Comparison of metal artifact reduction techniques in magnetic resonance imaging of carbon-reinforced PEEK and titanium spinal implants

Abstract

Background

Carbon-reinforced PEEK (C-FRP) implants are non-magnetic and have increasingly been used for the fixation of spinal instabilities.

Purpose

To compare the effect of different metal artifact reduction (MAR) techniques in magnetic resonance imaging (MRI) on titanium and C-FRP spinal implants.

Material and Methods

Rod-pedicle screw constructs were mounted on ovine cadaver spine specimens and instrumented with either eight titanium pedicle screws or pedicle screws made of C-FRP and marked with an ultrathin titanium shell. MR scans were performed of each configuration on a 3-T scanner. MR sequences included transaxial conventional T1-weighted turbo spin echo (TSE) sequences, T2-weighted TSE, and short-tau inversion recovery (STIR) sequences and two different MAR-techniques: high-bandwidth (HB) and view-angle-tilting (VAT) with slice encoding for metal artifact correction (SEMAC). Metal artifact degree was assessed by qualitative and quantitative measures.

Results

There was a much stronger effect on artifact reduction with using C-FRP implants compared to using specific MRI MAR-techniques (screw shank: P < 0.001; screw tulip: P < 0.001; rod: P < 0.001). VAT-SEMAC sequences were able to reduce screw-related signal loss artifacts in constructs with titanium screws to a certain degree. Constructs with C-FRP screws showed less artifact-related implant diameter amplification when compared to constructs with titanium screws (P < 0.001).

Conclusion

Constructs with C-FRP screws are associated with significantly less artifacts compared to constructs with titanium screws including dedicated MAR techniques. Artifact-reducing sequences are able to reduce implant-related artifacts. This effect is stronger in constructs with titanium screws than in constructs with C-FRP screws.

Keywords

Magnetic resonance imaging artifacts carbon spine

Introduction

After surgical stabilization of traumatic or unstable metastatic lesions of the spine, the presence of metallic implants limits diagnostic yield in postoperative imaging, especially in magnetic resonance imaging (MRI) (1,2). Metal implants cause susceptibility artifacts that may hinder appropriate assessment of the spinal canal including the spinal cord and nerves, of implant positioning and of pathologies of the disc (3). With increasing life expectancies due to better oncologic therapies, follow-up MRI has gained increasing importance in patients with primary vertebral tumors and spinal metastatic disease. Metal artifact reduction (MAR) after spinal instrumentation in advanced imaging modalities, i.e. computed tomography (CT) and MRI, has therefore become the focus of recent studies (4 –6).

There are two main strategies to reduce artifacts in MRI: either adapting MRI sequences using MAR techniques (4,6,7) or changing the amount and nature of artifact-provoking elements (e.g. implant geometry and material).

For metallic implants, several MAR sequences have been described and are being provided by the manufacturers of MR scanners (4). Despite significant reduction of artifacts, these sequences are, however, associated with a marked increase in scanning time, which has to be weighed against the potential increase in diagnostic yield (8).

Carbon fiber-reinforced PEEK (C-FRP) implants are radiolucent and non-magnetic, provide comparable mechanical strength in the fixation of spinal instabilities as metallic implants, and have been shown to be clinically safe (9,10). Some studies even suggest that C-FRP spinal implants could provide better biomechanical properties than titanium spinal implants, especially in the osteoporotic spine (11). Artifacts were reported to be significantly reduced in postoperative MRI of the femur when using C-FRP intramedullary femur nails (12). Depending on the model type, however, C-FRP implants also feature small amounts of metal alloy components in order to facilitate accurate positioning under fluoroscopic guidance (e.g. markers). Hence, metal artifacts can also occur with these implants but should be less when compared to standard metal implants.

Yet, so far, no study has investigated the effect of sophisticated MAR techniques on traditional titanium and most recent C-FRP spinal implants.

Thus, the aim of the present ex vivo study was to compare different MAR techniques in MRI of spinal instrumentation and to compare their effect on titanium and C-FRP spinal implants.

Material and Methods

Tissue phantoms

Six fresh-frozen cadaver lumbar spines including peri-spinal soft tissues from skeletally mature female Swiss White Alpine sheep were thawed at room temperature for 12 h and used for this investigation. The cadaver specimens were remainders from other studies at the AO Research Institute in Davos, Switzerland; the animals were sacrificed in accordance with the local laws and regulations on animal welfare by a certified veterinarian and no animal was sacrificed exclusively for this study. The local commission for animal welfare has issued a general waiver for studies involving animal cadavers.

Through a posterolateral approach, rod-pedicle screw constructs were applied bridging four spinal segments of the lumbar spine (L1 to L5) with pedicle screw placed into L1, L2, L4, and L5. Three specimens were instrumented with a total of eight titanium pedicle screws each, with a diameter of 5.5 mm (Legacy 5.5; Medtronic Int., Tolochenaz, Switzerland) and three specimens were instrumented with eight pedicle screws with a diameter of 5.5 mm made of C-FRP marked with an ultrathin titanium shell (CarboClear; Carbofix Orthopedic Ltd., Herzeliya, Israel). For each screw-configuration, either titanium (5.5 mm) or C-FRP rods (6 mm) were mounted resulting in four different implant configurations: (i) titanium screws/titanium rods (n = 3): two titanium rods linked to eight titanium pedicle screws; (ii) titanium screws/carbon rods (n = 3): two C-FRP rods linked to eight titanium pedicle screws; (iii) carbon screws/titanium rods (n = 3): two titanium rods linked eight C-FRP pedicle screws; and (iv) carbon screws/carbon rods (n = 3): two C-FRP rods linked to eight C-FRP pedicle screws.

The intervertebral disc L2/L3 was removed from anteriorly and water-soluble ultrasound transmission gel (Aquasonic 100; Parker Laboratories Inc., Fairfield, NJ, USA) was injected into the disc space at this level representing inflammatory edema found in a discitis.

The specimens were then placed into a thin-walled plastic container (20 L) and the container was filled with rapeseed oil simulating the surrounding fat tissue (Fig. 1).

Fig. 1.

Ovine cadaver phantom. (a) Screws and rod constructs were implanted into an ovine cadaver spine with surrounding soft tissues. (b) The phantom was then placed into rapeseed oil for scanning.

Tissue phantom scan

MRI scans were performed of each configuration on a 3-T scanner (Magnetom Skyra; Siemens Healthcare, Erlangen, Germany) equipped with 45 mT/m gradient strength at 200 mT/m/s slew rate.

An eight-channel spine coil was used for image acquisition. MR sequences included transaxial conventional T1-weighted (T1W) turbo spin echo (TSE), T2-weighted (T2W) TSE, and short-tau inversion recovery (STIR) sequences. Two different MAR techniques were applied to each of the standard sequences: the rather basic high-bandwidth (HB) and a sophisticated view-angle-tilting (VAT) with slice encoding for metal artifact correction (SEMAC) approach. Respective sequence parameters and acquisition times are given in Table 1.

Table 1.

MRI sequence parameters and acquisition times.

MR sequence type	FOV (mm)	TR (ms)	TE (ms)	Voxel size/slice thickness (mm)	Matrix	Scan time (min:s)	Flip angle	Bandwidth (Hz/px)
T1 TSE tra	256	579	9	0.8 × 0.8 × 4.0	320 × 126	0:48	150	252
T2 TSE tra	256	6450	103	0.8 × 0.8 × 4.0	320 × 126	0:53	150	252
T2 TSE STIR tra	256	9980	35	1.0 × 1.0 × 4.0	256 × 115	3:01	150	260
T1 TSE tra HB	256	580	8	0.8 × 0.8 × 4.0	320 × 126	0:48	150	601
T2 TSE tra HB	256	6470	105	0.8 × 0.8 × 4.0	320 × 126	0:53	150	601
T2 TSE STIR tra HB	256	9980	33	1.0 × 1.0 × 4.0	256 × 115	3:01	150	592
T1 TSE tra VAT-SEMAC (SES = 9)	256	580	9	0.8 × 0.8 × 4.0	320 × 126	6:36	150	601
T2 TSE tra VAT-SEMAC (SES = 9)	256	6500	101	0.8 × 0.8 × 4.0	320 × 126	6:06	150	601
T2 TSE STIR tra VAT-SEMAC (SES = 9)	256	5500	35	1.0 × 1.0 × 4.0	256 × 115	6:49	130	592
T2 TSE STIR sag	320	4290	35	1.0 × 1.0 × 4.0	320 × 208	1:36	150	260
T2 TSE STIR sag HB	320	4290	31	1.0 × 1.0 × 4.0	320 × 208	1:36	150	601
T2 TSE STIR sag VAT-SEMAC (SES = 9)	320	4000	33	1.0 × 1.0 × 4.0	320 × 208	6:10	130	601

FOV, field of view; HB, high bandwidth; MRI, magnetic resonance imaging; sag, sagittal; SEMAC, slice encoding for metal artifact correction; SES, slice encoding steps; STIR, short-tau inversion recovery; TE, echo time; TR, relaxation time; tra, transversal; TSE, turbo spin echo; VAT, view-angle-tilting.

Outcome variables

Metal artifact degree was assessed by qualitative and quantitative measures on axial slices.

Primary outcome was the extent of artifact-related implant diameter amplification expressed in the diameters of the screw shank, the screw tulip, and the rod (in mm). These were defined as follows: (i) shank diameter (in mm): maximum diameter of each screw’s shank in mm at the level of the corresponding pedicle; (ii) tulip diameter (in mm): maximum diameter of each screw’s tulip at the level the of corresponding pedicle; and (iii) rod diameter (in mm): maximum diameter of rod and respective halo, measured between two adjacent levels of screw placement.

Secondary outcome was the image quality in terms of clinically relevant findings expressed in a grading of geometric distortion and assessability of the spinal canal, the implants, and the presence of a discitis. These were defined as follows: (i) geometric distortion (0–3, 0 = no distortion, 1 = little distortion, 2 = moderate distortion, 3 = full distortion), graded for each screw; (ii) spinal canal (0–3, 0 = perfect assessability, 1 = slightly impaired assessability, 2 = strongly impaired assessability, 3 = no assessability), graded at each instrumented level; (iii) screw–bone interface (0–3, 0 = perfect assessability, 1 = slightly impaired assessability, 2 = strongly impaired assessability, 3 = no assessability), graded for each screw; (iv) screw placement (0–3, 0 = perfect assessability, 1 = slightly impaired assessability, 2 = strongly impaired assessability, 3 = no assessability), graded for each screw; (v) implant integrity (0–3, 0 = perfect assessability, 1 = slightly impaired assessability, 2 = strongly impaired assessability, 3 = no assessability), graded for each screw. In addition, the ability of recognizing the simulated discitis in the intervertebral disc space L2/L3 was graded (0–3, 0 = perfect assessability, 1 = slightly impaired assessability, 2 = strongly impaired assessability, 3 = no assessability) on mid-sagittal slice positions of sagittal STIR images.

All measurements and ratings were performed independently by two readers (FH, board-certified specialist in radiology, and LCG, medical student) using an image processing package (syngo.via “MR C-Spine tool,” Siemens Healthcare, Erlangen, Germany). Intraclass correlation coefficients (ICC; two-way random, single measure, consistency) were 0.38 (95% confidence interval [CI] = 0.32–0.44; P < 0.001) for shank diameter, 0.78 (95% CI = 0.75–0.81; P < 0.001) for tulip diameter, and 0.81 (95% CI = 0.79–0.84; P < 0.001) for rod diameter.

Statistical analysis

ICCs (two-way random, single measure, consistency) were calculated in order to assess interrater agreement between the two readers who performed the measurements. Post-test analysis was done using SPSS for Windows v 24.0 (IBM, Chicago, IL, USA). All data are reported as mean and standard deviation (SD). A two-factorial ANOVA with a Bonferroni post-hoc correction was used to compare differences across and between the different implant constructs and the different MR sequences. The level of significance was defined as P < 0.05.

Results

Quantitative analysis

ICCs were 0.38 (95% CI = 0.32–0.44; P < 0.001) for shank diameter, 0.78 (95% CI = 0.75–0.81; P < 0.001) for tulip diameter, and 0.81 (95% CI = 0.79–0.84; P < 0.001) for rod diameter.

The mean artifact-generated screw shank diameter across all constructs was 7.9 ± 2.3 mm (range = 3.2–19.1 mm), the mean tulip diameter was 20.6 ± 7.1 mm (range = 7.4–44.7 mm), and the mean rod diameter was 18.2 ± 10.5 mm (range = 5.1–48.9 mm).

VAT-SEMAC metal artifact reducing sequences were able to reduce screw-related signal loss artifacts in constructs with titanium screws to a certain degree (T2W TSE vs. T2W VAT-SEMAC: screw shank: P = 1.0, screw tulip: P = 1.0, rod: P = 0.023; TSE STIR vs. VAT-SEMAC STIR: screw shank: P = 0.379, screw tulip: P < 0.001, rod: P = 0.012). This effect was not seen for HB sequences (P > 0.05) (Fig. 2).

Fig. 2.

Results of the quantitative analysis. Extent of artifact-related implant diameter amplification expressed in the diameters of (a) the screw shank, (b) the screw tulip, and (c) the rod. In X/Y, X represents the screw material and Y the rod material. Ca, C-FRP (screws with an ultrathin titanium shell); Ti, titanium.

In constructs with C-FRP screws, the effect of metal artifact reducing sequences was not significant or very small (T2W TSE vs. T2W VAT-SEMAC: screw shank: P = 0.037, screw tulip: P = 1.0, rod: P = 1.0; T2W TSE STIR vs. VAT-SEMAC STIR: screw shank: P = 0.058, screw tulip: P = 1.0, rod: P = 1.0, HB: P > 0.05).

In general, metal artifact reducing effect of using C-FRP implants was much stronger compared to using specific MAR sequences (screw shank: P < 0.001; screw tulip: P < 0.001; rod: P < 0.001) (Fig. 2).

Implant constructs with C-FRP screws showed significantly less artifact-related implant diameter amplification when compared to constructs with titanium screws and this was independent of using a C-FRP or titanium rod (screw shank: P < 0.001, screw tulip: P < 0.001, rod: P < 0.001). Combining titanium screws with a C-FRP rod slightly reduced artifacts in the region posterior to the spinal canal (screw shank: P = 1.0, screw tulip: P < 0.001, rod: P = 0.005). In contrast, the use of only C-FRP implants was not of relevant advantage when compared to C-FRP screws in combination with a titanium rod (screw shank: P = 1.0, screw tulip: P = 1.0, rod: P = 0.497).

Qualitative analysis

The mean artifact-related geometric distortion was graded 2.1 ± 0.9 (range = 0–3), assessability of the bone–screw interface 2.1 ± 0.8 (range = 0–3), implant integrity 2.1 ± 0.9 (range = 0–3), screw placement 2.0 ± 0.9 (range = 0–3), spinal canal 1.7 ± 1.3 (range = 0–3), and presence of a discitis 1.6 ± 1.1 (range = 0–3).

Metal artifact-reducing sequences were also able to improve parameters of image quality in constructs with titanium screws to a certain degree (Fig. 3). This worked better with VAT-SEMAC sequences (TSE T1W vs. VAT-SEMAC T1W: geometric distortion: P = 0.021, bone–screw interface: P < 0.001, implant integrity: P = 1.0, screw placement: P < 0.001, spinal canal: P <0 .001, discitis: P = 1.0; TSE T2W vs. VAT-SEMAC T2W geometric distortion: P = 0.007; bone–screw interface: P < 0.001, implant integrity: P = 1.0, screw placement: P < 0.001; spinal canal: P < 0.001, discitis: P = 1.0; TSE STIR vs. VAT-SEMAC STIR spinal canal: P < 0.001, all other STIR sequences: P = 1.0) than with HB sequences (TSE T1W vs. HB T1W geometric distortion: P = 0.135, bone–screw interface: P = 0.002, implant integrity: P = 1.0, screw placement: P = 0.018, spinal canal: P = 1.0, discitis: P = 1.0; TSE T2W vs. HB T2W geometric distortion: P = 0.021, bone–screw interface: P = 1.0, implant integrity: P = 1.0, screw placement: P = 0.007, spinal canal: P = 1.0, discitis: P = 1.0).

Fig. 3.

Results of the qualitative analysis. Mean grading of (a) geometric distortion (0–3, no to full distortion), and the assessability of (b) the screw–bone interface (0–3, perfect to no assessability), (c) implant integrity (0–3, perfect to no assessability), (d) the simulated discitis in the intervertebral disc space L2/L3 (0–3, perfect to no assessability), (e) screw placement (0–3, perfect to no assessability), and (f) the spinal canal (0–3, perfect to no assessability). In X/Y, X represents the screw material and Y the rod material. Ca, C-FRP (screws with an ultrathin titanium shell); Ti, titanium.

Again, the effect of metal artifact reducing sequences was insignificant or very small in constructs with C-FRP screws both for HB (TSE T1W vs. HB T1W, TSE T2W vs. HB T2W, and TSE STIR vs. HB STIR for all parameters: P = 1.0) and for HB-WARP STIR sequences (TSE T1W vs. VAT-SEMAC T1W geometric distortion: P = 0.006, bone–screw interface: P = 1.0, implant integrity: P = 0.798, screw placement: P = 0.077, spinal canal: P = 1.0, discitis: P = 1.0; TSE T2W vs. VAT-SEMAC T2W for all parameters: P = 1.0; TSE STIR vs. VAT-SEMAC STIR geometric distortion: P = 1.0, bone–screw interface: P = 1.0, implant integrity: P = 0.219, screw placement: P = 0.018, spinal canal: P = 0.080, discitis: P = 1.0).

In addition, for the qualitative grading of image quality, there was a much stronger effect of using C-FRP implants when compared to using specific MR sequences (geometric distortion: P < 0.001; bone–screw interface: P < 0.001; implant integrity: P < 0.001; screw placement: P < 0.001; spinal canal: P < 0.001). The MR sequence had more effect on image quality only for detecting a simulated discitis on sagittal STIR sequences (P < 0.001).

Image quality was significantly better in constructs with C-FRP screws when compared to constructs with titanium screws and this, again, was independent of using a C-FRP or titanium rod (geometric distortion: P < 0.001, bone–screw interface: P < 0.001, implant integrity: P < 0.001, screw placement: P < 0.001, spinal canal: P < 0.001) (Figs. 3 –5). This effect was not seen for the detection of a discitis (P > 0.05).

Fig. 4.

Impact of MAR techniques. (a–c) Impact of (b) high bandwidth and (c) VAT-SEMAC techniques on severest susceptibility artifacts (arrowheads), compared to (a) conventional sequences, demonstrated in representative (1) T1W TSE, (2) T2W TSE, and (3) T2W TSE STIR images of a phantom with titan screws and titan rods. MAR, metal artifact reduction; SEMAC, slice encoding for metal artifact correction; T1W/T2W, T1-/T2-weighted; TSE, turbo spin echo; VAT, view-angle-tilting.

Fig. 5.

Impact of implant material. Exemplary T2w TSE STIR images of (1) titanium and (2) C-FRP screws (arrowheads) in (b) VAT-SEMAC sequences, compared to (a) conventional sequences.

Combining titanium screws with a C-FRP rod did not reduce artifacts in a way improving assessability of the spinal canal and the implants (geometric distortion: P = 1.0, bone–screw interface: P = 1.0, screw placement: P = 0.982, spinal canal: P = 0.450). Only implant integrity (P = 0.036) and detection of a discitis (P = 0.002) were facilitated.

While geometric distortion (P < 0.001) and assessability of the bone–screw interface (P < 0.001) and implant integrity (P = 0.001) were improved with the use of all C-FRP implants compared to C-FRP screws with a titanium rod, this effect was not seen for the assessability of screw placement (P = 0.168), the spinal canal (P = 1.0), and the detection of a discitis (P = 0.500).

Discussion

The aim of this ex vivo study using ovine cadaver spine phantoms was to compare different MAR techniques in MRI of spinal instrumentation and to compare their effect on titanium and C-FRP spinal implants. It was shown that constructs with C-FRP screws are associated with significantly less artifacts compared to constructs with titanium screws and this effect on MAR was much stronger than any advanced MAR MRI technique. The implant material of the rod as opposed to tulip and screw was only of minor relevance with regard to the artifact degree and associated parameters investigated.

As shown previously (3,4), this study confirmed the beneficial effect of the used MAR MRI techniques HB, VAT, and SEMAC. This especially accounts for VAT-SEMAC sequences and – as expected – the effect of these sequences was, in particular, seen in constructs with titanium screws rather than in constructs with C-FRP screws, as there is very little paramagnetic material in place leading to few “pile-up” and distortion effects that can be corrected. HB sequences reduce metallic artifacts by increasing readout receiver bandwidth. A correction of in-plane geometric distortions can be achieved by VAT, where the slice selection gradient is being repeated during signal readout (13). However, VAT may cause blurring effects and geometric through-plane distortion. SEMAC combines the VAT technique with variable slice encoding steps for a better correction of through-plane artifacts. Yet, sophisticated MAR strategies in MRI are also associated with a marked increase in scan time (3 min 1 s for standard transaxial STIR vs. 6 min 41 s for VAT-SEMAC STIR). This has to be weighed against a potential increase in image quality and thus diagnostic yield.

In both the quantitative and qualitative analyses, it was shown that changing the implant material, i.e. using C-FRP implants instead of titanium implants, had a greater effect on image quality than using an advanced MAR technique versus a conventional MR sequence. This is in line with previously published data on MRI artifacts with titanium and C-FRP pedicle screws (1,14). Titanium implants are made of ferromagnetic alloys. Consequently, severe metal-induced inhomogeneities of the magnetic field with artifacts in MRI scans are produced that result in geometric distortion and deterioration of the image contrast. Even in C-FRP screws – with only a small amount of metal alloy in order to facilitate fluoroscopically guided screw placement – significant local magnetic field inhomogeneities are induced by the implant. However, this effect’s magnitude generally correlates with the amount of metal volume in place and is thus less in C-FRP implants compared to full-titanium instrumentations.

Using C-FRP intramedullary nails instead of titanium implants can significantly reduce artifacts in postoperative MRI and CT of the femur (12). Recent studies could confirm the advantage of C-FRP implants for MRI after instrumentation of the spine (15,16). Ringel et al. (16) investigated pedicle screws with C-FRP shanks and titanium tulips. In their study, the metallic tulips had only little effect with 1.5-T MRI scans after pedicle screw fixation of spinal metastases but produced relevant artifacts with 3-T MRI. Screws made completely of C-FRP reinforced polymers with only a thin titanium shell do not seem to follow this pattern as could be shown in the present study where 3-T MRI scans were conducted.

The present study has some limitations, which are inherent to the design of a phantom study. The conditions in an animal model can only approximate the relative content of water and fat of a living human. However, sheep are generally considered a realistic animal model for the human spine. Second, all measurements were conducted on axial MR images except for qualitative assessment of discitis on sagittal STIR images. This mirrors clinical practice where physicians commonly assess implant position and compression of the spinal cord on axial images. Third, we investigated the effect of different MAR strategies on standard 2D TSE MR sequences. Although 3D gradient-echo MR sequences have become popular as they deliver high spatial-resolution and contrast, 2D TSE are still considered the backbone in musculoskeletal imaging protocols due to their robustness. Lastly, we have investigated metal artifacts on a 3-T state-of-the-art MR unit. Lower field strengths, e.g. 1.5 T, may further contribute to a significant artifact reduction, yet the advantageous effect of C-FRP versus titanium implants in terms of artifact reduction does also apply to lower field strengths.

In conclusion, constructs with C-FRP screws are associated with significantly less artifacts compared to constructs with titanium screws including dedicated MAR techniques. Artifact reducing sequences are able to reduce implant-related artifacts. This effect is stronger in constructs with titanium screws than in constructs with C-FRP screws.

Footnotes

Declaration of conflicting interests

The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: The authors of this manuscript declare relationships with the following companies: Georg Osterhoff has given paid lectures for Medtronic; The Department of Trauma of the University Hospital Zurich has received funding for events from Medtronic and Carbofix.

Funding

The author(s) received the following financial support for the research, authorship, and/or publication of this article: This study received funding from Medtronic (implants in kind), Carbofix (implants in kind), and OrthoContor (sheep specimens).

ORCID iDs

Georg Osterhoff

Florian A Huber

References

Malik

Boyko

Aktar

, et al. A comparative study of MR imaging profile of titanium pedicle screws. Acta Radiol 2001; 42:291–293.

Imai

Tanaka

Nomura

, et al. Three-dimensional quantification of susceptibility artifacts from various metals in magnetic resonance images. Acta Biomater 2013; 9:8433–8439.

Viano

Gronemeyer

Haliloglu

, et al. Improved MR imaging for patients with metallic implants. Magn Reson Imaging 2000; 18:287–295.

Padua

Goerner

, et al. SEMAC-VAT and MSVAT-SPACE sequence strategies for metal artifact reduction in 1.5T magnetic resonance imaging. Invest Radiol 2012; 47:267–276.

Guggenberger R, Winklhofer S, Osterhoff G, et al. Metallic artefact reduction with monoenergetic dual-energy CT: Systematic ex vivo evaluation of posterior spinal fusion implants from various vendors and different spine levels. Eur Radiol 2012; 22: 2357–2364.

Agten

Del Grande

Fucentese

, et al. Unicompartmental knee arthroplasty MRI: Impact of slice-encoding for metal artefact correction MRI on image quality, findings and therapy decision. Eur Radiol 2015; 25:2184–2193.

Filli

Luechinger

Frauenfelder

, et al. Metal-induced artifacts in computed tomography and magnetic resonance imaging: Comparison of a biodegradable magnesium alloy versus titanium and stainless steel controls. Skeletal Radiol 2015; 44:849–856.

Fritz

Thawait

, et al. Advanced metal artifact reduction MRI of metal-on-metal hip resurfacing arthroplasty implants: compressed sensing acceleration enables the time-neutral use of SEMAC. Skeletal Radiol 2016; 45:1345–1356.

Boriani

Tedesco

Ming

, et al. Carbon-fiber-reinforced PEEK fixation system in the treatment of spine tumors: A preliminary report. Eur Spine J 2018; 27:874–881.

10.

Lindtner

Schmid

Nydegger

, et al. Pedicle screw anchorage of carbon fiber-reinforced PEEK screws under cyclic loading. Eur Spine J 2018; 27:1775–1784.

11.

Oikonomidis

Greven

Bredow

, et al. Biomechanical effects of posterior pedicle screw-based instrumentation using titanium versus carbon fiber reinforced PEEK in an osteoporotic spine human cadaver model. Clin Biomech 2020; 80:105153.

12.

Zimel

Hwang

Riedel

, et al. Carbon fiber intramedullary nails reduce artifact in postoperative advanced imaging. Skeletal Radiol 2015; 44:1317–1325.

13.

Deligianni

Bieri

Elke

, et al. Optimization of scan time in MRI for total hip prostheses: SEMAC tailoring for prosthetic implants containing different types of metals. Rofo 2015; 187:1116–1122.

14.

Fleege

Makowski

Rauschmann

, et al. Carbon fiber-reinforced pedicle screws reduce artifacts in magnetic resonance imaging of patients with lumbar spondylodesis. Sci Rep 2020; 10:16094.

15.

Ernstberger

Buchhorn

Heidrich

Artifacts in spine magnetic resonance imaging due to different intervertebral test spacers: An in vitro evaluation of magnesium versus titanium and carbon-fiber-reinforced polymers as biomaterials. Neuroradiology 2009; 51:525–529.

16.

Ringel

Ryang

Y-M

Kirschke

, et al. Radiolucent carbon fiber-reinforced pedicle screws for treatment of spinal tumors: advantages for radiation planning and follow-up imaging. World Neurosurg 2017; 105:294–301.