Vol.1, No.3, 207-210 (2009)
doi:10.4236/health.2009.13035
SciRes Copyright © 2009 http://www.scirp.org/journal/HEALTH/
Health
Openly accessible at
Implant-related MRI artifacts of determined interbody
test spacers: artifact calculations due to implant
parameters in a porcine spine model
Thorsten Ernstberger
Clinic for Spinal Surgery, Klinikum Bad Bramstedt, Bad Bramstedt, Germany;
Received 21 September 2009; revised 19 October 2009; accepted 20 October 2009
ABSTRACT
Aim: Intervertebral spacers for anterior spine
fusion are made of different materials, which
can affect the post-fusion MRI scans. Suscep-
tibility artifacts specially for implants made of
titanium alloys can decrease the image quality.
This study focused on the influence of deter-
mined implant parameters like shape and implant
volume in MRI artifacting independent from se-
lected MRI-sequences. Methods: In this study
the post-implantation MRI scans of determined
cuboids and cylinders were evaluated. All in-
terbody test implants were made of titanium
alloys. MRI scans were carried out by using T1
TSE sequences. The total artifact volume (TAV)
of all examined implants were calculated for sta-
tistical t-test correlation and implant volume
(IV)/TAV-relation. Results: Considering all ex-
amined test implants with an increasing implant
size the TAV became significant larger (p<0,001)
with simultaneous reduction of the respective
IV/TAV-relation. According to an intergroup TAV-
correlation for cylinders and cuboids with an
equivalent implant volume the cylindric test im-
plants demonstrated a significant smaller arti-
fact range (p<0,05). Conclusions: Based on
these results the MRI artifacts of larger test im-
plants were more limited to the to the implant’s
direct surroundings. In this connection for im-
plants with identical material volumes a cylin-
dric shape demonstrated more advantages con-
sidering MRI artifacting than cubic forms.
Keywords: Magnetic Resonance Imaging;
Geometrical Test Implants; Susceptibility
Effect; Titanium Interbody Spacer
1. INTRODUCTION
In the preoperative diagnostics of spinal diseases, ma g -
netic resonance imaging (MRI) is used as a standard
procedure that can visualize disc pathologies and neuro-
logical changes of the spinal canal with high precision.
When anterior spine fusion proves indicated, implanta-
tion of interbody spacers often represents the treatment of
choice. Interbody spacers made of various materials can
be used as stand-alone cages for exclusively anterior
fusion or in combination with dorsal instrumentation for
dorsoventral fusion.
When postoperative complications arise secondary to
vertebra fusion, MRI scans are frequently necessary to
evaluate implant position and demonstrate any clinically
relevant abnormalities and to direct further surgical deci-
sion-making [1]. However, implant-related susceptibility
artifacts can negatively impact the complex post-fusion
evaluation of MRI scans. Depending on the spacer material,
a local magnetic field gradient of varying susceptibility
results in the area between structures. In these border areas,
the respective spins gyrate with different frequencies and
cause image distortions and susceptibility artifacts [2-4].
The MRI imaging behavior of metallic spinal implants
is well documented in the literature [5-11]. However, the
aims of the published studies differed in that most focused
on determining sequence-related artifact size.
Studies on metallic artifacts in MRI of the anterior thor-
acic and lumbar spine have been conducted by Vaccaro et
al. [11] and Wang et al. [12]. In one cadaveric study,
Vaccaro et al. [11] examined the MRI artifact rates of
different metal particles introduced in predefined in-
tervertebral drill holes and subsequently embedded in
paraffin. Vaccarro could not demonstrate any significant
artifacts in T1- or T2 SE sequences, probably due to the
fact that the particle density was lower than that produced
by metal implants commonly used in clinical practice.
In another cadaveric artifact study, Wang et al. [12]
described the MRI behavior of an intervertebral spacer
made of titanium. Using T1 SE sequences, the implant-
related artifact rate of the titanium spacer was primarily
limited to the implant’s direct surroundings and anatomic
neighboring structures were clearly distinguishable. In a
phantom study by Rudisch et al. [7]. the relevance of

T. Ernstberger / HEALTH 1 (2009) 207-210
SciRes Copyright © 2009 http://www.scirp.org/journal/HEALTH/
208
Table 1. Test implant’s parameters and artifact calculations.
Implant parameters
Cylinder:
Height x base diameter
Cuboid:
Height x depth x width
(cm)
Implant volume
(IV)
(cm3)
Total artifact vol-
ume
(TAV)
(+ s.d.)
(cm3)
TAV
p-value
Relation
(IV / TAV)
Small: 1,5 x 1,0 x 1,0 1,5 7,0 (+ 0,14) 1 : 4,7
Medium: 2,0 x 1,2 x 1,2 2,9 11,2 (+ 0,11) 1 : 3,9
Implant
group 1
Large: 2,5 x 1,4 x 1,4 4,9 15,2 (+ 0,16)
p<0,001
1 : 3,1
Small: 1,5 x 1,0 1, 6,2 (+ 0,19) 1 : 5,2
Medium: 2,0 x 1,2 2,3 9,1 (+ 0,23) 1 : 3,9
Implant
group 2
Large: 2,5 x 1,4 3,9 11,4 (+ 0,11)
p<0,001
1 : 2,9
Small: 1,91 x 1,0 1,5 6,5 (+ 0,23) 1 : 4,3
Medium: 2,55 x 1,2 2,9 9,7 (+ 0,23) 1 : 3,4
Implant
group 3
Large: 3,18 x 1,4 4,9 14,3 (+ 0,26)
p<0,001
1 : 2,9
a: Cuboids and cylinders with equivalent heights (Implant group 1 + 2).
b: Cylinders with equivalent cuboid volumes (Implant group 3).
Figure 1. a + b: Titanium test implants.
Openly accessible at
Figure 2. Cadavaric porcine spine model
with an implanted large titanium cuboid.
metallic artifacts and implant-related characteristics, such as
implant material, shape and position, was demonstrated in
addition to an impact by the selected MRI sequence. In spite
of the use of optimum MRI sequences, variability in the
amount of susceptibility artifacts must be accounted for
when evaluating MRI scans of metallic spine implants. This
experimental study mainly focused on the influence of the
implant parameters shape and volume in MRI artifacting.
Therefore in an in vitro spine model, we evaluated the
post-implantation MRI scans of determined cubic and
cylindric test implants made of titanium alloy. Our
hypothesis was that defined implant parameters of geom-
etrical implants would have a clear affect to the range of
susceptibility artifacts independent from the selected MRI
sequences or used implant materials.
2. MATERIAL AND METHODS
In this MRI-study we assessed 3 cuboids (Implant group 1)
and 3 cylinders (Implant group 2) (Figure 1a) made of
titanium-aluminium-vanadium alloy (TiAL6V4). The test
implants of both groups were divided in small, medium and
large on account to the implant height. Considering the
implant volumes of the cuboids in a third group titanium
cylinders with equivalent volumes (Implant group 3) were
examined additionally (Figure 1b). The implant parame-
ters of all test implants were listed in Table 1. To visualize
differences of the artifact range the respective test implants
were exactly placed between adjacent vertebras of a ca-
daveric Göttingen mini pig spine model (Figure 2). The
porcine spine model was completely coated with a soft
tissue mass and later stored into a plastic container. For
comparable trial conditions markings were drawn to the
container wall to reproduce the spine/implant position.
After completion of the preparation the MRI investigation
followed. For less MRI artifacting MRI scans were carried
out by using T1 TSE sequences [13,14].
2.1. Magnetic Resonance Imaging
MRI was performed with a 1.5T MRI (Magnetom-

T. Ernstberger / HEALTH 1 (2009) 207-210
SciRes Copyright © 2009 http://www.scirp.org/journal/HEALTH/
209
Table 2. TAV inter group correlation of implant group 1 + 3.
Implant
group
Implant volume
(cm3)
Total artifact volume
(TAV)
(+ s.d.)
(cm3)
TAV
(t-test)
p-value
1 7,0 (+ 0,14)
3 Small: 1,5 6,5 (+ 0,23) 0,027
1 11,2 (+ 0,11)
3 Medium: 2,9 9,7 (+ 0,23) 0,002
1 15,2 (+ 0,16)
3 Large: 4,9 14,3 (+ 0,26)
0,006
Symphony,Siemens AG Medical Solutions, Erlangen,
Germany. T1w-TSE sequences (TR: 600 + 2260, TE: 14,
Flip angle: 15, Band width: 150) were used to acquire a
slice thickness of 3 mm (Figure 3a-c). We selected a
matrix with 512 x 512 combined with a Field of View
(FOV) of 500 mms.
All MRI scans were evaluated 5 times to determine
the implant-related TAV by using an actual version of the
DICOM reader software. Based on the multisection slice
technique for cardiovolumetric MRI analysis (15), the
respective TAV were calculated. The artifact area of
every MRI slice were determined and later added regar-
ding a slice thickness of 3 mm. Additionally we
cal-culated the implant volume/TAV-relation.
2.2. Statistical Analysis
Considering a reliable analysis the TAV as well as implant
volume (IV)/ TAV-relation of the cuboids and cylinders
were determined. To calculate significant differences of the
respective TAV within every implant group Newman-
Keuls multiple comparisons were carried out (Table 1).
Additionally the TAV of cylindric and cubic test implants
with an equivalent implant volume (Implant group 1+3)
were tested for statistical t-test correlation (Table 2).
3. RESULTS
Concerning the repeated artifact measurements for each
implant group at the 0,05 significans level no significant
differences could be demonstrated. Independent from the
implant shape with an increasing implant size the TAV
became significant larger (p<0,001) with simultaneous
reduction of the IV /TAV relation within the respective
implant group (Table 1). In this context with an increas-
ing implant size the artifact range were more limited to
the implant´s surrounding. Considering an intergroup
TAV correlation a statistical significance between cuboids
and cylinders with equivalent implant volumes (Implant
group 1+3) could be demonstrated (p<0,05) (Table 2).
All cylinders of implant group 3 represented a significant
smaller artifact range than the correponding cubic test
implants of group 1. In this connection for implants with
identical material volumes a cylindric shape demon-
strated more advantages considering MRI artifacting than
cubic forms. Additional correlations did not result in
a: Cylinder (small).
b: Cuboid (large).
c: Cylinder (medium)with equivalent cuboid volume.
Figure 3. a-c: Median MRI artifact range depicted in a
selection of 3 test implants.
further information.
4. DISCUSSIONS
The disadvantages associated with bone grafting alone
has led to the development of interbody spacers to en-
hance anterior spinal fusion [16,17]. The use of inter-
body spacers of different designs and materials has thus
become increasingly widespread in clinical routine be-
cause they offering immediate load transmission with
direct primary stability. In this context the implant
design of many interbody spacers derived from a cylin-
drical or cubic prototype.
In our department, MRI is the radiological diagnostic
method of choice for clarifying post-fusion questions
regarding the involvement of osseous and soft tissue str-
uctures in relation to implant position. In these indica-
tions, MRI is better suited than multisection CT to dem-
onstrate myelopathies, inflammatory and infectious pro -
cesses and any neurodegenerative changes. When posto-
perative complications arise secondary to vertebra fusion,
MRI scans are frequently necessary to demonstrate any
clinically relevant abnormalities and to direct further
surgical decision-making [4].
The MRI imaging behavior of spinal implants has
been widely studied [5-12]. However, the aims of the
Openly accessible at

T. Ernstberger / HEALTH 1 (2009) 207-210
SciRes Copyright © 2009 http://www.scirp.org/journal/HEALTH/Openly accessible at
210
published studies differed in that most focused on deter-
mining sequence-related artifact size.
In spite of the use of optimum MRI sequences, vari-
ability in the amount of susceptibility artifacts must be
accounted for when evaluating MRI scans of metallic
spine implants.
Ernstberger et al [18] evaluated in an in vitro study the
post-implantation MRI scans of 3 intervertebral disc
spacers that differed in shape, material (Titanium,
Carbon, Cobalt-chromium), surface qualities and imp-
lantation technique. A spacer made of human cortical
bone was used as a control. The respective artifact-
affected image quality was rated on a developed score.
Turbo spin echo sequences produced the best scores for
all spacers and the control. Only the control achieved a
score of 100%. The carbon, titanium and cobalt-chrome
spacers scored 83.3%, 62.5% and 50%, respectively.
In this study geometrical implant forms like cylinders
and cuboids were chosen to prove possible coherences
between the range of MRI artifacting and determined
implant related factors like shape and implant volume.
For the cuboids as well as cylinders the range of MRI
artifacts was directly affected by the implant size and
volume. In this context the smaller the implant the
smaller the range of susceptibility artifacts. For implants
characterized by equivalent materials and implant
volumes the influence of the implant shape on the
artifact size has to be proofed with priority on an artifact
volumetric analysis. In this context significant diff-
erences could be determined in favor of a cylindric
implant shape. For all examined implant groups a more
advantageous IV/TAV-relation could be recognized the
larger the implants. In this connection using T1 TSE
sequences, the expected implant-related artifact rate are
more limited to the implant’s direct surroundings. On the
basis of our study results the range of artifacts of our
used test implants were influenced by material volume
as well as implant shape. Considering the implant design
of current intervertebral spacers further studies are nec-
cessary to interpreted the influence and concurrency of
implant characteristics in MRI artifacting.
REFERENCES
[1] J. W. Van Goethem, P. M. Parizel, J. R. Jinkins, (2002)
Review article: MRI of the postoperative lumbar spine.
Neuroradiology, 44, 723-39.
[2] C. Fellner, M. Behr, F. Fellner, P. Held, G. Handel, S.
Feuerbach, (1997). Artifacts in MR imaging of the tem-
poromandibular joint caused by dental alloys: A phantom
model study at T1.5. Rofo, 166,421-8.
[3] S. Fritzsche, R. Thull, A. Haase, (1994) Reduction of
artifacts in magnetic resonance images by using optimized
materials for diagnostic devices and implants. Biomed
Tech (Berl), 39, 42-6.
[4] J. F. Schenck, (1996) The role of magnetic susceptibility
in magnetic resonance imaging: MRI magnetic compati-
bility of the first and second kinds. Med Phys, 23, 815-50.
[5] A. S. Malik, O. Boyko, N. Atkar, W. F. Young, (2001) A
comparative study of MR imaging profile of titanium
pedicle screws. Acta Radiol, 42, 291-3.
[6] C. A. Petersilge, J. S. Lewin, J. L. Duerk, J. U. Yoo, A. J.
Ghaneyem, (1996) Optimizing imaging parameters for
MR evaluation of the spine with titanium pedicle screws.
AJR Am J Roentenol, 166, 1213-8.
[7] A. Rudisch, C. Kremser, S. Peer, A. Kathrein, W. Jud-
maier, H. Daniaux, (1998) Metallic artifacts in magnetic
resonace imaging of patients with spinal fusion. A com-
parison of implant materials and implant sequences. Spine,
23, 692-9.
[8] R. Rupp, N. A. Ebraheim, E. R. Savolaine, W. T. Jackson,
(1993) Magnetic resonance imaging evaluation of the spine
with metal implants. General safety and superior imaging
with titanium. Spine, 18, 379-85.
[9] M. Thomsen, U. Schneider, S. J. Breusch, J. Hansmann, M.
Freund, (2001) Artefacts and ferromagnetism dependent on
different metal alloys in magnetic resonance imaging. An
experimental study. Orthopade, 30, 40-4.
[10] J. C. Wang, H. S. Sandhu, M. D. Yu, J. T. Minchew, R. B.
Delamarter, (1997) MR parameters for imaging titanium
spinal instrumentation. J Spinal Disord, 10, 27-32
[11] A. R. Vaccaro, R. M. Chesnut, G. Scuderi, J. F. Healy, J. B.
Massie, S. R. Garfin, (1994) Metallic spinal artifacts in
magnetic resonance imaging. Spine, 19, 1237-42.
[12] J. C. Wang, W. D. Yu, H. S. Sandhu, V. Tam, R. B.
Delamarter, (1998) A comparison of magnetic resonance
and computed tomographic image quality after the implan-
tation of tantalum and titanium spinal instrumentation.
Spine, 23, 1684-8.
[13] C. B. Henk, W. Brodner, S. Grampp, M. Breitenseher, M.
Thurnher, G. H. Mostbeck, H. Imhof, (1999) The postop-
erative spine. Top Magn Reson Imaging, 10, 247-64.
[14] T. Herold, W. C. Caro, G. Heers, L. Perlick, J. Grifka, S.
Feuerbach, W. Nitz, M .Lenhart, (2004) Influence of se-
quence type on the extent of the susceptility artifact in MRI-
a shoulder specimen study after suture anchor repair. Rofo,
176, 1296-301.
[15] J. F. Debatin, S. N. Nadel, H. D. Sostman, C. E. Spritzer, A.
J. Evans, T. M. Grist, (1992) Magnetic resonance imaging
cardiac ejection fraction measurements. Phantom study
comparing four different methods. Invest Radiol, 27, 198-
204.
[16] B. N. Summers and S. M. Eisenstein, (1989) Donor site
pain from the ilium. A complication of lumbar spine fusion.
J Bone Joint Surg Br, 71, 677-80.
[17] J. A. Goulet, L. E. Senunas, G. L. DeSilva, M. L. Green-
field, (1997) Autogenous iliac crest bone graft. Complica-
tions and functional assessment. Clin Orthop, 339, 76-81.
[18] T. Ernstberger, G. Heidrich, T. Bruning, S. Krefft, G.
Buchhorn, H. M. Klinger, (2007) The inter-observer vali-
dated relevance of intervertebral spacer materials in MRI
artifacting. Eur Spine J, 16, 179-85.