2-weighted MR images and conventional radiograph films in detecting fibrous tissue at the vertebrae-implant interfaces. T1-weighted images have statistically larger AUC than both T2-weighted images (P < 0.001) and conventional radiograph films (P = 0.002); however, there is no significant difference between T2-weighted images and conventional radiograph films (P = 0.060).
Since the first posterior interbody fusion reported by Cloward in 1940s, spinal interbody fusion techniques have continued to evolve with an increasing number of interbody fusion devices available.15,20–23 These devices have varying geometric configurations and wall thicknesses, and they are made of various materials, such as titanium, tantalum, and polymers. Successful arthrodesis is important to have a clinically satisfactory outcome; however, generally accepted objective guidelines that can be used in clinical practice to evaluate spinal fusion are still not well documented.
Although dynamic radiographs have been widely used for spinal fusion assessment, they tend to significantly overestimate the presence of a solid fusion. In a study using sheep, Sandhu et al24 found that, although all sheep that accepted interbody fusion using cage showed evidence of fusion at 6 month-time point on plain radiograph films, only 33% were subsequently judged to be fused on histological analysis. In addition, a consensus concerning the degree of interbody motion that implies failure of fusion has not been reached.25 Hipp and Santos et al26,27 found that the apparent pseudoarthrosis rate was highly dependent on the threshold of motion used to define pseudoarthrosis. In fact, the degree of the interbody motion depends not only on the status of fusion but also on the elastic modulus of implantation, the height of interbody space, and rigidity of fixation. For these reasons, many spine surgeons assessing fusion now use CT scanning28,29; however, CT images are affected by artifact interference through and adjacent to the fusion device. This artifact impairs one's ability to interpret the CT scans. The severity of artifacts on CT imaging depends mostly on the innate material properties of implants. Porous tantalum, a new low modulus metal with a characteristic appearance similar to cancellous bone, is currently available to be applied in several orthopedic applications, such as arthroplasty, spine surgery, and bone graft substitute. This transition metal maintains several interesting biomaterial properties, including a high volumetric porosity (70%–80%), low modulus of elasticity, and high frictional characteristics.30 Unfortunately, tantalum creates more artifacts on CT scan than other metals, such as titanium.10 Furthermore, significant concerns remain with regard to the effects of high radiation exposure. Although there is no consensus with regard to an ideal radiologic assessment tool for evaluating interbody fusion,27 reducing or eliminating metallic artifact has been considered as a goal for both clinicians and radiologists, since the clinical application of MR imaging techniques. Some studies have shown that the tantalum seems to perform better on MR images than other metals including titanium.31,32
In this study, we carried out anterior lumbar interbody fusion in pigs using porous tantalum cages, and investigated the fusion status by using radiography, MR imaging and histology. The 12 weeks old Danish Landrace pig equals an adolescent in human beings. Although the exact age of Danish Landrace pigs is unknown in humans, the pig model was employed because of the availability and size suitability of the pigs for the cages. Human surgical techniques were used with no modification of the standard instruments. We carefully excised the cranial and caudal endplate of the fusion level with the guidance of a self-designed instrument, and fed the pigs with controlled diet to avoid rapid growth of the vertebra. In the previous study, the mean total body weight increased only 21.3% after 3 months. We demonstrated that the fibrosus tissue surrounding the implants, verified by histology, were consistent with the high SI band on T1-weighted MR images. On the basis of the “gold standard” of histology, we validated the diagnostic values of radiographs and MR images in detecting the fibrous tissue surrounding the tantalum interbody fusion devices according to ROC curves. The results showed that the sensitivity and specificity of both T1- and T2- weighted MR images were superior to that of radiographs, especially those of T1-weighted MR images. This indicated that T1-weighted MR images may be a primary alternative for monitoring the progression of interbody fusion carried out with tantalum implants.
Stradiotti et al33 demonstrated that the fast spin echo pulse sequence was the best MR sequence for artifact reduction. Taniyam et al34 verified that the cause of artifact in clinical cases is ring-shaped implants. Changing the direction of the signal magnetic field is effective in reducing or eliminating such artifact.35,36 Nevertheless, the most effective method in artifact reduction is substituting ferromagnetic implants with nonferromagnetic implants, such as tantalum which has been confirmed to have minor artifact on MR imaging. In this study, the data confirmed that tantalum has little artifact on MR, which was an additional advantage for the application of the porous tantalum metal.
There were also several limitations in this study. One of the limitations was the lack of data from CT scans. As previously mentioned, tantalum produces more artifact on CT scan than other metals, which makes it difficult to investigate the progression of spinal fusion carried out using tantalum metal devices.10,11 Although the use of CT for monitoring the progression of interbody fusion is now common, we did not include CT scan into our protocol. Another limitation was that we did not compare the specificity and sensitivity of different combinations of imaging tests. Although T1-weighted MR imaging has shown better performance in the observation of the spinal fusion in this study, further study is warranted to establish a more precise, ideal technique to evaluate fusion status.
This study suggests that MR imaging seems to be both an effective and noninvasive way of assessing the progression of spinal interbody fusion carried out with porous tantalum metal devices. T1-weighted spin-echo MR images are more sensitive and specific in detecting lucency between the vertebral body and porous tantalum interbody devices, compared withT2-weighted spin-echo MR images and conventional radiograph films.
The authors would like to thank Dr. Mohammed Khaleel and Dr. Aziz Hammoud (from Department of Orthopaedic Surgery, UT Southwestern Medical Center) for their contribution to this work. Zhiyu Zhou, Fuxin Wei, and Sheng Huang contributed equally to this work.
1. Bobyn JD, Poggie RA, Krygier JJ, et al. Clinical validation of a structural porous tantalum
biomaterial for adult reconstruction. J Bone Joint Surg Am 2004;86-A(Suppl 2):123–9.
2. Bobyn JD, Stackpool GJ, Hacking SA, et al. Characteristics of bone ingrowth and interface mechanics of a new porous tantalum
biomaterial. J Bone Joint Surg Br 1999;81:907–14.
3. Zardiackas LD, Parsell DE, Dillon LD, et al. Structure, metallurgy, and mechanical properties of a porous tantalum
foam. J Biomed Mater Res 2001;58:180–7.
4. Wigfield C, Robertson J, Gill S, et al. Clinical experience with porous tantalum
cervical interbody implants in a prospective randomized controlled trial. Br J Neurosurg 2003;17:418–25.
5. Lofgren H, Engquist M, Hoffmann P, et al. Clinical and radiological evaluation of trabecular metal and the Smith-Robinson technique in anterior cervical fusion for degenerative disease: a prospective, randomized, controlled study with 2-year follow-up. Eur Spine
6. Matejka J, Zeman J, Belatka J. Mid-term results of 360-degree lumbar spondylodesis with the use of a tantalum implant for disc replacement. Acta Chir Orthop Traumatol Cech 2009;76:388–93.
7. Kasliwal MK, Baskin DS, Traynelis VC. Failure of porous tantalum
cervical interbody fusion devices: two-year results from a prospective, randomized, multicenter clinical study. J Spinal Disord Tech 2013;26:239–45.
8. Moreland DB, Asch HL, Clabeaux DE, et al. Anterior cervical discectomy and fusion with implantable titanium cage: initial impressions, patient outcomes and comparison to fusion with allograft. Spine
9. Phillips FM, Carlson G, Emery S, et al. Anterior cervical pseudarthrosis. Spine
(Phila Pa 1976) 1997;22:1585–9.
10. Levi AD, Choi WG, Keller PJ, et al. The radiographic and imaging characteristics of porous tantalum
implants within the human cervical spine
(Phila Pa 1976) 1998;23:1245–50, 1251.
11. Wang JC, Yu WD, Sandhu HS, et al. A comparison of magnetic resonance
and computed tomographic image quality after the implantation of tantalum and titanium spinal instrumentation. Spine
(Phila Pa 1976) 1998;23:1684–8
12. Kroner AH, Eyb R, Lange A, et al. Magnetic resonance
imaging evaluation of posterior lumbar interbody fusion
(Phila Pa 1976) 2006;31:1365–71.
13. Uffen MP, Krijnen MR, Hoogendoorn RJ, et al. Tissue identification with micro-magnetic resonance
imaging in a caprine spinal fusion model. Eur Spine
14. Kwon DY, Kim BJ, Park KW. MRI artifact mimicking root compression by interbody cage displacement. Am J Phys Med Rehabil 2008;87:423–424.
15. Blumenthal SL, Ohnmeiss DD. Intervertebral cages for degenerative spinal diseases. Spine
16. Zou X, Li H, Teng X, et al. Pedicle screw fixation enhances anterior lumbar interbody fusion
with porous tantalum
cages: an experimental study in pigs
(Phila Pa 1976) 2005;30:E392–9.
17. Zou X, Li H, Bünger M, et al. Bone ingrowth characteristics of porous tantalum
and carbon fiber interbody devices: an experimental study in pigs
18. Altman DG. Practical Statistics for Medical Research. London: Chapman & Hall; 1991.
19. Barnhart HX, Song J, Haber MJ. Assessing intra, inter and total agreement with replicated readings. Stat Med 2005;24:1371–84.
20. Cloward RB New treatment for ruptured intervertebral disc. Paper presented at: Annual meeting of Hawaii Territorial Medical Association; May 1945.
21. Cloward RB The treatment of ruptured lumbar discs by intervertebral fusion. A report of 100 cases. Paper presented at: The meeting of the Harvey Cushing Society; November 15, 1947; Hot Springs, VA.
22. Zdeblick TA, Phillips FM. Interbody cage devices. Spine
(Phila Pa 1976) 2003;28:S2–7.
23. McAfee PC. Interbody fusion cages in reconstructive operations on the spine
. J Bone Joint Surg Am 1999;81:859–80.
24. Sandhu HS, Toth JM, Diwan AD, et al. Histologic evaluation of the efficacy of rhBMP-2 compared with autograft bone in sheep spinal anterior interbody fusion. Spine
(Phila Pa 1976) 2002;27:567–75.
25. McAfee PC, Boden SD, Brantigan JW, et al. Symposium: a critical discrepancy: a criteria of successful arthrodesis following interbody spinal fusions. Spine
(Phila Pa 1976) 2001;26:320–34.
26. Hipp JA, Reitman CA, Wharton N. Defining pseudoarthrosis in the cervical spine
with differing motion thresholds. Spine
(Phila Pa 1976) 2005;30:209–10.
27. Santos ER, Goss DG, Morcom RK, et al. Radiologic assessment of interbody fusion using carbon fiber cages. Spine
(Phila Pa 1976) 2003;28:997–1001.
28. Williams AL, Gornet MF, Burkus JK. CT evaluation of lumbar interbody fusion
: current concepts. AJNR Am J Neuroradiol 2005;26:2057–66.
29. Sethi A, Craig J, Bartol S, et al. Radiographic and CT evaluation of recombinant human bone morphogenetic protein-2-assisted spinal interbody fusion. AJR Am J Roentgenol 2011;197:W128–W133.
30. Levine BR, Sporer S, Poggie RA, et al. Experimental and clinical performance of porous tantalum
in orthopedic surgery. Biomaterials 2006;27: 4671–81.
31. Lenhart M, Volk M, Manke C, et al. Stent appearance at contrast-enhanced MR angiography: in vitro examination with 14 stents. Radiology 2000;217:173–8.
32. O'Brien BJ, Stinson JS, Boismier DA, et al. Characterization of an NbTaWZr alloy designed for magnetic resonance
angiography compatible stents. Biomaterials 2008;29:4540–5.
33. Stradiotti P, Curti A, Castellazzi G, et al. Metal-related artifacts in instrumented spine
: techniques for reducing artifacts in CT and MRI: state of the art. Eur Spine
J 2009;18(Suppl 1):102–108.
34. Taniyama T, Sohmura T, Etoh T, et al. Metal artifacts in MRI from non-magnetic dental alloy and its FEM analysis. Dent Mater J 2010;29:297–302.
35. Guermazi A, Miaux Y, Zaim S, et al. Metallic artefacts in MR imaging: effects of main field orientation and strength. Clin Radiol 2003;58:322–8.
36. Ganapathi M, Joseph G, Savage R, et al. MRI susceptibility artefacts related to scaphoid screws: the effect of screw type, screw orientation and imaging parameters. J Hand Surg Br 2002;27:165–70.
- This study was to investigate the use of magnetic resonance imaging in the assessment of lumbar interbody fusion using porous tantalum implants in a porcine spinal fusion model.
- Twelve female Danish Landrace pigs underwent 3 levels of anterior lumbar interbody fusion. Six months postoperatively, different MR sequences, conventional radiographs, and histology were carried out to evaluate the fusion status.
- According to the ROC curve, T1-weighted magnetic resonance (MR) images carried out better than both T2-weighted MR images and conventional radiograph films in detecting fibrous tissue at the vertebrae-implant interfaces.
- T1-weighted spin-echo MR imaging is more sensitive and specific in detecting nonunion via the lucency between the vertebral body and tantalum metal device, compared with T2-weighted MR imaging and conventional radiograph.