Original Article

Biomechanical Analysis of Porous Additive Manufactured Cages for Lateral Lumbar

Interbody Fusion: A Finite Element Analysis

Zhenjun Zhang1,2, Hui Li3, Guy R. Fogel4, Zhenhua Liao2, Yang Li1,2, Weiqiang Liu1,2

-BACKGROUND: A porous additive manufactured (AM)cage may provide stability similar to that of traditionalsolid cages and may be beneficial to bone ingrowth. Thebiomechanical influence of various porous cages on sta-bility, subsidence, stresses in cage, and facet contactforce has not been fully described. The purpose of thisstudy was to verify biomechanical effects of porous AMcages.

-METHODS: The surgical finite element models withvarious cages were constructed. The partially porous ti-tanium (PPT) cages and fully porous titanium (FPT) cageswere applied. The mechanical parameters of porous ma-terials were obtained by mechanical test. Then the porousAM cages were compared with solid titanium (TI) cage andsolid polyetheretherketone (PEEK) cage. The 4 motionmodes were simulated. Range of motion (ROM), cagestress, end plate stress, and facet joint force (FJF) werecompared.

-RESULTS: For all the surgical models, ROM decreasedby >90%. Compared with TI and PPT cages, PEEK and FPTcages substantially reduced the maximum stresses in cageand end plate in all motion modes. Compared with PEEKcages, the stresses in cage and end plate for FPT cagesdecreased, whereas the ROM increased. Comparing FPTcages, the stresses in cage and end plate decreasedwith increasing porosity, whereas ROM increased with

Key words- Biomechanics- Facet joint force (FJF)- Finite element analysis (FEA)- Lumbar interbody fusion- Porous cage- Range of motion (ROM)- Subsidence

Abbreviations and AcronymsAM: Additive manufacturedFE: Finite elementFJF: Facet joint forceFPT: Fully porous titaniumIDP: Intervertebral disc pressurePEEK: PolyetheretherketonePPT: Partially porous titanium

WORLD NEUROSURGERY 111: e581-e591, MARCH 2018

increasing porosity. After interbody fusion, FJF was sub-stantially reduced in all motion modes except for flexion.

-CONCLUSIONS: Fully porous cages may offer an alter-native to solid PEEK cages in lateral lumbar interbodyfusion. However, it may be prudent to further increase theporosity of the cage.

INTRODUCTION

umbar interbody fusion has been used in the treatment oflumbar diseases such as spondylolisthesis, trauma, and

Ldegenerative disc degeneration. Successful clinical out-

comes depend on fusion healing. The best opportunity for healingmay be anterior interbody fusion with better loading of the graftand the largest surface area for fusion.1-4 Laterally inserted inter-body cages may decrease range of motion (ROM) compared withother cages.5-8 The interbody cage fusion improves the loadingcapacity of the anterior column.2,3 In addition, supplemental fix-ation favorably influences the healing of the lumbar fusion;therefore, bilateral pedicle screw fixation is often used to sup-plement the interbody cage because it can provide multiplanarstability.4,5,9-11

Traditional solid cages made of titanium or poly-etheretherketone (PEEK) have been widely used in lumbarfusion.12,13 However, solid cages have high mechanical stiffness,which may affect the loading mechanism of the lumbar spine.14,15

ROM: Range of motionTI: Titanium

From the 1Department of Mechanical Engineering, Tsinghua University, Beijing;2Biomechanics and Biotechnology Lab, Research Institute of Tsinghua University inShenzhen, Shenzhen; 3Naton Science and Technology Group, Beijing, China; and 4Spine PainBegone Clinic, San Antonio, TX, USA

To whom correspondence should be addressed: Weiqiang Liu, Ph.D.[E-mail: weiqliu@hotmail.com]

Citation: World Neurosurg. (2018) 111:e581-e591.https://doi.org/10.1016/j.wneu.2017.12.127

Journal homepage: www.WORLDNEUROSURGERY.org

Available online: www.sciencedirect.com

1878-8750/$ - see front matter ª 2017 Elsevier Inc. All rights reserved.

www.WORLDNEUROSURGERY.org e581

ORIGINAL ARTICLE

ZHENJUN ZHANG ET AL. BIOMECHANICAL ANALYSIS OF POROUS LUMBAR FUSION CAGES

A porous additive manufactured (AM) cage can reduce themechanical stiffness, may provide stability similar to that oftraditional solid cages, and may be beneficial to bone ingrowth.16

Currently, there have been a variety of porous AM cages, such asthe Posterior Spine Truss System (4WEB Medical, Frisco, Texas,USA), CASCADIA Lateral 3D Interbody System (K2M GroupHoldings, Inc., Leesburg, Virginia, USA), EndoLIF O-Cage(joimax GmbH, Karlsruhe, Germany), Tritanium PL PosteriorLumbar Cage (Stryker Corporation, Kalamazoo, Michigan, USA),and 3D ACT (AK Medical, Beijing, China).Although some factors such as lumbar instability, bone quality

of vertebral trabecular and end plate, and previous operationhistory are known to affect the choice of cage, the biomechanicsof cages made of various materials are not fully investigated.Compared with solid cages, partially porous titanium (PPT)cages and fully porous titanium (FPT) cages have a porousstructure, along with rough surfaces, to allow for bony integra-tion throughout the implant. Some previous studies haveinvestigated the mechanical performance of porous AM cagesand evaluated their biomechanics. Lee et al.17 developed aporous cage with 50% porosity and compared 3 lumbar fusiontechniques by using an L3-4 bone implant model. Kang et al.18

investigated the porous biodegradable lumbar interbody fusioncage using a finite element (FE) model of mini pig L2-5 lum-bar spine. Additionally, Tsai et al.,19 in their recent numericaland experimental study, found that porous AM cages with aporosity of between 69% and 80% could provide betterbiomechanical performances. However, the influence of thevarious porous cages (solid cage, PPT cage, and FPT cage) andtheir different porosities on lumbar stability, cage stress, endplate stress, and facet joint force (FJF) have not been fullydescribed.Using an FE model to systematically study the biomechanical

effects of various cages may be valuable. Further information to bedetermined is the influence of various porous structures on thesubsidence, which is induced by the changed stiffness in the indexdisc space associated with the interbody cage and supplementedfixation. The aim of this study was to evaluate the lateral lumbarinterbody fusion cages by comparing the biomechanical propertiesof porous cages with that of solid cages, and to verify thebiomechanical effects of cage porosity.

MATERIALS AND METHODS

Computed tomography images of intact lumbar spine with intervalof 0.7 mm were obtained from a 36-year-old woman (weight, 52kg; height, 158 cm; excluded for lumbar disease based on visualand radiographic examination). A total of 492 computed tomog-raphy images were imported into Mimics (Materialise Inc.,Leuven, Belgium). In Mimics, the 3-dimensional geometry struc-ture was constructed, which consists of vertebrae, intervertebraldisc, and cartilage end plate. The geometric structure was meshedusing Hypermesh (Altair Technologies, Inc., Fremont, California,USA). Finally, the mesh model was imported into Abaqus(Simulia, Inc., Providence, Rhode Island, USA) to perform FEanalysis. The computer for the simulation was the ThinkStation(Lenovo, Beijing, China), configured with 24 processors and 64GB memory.

e582 www.SCIENCEDIRECT.com WORLD NE

Figure 1 shows the FE model of intact lumbar spine L1-5. Thevertebral body was divided into 3 parts: cortical bone, cancellousbone, and posterior bone. The intervertebral disc was divided intonucleus pulposus and annulus ground. The intact model included7 kinds of ligaments: anterior longitudinal ligament, posteriorlongitudinal ligament, ligamenta flava, interspinous ligament,supraspinal ligament, intertransverse ligament, and capsular lig-ament. The cortical bone was 1.0-mm thick, and the end plate was0.5-mm thick. All the ligaments were modeled with 3-dimensionaltruss elements (T3D2) which had a tension-only property. The FEmodel was meshed using the 3-dimensional tetrahedral elements,except for the ligaments. There were 195,533 nodes and 841,038elements contained in the intact model, which could effectivelyeliminate the influence of meshing on the accuracy of calculation.The PPT cage was configured as both a solid and porous

structure. The outer part was solid titanium (TI) structure, and theinner part was porous TI structure. The porosity (vol%) and poresize (mm) were chosen according to the previous literature.19,20

There were 3 kinds of PPT cages: PPT cage with 65% porosity,PPT cage with 75% porosity, and PPT cage with 80% porosity. TheFPT cage was configured as a porous structure. There were 3 kindsof FPT cages: FPT cage with 65% porosity, FPT cage with 75%porosity, and FPT cage with 80% porosity. The average pore size ofall the porous cages was 350e400 mm. Including the solid TI cageand solid PEEK cage, there were 8 kinds of cages. All cage modelswere 30-mm long, 10-mm wide, and 7-mm high. The PEEK cagewas made of PEEK, and other cages were made of TI alloy(Ti6Al4V). The bilateral pedicle screw system was modeled basedon the EXPEDIUM 5.5 System (DePuy Synthes Spine, Inc., Rayn-ham, Massachusetts, USA). The diameter of the pedicle screw was5.5 mm. The material of the pedicle screws was titanium alloy(Ti6Al4V).Figure 2 shows the mechanical test for the porous materials

with different porosities. The static compression testing wascarried out according to the ISO 13314:2011 standard by using auniversal mechanical testing machine.21-24 The machine for thecompression testing was the Instron 8874 (Naton, Beijing, China).The average pore size of all the test samples was 350e400 um. Theporosities of the 3 groups of test samples (5 per group) were 65%,75%, and 80%. The test results were as follows: for the sampleswith 65% porosity, the average Young modulus was 2653 MPa andthe average plateau stress was 57.2 MPa; for the samples with 75%porosity, the average Young modulus was 1551 MPa and theaverage plateau stress was 30 MPa; for the samples with 80%porosity, the average Young modulus was 675 MPa and the averageplateau stress was 19.1 MPa. The test results were comparable withthe previous literature.20 The Young modulus and the plateaustress decreased with increasing porosity. The materialproperties of the components are shown in Table 1.25-33

To validate the intact FE model, the analysis study included 2steps of simulation. The predicted results were compared with theprevious experimental data. The contact surfaces of the vertebraeand discs were set as tie constraints. The contact between the facetjoints was simulated as frictionless surfaces.25-29,34 First, the ROMof the intact lumbar spine L1-5 under pure moment was predicted.Three different moments (2.5, 5.0, and 7.5 Nm) were applied tothe upper surface of L1 while the bottom of L5 was fixed. TheROM of the lumbar spine was compared with previous in vitro

UROSURGERY, https://doi.org/10.1016/j.wneu.2017.12.127

Figure 1. Finite element model of the intact lumbar spine. L, left; R, right.

ORIGINAL ARTICLE

ZHENJUN ZHANG ET AL. BIOMECHANICAL ANALYSIS OF POROUS LUMBAR FUSION CAGES

results.28,34,35 Then the compression displacement and interver-tebral disc pressure (IDP) of the motion segment L4-5 under purecompression were calculated. The upper surface of L4 was loadedwith 4 preload values (100, 200, 300, and 400 N) as described byBerkson et al.36 The compression displacement and IDP of L4-5were compared with previous results.28,36,37

For simulation of the surgical models, the segment L2-5 waschosen to predict the biomechanics changes of the surgical level.The surgical conditions were as follows: the interbody cage was

Figure 2. Mechanical test for the porous materials withdifferent porosities: (A) test samples with different

WORLD NEUROSURGERY 111: e581-e591, MARCH 2018

inserted at the L3-4 disc space laterally, and supplemented withbilateral pedicle screw fixation. The FE models of interbody fusionwith various cages are shown in Figure 3. All the surgical FEmodels were constructed based on the validated intact model.The surface contact between the vertebrae and discs, and thecontact between the facet joints, were consistent with that of thevalidated model. The interfaces of the vertebrae and cages werealso assigned to tie constraints. The bottom of L5 was fixed inall directions. The compressive load of 280 N and the moment

porosities, (B) mechanical testing machine, and (C) testresults for each group of samples.

www.WORLDNEUROSURGERY.org e583

Table 1. Material Properties Used in the Finite Element Models

ComponentsYoung

Modulus (MPa)PoissonRatio

Cross-SectionalArea (mm2) Reference(s)

Cortical bone 12,000 0.3 — Shirazi-Adl et al., 198625; Zhong et al., 200626

Cancellous bone 100 0.2 — Shirazi-Adl et al., 198625; Zhong et al., 200626

Posterior bone 3500 0.25 — Shirazi-Adl et al., 198625; Zhong et al., 200626

End plate 4000 0.3 — Schmidt et al., 200727

Annulus ground 4.2 0.45 — Shirazi-Adl et al., 198625; Zhong et al., 200626

Nucleus pulposus 1 0.49 — Dreischarf et al., 201428; Ayturk and Puttlitz, 201129

Anterior longitudinal ligament 20 0.3 63.7 Zhong et al., 200626; Faizan et al., 201430

Posterior longitudinal ligament 20 0.3 20 Zhong et al., 200626; Faizan et al., 201430

Ligamentum flavum 19.5 0.3 40 Zhong et al., 200626; Faizan et al., 201430

Interspinous ligament 11.6 0.3 40 Zhong et al., 200626; Faizan et al., 201430

Supraspinous ligament 15 0.3 30 Zhong et al., 200626; Faizan et al., 201430

Transverse ligament 58.7 0.3 3.6 Zhong et al., 200626; Faizan et al., 201430

Capsular ligament 32.9 0.3 60 Zhong et al., 200626; Faizan et al., 201430

Cage (PEEK) 3500 0.3 — Xiao et al., 201231; Vadapalli et al., 200632

Cage (TI) 110,000 0.3 — Xiao et al., 201231; Chosa et al., 200433

Cage (porosity 65%) 2653 0.3 — Mechanical test

Cage (porosity 75%) 1551 0.3 — Mechanical test

Cage (porosity 80%) 675 0.3 — Mechanical test

Pedicle screws (TI alloy) 110,000 0.3 — Xiao et al., 201231; Chosa et al., 200433

PEEK, polyetheretherketone; TI, titanium.

ORIGINAL ARTICLE

ZHENJUN ZHANG ET AL. BIOMECHANICAL ANALYSIS OF POROUS LUMBAR FUSION CAGES

of 7.5 Nm were applied to the upper surface of L2 as in previousliterature.38,39 The compressive load of 280 N corresponded tothe partial weight of a human body, and the moment of 7.5 Nmsimulated the motion modes occurring in different conditionssuch as flexion, extension, lateral bending, and axial rotation.Considering the symmetry of the sagittal plane, this study simu-lated the biomechanical properties of surgical FE models in 4motion modes: flexion, extension, bending left, and rotation left.The ROM, cage stress, end plate stress, and FJF were analyzed andexported. The predicted results of porous cages were comparedwith that of solid cages. The ROM data were normalized to theintact ROM data. Under combined loading, the intact L2-5 modelwas recalculated. In total, 36 simulation calculations for 9 modelsand 4 motion modes were performed. Simulation results were inaccordance with the requirements of visualization, and mechanicsdata were expressed using von Mises stress contours.

RESULTS

Model ValidationUnder the pure moment of 7.5 Nm, the L1-5 ROM was within therange of the previous FE and in vitro experimental studies.28,35 Thetotal movement angles were 24.75� in flexion-extension, 26.66� inlateral bending, and 18.43� in axial rotation. As shown in

e584 www.SCIENCEDIRECT.com WORLD NE

Figure 4A, all values in different conditions were within the FE andin vitro ranges. The load-deflection curves are shown in Figure 4B,which were comparable with the existing results of previousstudies.28,34,35 The rotation angles were 15.04� in flexion, 9.71�

in extension, 13.45� in bending left, and 10.33� in rotation left.The compression-displacement curves are shown in Figure 4C,

which indicate that the axial displacement of the L4-5 segmentincreased almost linearly with the applied axial compressiveloading. The results were comparable with the previous in vitroexperimental study.36 The compression-IDP curves are displayedin Figure 4D, which are also compared with the previous FE andin vitro experimental studies.28,37 The predicted results demon-strated that IDP of the L4-5 segment increased almost linearly withthe applied axial compressive loading.

ROMUnder the combined loading of 280 N and 7.5 Nm, the ROM ofsurgical models is shown in Figure 5. After inserting the interbodycage, the predicted ROMs for all surgical models decreased by>90% compared with the intact case. ROMs for the 3 PPT cageswere slightly more than that for the TI cage, and the porosity ofthe PPT cage did not substantially alter the ROM in all motionmodes. ROMs for the 3 FPT cages were more than that for thePEEK cage, and they increased with increasing porosity.

UROSURGERY, https://doi.org/10.1016/j.wneu.2017.12.127

Figure 3. Finite element models of lateral lumbarinterbody fusion with various cages. The dimensions ofthe solid titanium cage, solid polyetheretherketone

cage, and fully porous titanium cages are the same asthe partially porous titanium cages.

ORIGINAL ARTICLE

ZHENJUN ZHANG ET AL. BIOMECHANICAL ANALYSIS OF POROUS LUMBAR FUSION CAGES

Comparing all the surgical models, the ROM for the TI cage wasthe minimum and the ROM for the FPT cage with 80% porositywas the maximum in all motion modes. Compared with thePEEK cage, the ROMs for the FPT cage with 80% porosityincreased by 2.55% in flexion, 3.43% in extension, 2.62% inbending left, and 2.22% in rotation left.

Cage StressThe maximum stress in cage (cage stress) is displayed inFigure 6A. Cage stresses for the 3 PPT cages were more than thatfor the TI cage, and the porosity of the PPT cage did notsubstantially alter cage stress in all motion modes. Cage stressesfor the 3 FPT cages were less than that for the PEEK cage, andthey decreased with increasing porosity. Comparing all thesurgical models, cage stress for the PPT cage with 80% porositywas the maximum and cage stress for the FPT cage with 80%porosity was the minimum in all motion modes. Compared withthe PEEK cage, cage stress for the FPT cage with 80% porositywas reduced by 36.18% in flexion, 49.55% in extension, 33.84%in bending left, and 28.62% in rotation left. To compare thestress distribution of the 2 porous structures, the contour plotsof von Mises stress in the PPT cage with 75% porosity and FPTcage with 75% porosity are shown in Figure 6B. The PPT cagemay result in a stress shielding effect because of the

WORLD NEUROSURGERY 111: e581-e591, MARCH 2018

configuration of both solid and porous structures. In contrast,the stress for the FPT cage was more evenly distributed in thecage in all motion modes because of the fully porous structure.

End Plate StressFigure 7A depicts the maximum stress in the L3 bottom end plate(end plate stress). After interbody fusion, the predicted end platestress increased in all motion modes except for the FPT cage with75% porosity and FPT cage with 80% porosity in rotation left. Endplate stresses for the 3 PPT cages were more than that for the TIcage, and the porosity of the PPT cage did not substantially alterend plate stress in all motion modes. End plate stresses for the3 FPT cages were less than that for the PEEK cage, and theydecreased with increasing porosity. Comparing all the surgicalmodels, end plate stress for the PPT cage with 80% porositywas the maximum and end plate stress for the FPT cage with80% porosity was the minimum in all motion modes. Comparedwith the PEEK cage, end plate stress for the FPT cage with 80%porosity was reduced by 37.63% in flexion, 52.40% in extension,38.51% in bending left, and 33.12% in rotation left. Comparedwith the intact case, end plate stress for the FPT cage with 80%porosity remained the same in flexion, increased by 15.75% inextension, increased by 54.48% in bending left, and decreasedby 26.26% in rotation left. The contour plots of von Mises stress

www.WORLDNEUROSURGERY.org e585

Figure 4. Predicted total range of motion (A) and load-deflection curves (B)under pure moments, and L4-5 compression-displacement curves (C) and

intervertebral disc pressure (D) under axial compression. Deg, degrees; FE,finite element; IDP, intervertebral disc pressure; ROM, range of motion.

Figure 5. Range of motion at the surgical level forvarious cages in 4 motion modes. FPT65, fully poroustitanium cage with 65% porosity; FPT75, fully poroustitanium cage with 75% porosity; FPT80, fully poroustitanium cage with 80% porosity; L, left; PEEK, solid

polyetheretherketone; PPT65, partially porous titaniumcage with 65% porosity; PPT75, partially poroustitanium cage with 75% porosity; PPT80, partiallyporous titanium cage with 80% porosity; R, right;ROM, range of motion; TI, solid titanium.

e586 www.SCIENCEDIRECT.com WORLD NEUROSURGERY, https://doi.org/10.1016/j.wneu.2017.12.127

ORIGINAL ARTICLE

ZHENJUN ZHANG ET AL. BIOMECHANICAL ANALYSIS OF POROUS LUMBAR FUSION CAGES

Figure 6. Maximum stress in the cage for various cagesin 4 motion modes (A) and contour plots of von Misesstress in porous cages (B). FPT65, fully porous titaniumcage with 65% porosity; FPT75, fully porous titaniumcage with 75% porosity; FPT80, fully porous titaniumcage with 80% porosity; L, left; Max, maximum; PEEK,

solid polyetheretherketone; PPT65, partially poroustitanium cage with 65% porosity; PPT75, partiallyporous titanium cage with 75% porosity; PPT80,partially porous titanium cage with 80% porosity; R,right; TI, solid titanium.

ORIGINAL ARTICLE

ZHENJUN ZHANG ET AL. BIOMECHANICAL ANALYSIS OF POROUS LUMBAR FUSION CAGES

in the L3 bottom end plate for the PPT cage with 75% porosity andFPT cage with 75% porosity are shown in Figure 7B. The PPT cagemay result in a stress shielding effect, whereas the end plate stressfor the FPT cage was more evenly distributed in the end plate in allmotion modes.

FJFIn Figure 8, the FJF of intact and surgical models is displayed.After interbody fusions, FJF at surgical level L3-4 decreased sub-stantially in all motion modes except for flexion. Comparing allthe surgical models, FJF for various cages was not substantiallychanged in all motion modes except for rotation. Compared withthe TI cage, FJF for the PPT cage was not substantially changed inall motion modes. Compared with the PEEK cage, FJF for the FPTcage with 80% porosity was reduced by 7.57% in flexion, 11.90% in

WORLD NEUROSURGERY 111: e581-e591, MARCH 2018

extension, and 13.13% in bending left, whereas it increased by19.24% in rotation left.

DISCUSSION

We used an FE model to study the biomechanical effect of variouscages on ROM, stresses in cage and end plate, and FJF in 4loading modes (flexion, extension, bending left, and rotation left).Such an FE investigation is not feasible in a cadaver modelbecause of the inability to estimate the stresses in various spinalstructures.32 The predicted ROMs with various cages werecomparable with the previous studies.4,5,10,40 In addition, thecurrent study showed the cage stress, end plate stress, and FJF. Asshown in Figure 5, the ROM data were normalized to the intactROM data. The predicted ROMs for all surgical models were<10% of the intact model in all motion modes. That is the

www.WORLDNEUROSURGERY.org e587

Figure 7. Maximum stress in the L3 bottom end platefor various cages in 4 motion modes (A) and contourplots of von Mises stress in end plate for various cages(B). FPT65, fully porous titanium cage with 65%porosity; FPT75, fully porous titanium cage with 75%porosity; FPT80, fully porous titanium cage with 80%

porosity; L, left; Max, maximum; PEEK, solidpolyetheretherketone; PPT65, partially porous titaniumcage with 65% porosity; PPT75, partially poroustitanium cage with 75% porosity; PPT80, partiallyporous titanium cage with 80% porosity; R, right; TI,solid titanium.

ORIGINAL ARTICLE

ZHENJUN ZHANG ET AL. BIOMECHANICAL ANALYSIS OF POROUS LUMBAR FUSION CAGES

maximum stiffness or stability that the cage may provide for thelumbar spine. However, the ROM tends to increase further withincreasing porosity because of the lower stiffness and greaterdeformation of cages. As shown in Figures 6 and 7, stresses incage and end plate were sensitive to the various cages.Compared with the TI and PPT cages, the PEEK and FPT cagessubstantially reduced the maximum stresses in cage and endplate in all motion modes. Compared with the PEEK cages, thestresses in cage and end plate for the FPT cages decreased, andthe stresses decreased with increasing porosity of cages. Thevariations of the maximum stresses in the L3 bottom end plateafter interbody fusion are listed in Table 2. The end platestresses were normalized to the intact stress data. Comparedwith the intact model, the end plate stress for the PPT cage with

e588 www.SCIENCEDIRECT.com WORLD NE

80% porosity increased by <2-fold in all motion modes. Asshown in Figure 8, the FJF at the surgical level decreasedsubstantially compared with the intact conditions after insertingthe cage at L3-4 in all motion modes except for flexion. FJF wasnot substantially changed with various cages except for rotationleft.ROM, cage stress, and end plate stress at the surgical level were

directly affected in all motion modes and changed with variouscages. By comparing the biomechanics of interbody fusion usingdifferent cages, it was shown that the FPT cage displayed someadvantages, such as minimum cage stress and minimum end platestress, in all motion modes. The FPT cage reduced the cage stressand end plate stress, which may decrease the risk of subsidence ofthe cages into the end plate and the adjoining vertebral bone over

UROSURGERY, https://doi.org/10.1016/j.wneu.2017.12.127

Figure 8. Facet joint force at the surgical level for various cages in 4 motionmodes. FJF, facet joint force; FPT65, fully porous titanium cage with 65%porosity; FPT75, fully porous titanium cage with 75% porosity; FPT80, fullyporous titanium cage with 80% porosity; L, left; PEEK, solid

polyetheretherketone; PPT65, partially porous titanium cage with 65%porosity; PPT75, partially porous titanium cage with 75% porosity; PPT80,partially porous titanium cage with 80% porosity; R, right; TI, solid titanium.

ORIGINAL ARTICLE

ZHENJUN ZHANG ET AL. BIOMECHANICAL ANALYSIS OF POROUS LUMBAR FUSION CAGES

time.32 It was indicated that in performing lateral lumbarinterbody fusion, the FPT cage may offer an alternative to thePEEK cage.

Table 2. Variation of the Maximum Stress in the L3 Bottom EndPlate After Interbody Fusion with Respect to Intact Model

End PlateStress Flexion Extension Bending-L Rotation-L

Intact 1.00 1.00 1.00 1.00

TI 2.36 4.59 3.19 1.65*

PPT65 2.45 4.75 3.26 1.79*

PPT75 2.46 4.77 3.26 1.80*

PPT80 2.48 4.79 3.27 1.82*

PEEK 1.60* 2.43 2.51 1.10*

FPT65 1.49* 2.22 2.37 1.02*

FPT75 1.30* 1.79* 2.06 0.87*

FPT80 1.00* 1.16* 1.54* 0.74*

L, left; TI, solid titanium; PPT65, partially porous titanium cage with 65% porosity; PPT75,partially porous titanium cage with 75% porosity; PPT80, partially porous titaniumcage with 80% porosity; PEEK, solid polyetheretherketone; FPT65, fully porous tita-nium cage with 65% porosity; FPT75, fully porous titanium cage with 75% porosity;FPT80, fully porous titanium cage with 80% porosity.

*End plate stress increased by <2-fold (The end plate stresses were normalized to theintact stress data).

WORLD NEUROSURGERY 111: e581-e591, MARCH 2018

In the current study, the porous cages (PPT and FPT) werecompared with solid cages (TI and PEEK). The Young modulus ofsolid TI was much larger than that of the porous TI. As waspredicted in Figures 6 and 7, the PPT cage may result in a stressshielding effect because of the configuration of both solid andporous structures. The FPT cage has advantages in cage stressand end plate stress, which may lead to a smaller risk ofsubsidence.32 In addition, the FPT cage may lead to betterfusion healing because the porous structure was beneficial tobone ingrowth.16 According to Figure 5, although ROMs for allthe cages decreased by >90% compared with the intact case,the ROM for the FPT cage increased with increasing porosity.Higher porosity may result in greater bone ingrowth butdiminish mechanical properties.41 Further increased ROM mayreduce the stability to affect the healing response of the fusion.4

Clinically, the cages with higher porosity may have better boneingrowth into the cage structure and an enhanced bone fusionhealing. The improved bone healing could be a reasonableexchange for increased ROM of porous cages. In summary, theporous AM cage may offer an alternative to PEEK cages inlateral lumbar interbody fusion, whereas it may be prudent tofurther increase the porosity of the cage.There are some limitations in this study, such as validating the

model by comparing it with other models, using a unique lumbarmodel, simplifying the material properties of some tissues, andignoring the role of muscles. Another limitation is that the vali-dation of the model is done by comparing it with other models,and not to a biologic model. This intact spine model was scannedfrom a healthy volunteer, and was validated by comparing it withthe other FE models and in vitro models. The experimental studies

www.WORLDNEUROSURGERY.org e589

ORIGINAL ARTICLE

ZHENJUN ZHANG ET AL. BIOMECHANICAL ANALYSIS OF POROUS LUMBAR FUSION CAGES

on a biologic model are useful to further validate the current FEmodel. Another limitation is the FE analysis standard technique ofusing a unique lumbar model. The geometric model of the lumbarspine varies from person to person, such as the intervertebral discspace and the gaps between facet joints. However, only one modelof the lumbar spine was chosen in this study. Furthermore, thematerial properties of human tissues are poorly understood. In thepresent model, the material properties were simplified as linearelastic although the components of the lumbar spine are nonlinearin reality. However, many FE analyses on the lumbar spine haveassumed that the components of the spine are linear to improvethe calculation efficiency.26,32,39,42,43 In addition, the muscles werenot considered in this study although the muscles play animportant role in supporting the stability of the lumbar spine.

e590 www.SCIENCEDIRECT.com WORLD NE

However, the tendency of the predicted results with various cageswould not be substantially changed depending on the individualgeometric model, simplified material properties, and model of themuscles.In conclusion, the predicted results showed that the structure of

a porous cage can affect the biomechanics of lumbar interbodyfusion noticeably. Comparing the surgical models with differentcages, a porous AM cage showed advantages in cage stress andend plate stress. Compared with the PEEK cage, the porous cagehas some advantages in biomechanics and may lead to betterfusion healing in clinical practice. Further clinical studies on theeffect of porous cages on stability, subsidence, and other clinicallyrelevant parameters are necessary to validate the observations ofthis FE study.

REFERENCES

1. Oxland TR, Lund T. Biomechanics of stand-alonecages and cages in combination with posteriorfixation: a literature review. Eur Spine J. 2000;9:S095-S101.

2. Phillips FM, Cunningham B, Carandang G,Ghanayem AJ, Voronov L, Havey RM, et al. Effectof supplemental translaminar facet screw fixationon the stability of stand-alone anterior lumbarinterbody fusion cages under physiologiccompressive preloads. Spine. 2004;29:1731-1736.

3. Tsantrizos A, Andreou A, Aebi M, Steffen T.Biomechanical stability of five stand-alone ante-rior lumbar interbody fusion constructs. Eur SpineJ. 2000;9:14-22.

4. Fogel GR, Parikh RD, Ryu SI, Turner AW.Biomechanics of lateral lumbar interbody fusionconstructs with lateral and posterior plate fixa-tion: laboratory investigation. J Neurosurg Spine.2014;20:1-7.

5. Cappuccino A, Cornwall GB, Turner AW,Fogel GR, Duong HT, Kim KD, et al. Biome-chanical analysis and review of lateral lumbarfusion constructs. Spine. 2010;35:S361-S367.

6. Laws CJ, Coughlin DG, Lotz JC, Laws CJ,Coughlin DG, Lotz JC, et al. Direct lateralapproach to lumbar fusion is a biomechanicallyequivalent alternative to the anterior approach: anin vitro study. Spine. 2012;37:819-825.

7. Perez-Orribo L, Kalb S, Reyes PM, Chang SW,Crawford NR. Biomechanics of lumbar corticalscrew-rod fixation versus pedicle screw-rod fixa-tion with and without interbody support. Spine.2013;38:635-641.

8. Pimenta L, Turner AW, Dooley ZA, Parikh RD,Peterson MD. Biomechanics of lateral interbodyspacers: going wider for going stiffer. Scienti-ficWorldJournal. 2012;2012:381814.

9. Kornblum MB, Turner AW, Cornwall GB,Zatushevsky MA, Phillips FM. Biomechanicalevaluation of stand-alone lumbar polyether-ether-ketone interbody cage with integrated screws.Spine J. 2013;13:77-84.

10. Slucky AV, Brodke DS, Bachus KN, Droge JA,Braun JT. Less invasive posterior fixation method

following transforaminal lumbar interbody fusion:a biomechanical analysis. Spine J. 2006;6:78-85.

11. Ambati DV, Wright EK, Lehman RA, Kang DG,Wagner SC, Dmitriev AE. Bilateral pedicle screwfixation provides superior biomechanical stabilityin transforaminal lumbar interbody fusion: a finiteelement study. Spine J. 2015;15:1812-1822.

12. Cole CD, McCall TD, Schmidt MH, Dailey AT.Comparison of low back fusion techniques:transforaminal lumbar interbody fusion (TLIF) orposterior lumbar interbody fusion (PLIF) ap-proaches. Curr Rev Musculoskelet Med. 2009;2:118-126.

13. Yang SD, Chen Q, Ding WY, Zhao JQ, Zhang YZ,Shen Y, et al. Unilateral pedicle screw fixationwith bone graft vs. bilateral pedicle screw fixationwith bone graft or cage: a comparative study. MedSci Monit. 2016;22:890-897.

14. Kim DH, Jeong ST, Lee SS. Posterior lumbarinterbody fusion using a unilateral single cage anda local morselized bone graft in the degenerativelumbar spine. Clin Orthop Surg. 2009;1:214-221.

15. Melnyk AD, Wen TL, Kingwell SF, Chak JD,Singh VM, Cripton PA, et al. Load transfer char-acteristics between posterior spinal implants andthe lumbar spine under anterior shear loading: anin vitro investigation. Spine. 2012;37:E1126-E1133.

16. Jones AC, Arns CH, Sheppard AP,Hutmacher DW, Milthorpe BK, Knackstedt MA.Assessment of bone ingrowth into porous bio-materials using MICRO-CT. Biomaterials. 2007;28:2491-2504.

17. Lee YH, Chung CJ, Wang CW, Peng YT,Chang CH, Chen CH, et al. Computational com-parison of three posterior lumbar interbody fusiontechniques by using porous titanium interbodycages with 50% porosity. Comput Biol Med. 2016;71:35-45.

18. Kang H, Hollister SJ, Marca FL, Park P, Lin CY.Porous biodegradable lumbar interbody fusioncage design and fabrication using integratedglobal local topology optimization with laser sin-tering. J Biomech Eng. 2013;135:101013.

19. Tsai PI, Hsu CC, Chen SY. Biomechanical inves-tigation into the structural design of porous ad-ditive manufactured cages using numerical and

UROSURGERY, http

experimental approaches. Comput Biol Med. 2016;76:14-23.

20. Lewis G. Properties of open-cell porous metalsand alloys for orthopaedic applications. J Mater SciMater Med. 2013;24:2293-2325.

21. Wauthle R, Stok JV, Yavari SA. Additively manu-factured porous tantalum implants. Acta Biomater.2015;14:217.

22. Campoli G, Borleffs MS, Yavari SA. Mechanicalproperties of open-cell metallic biomaterialsmanufactured using additive manufacturing. MaterDes. 2013;49:957-965.

23. Wang L, Kang JF, Sun CN, Li DC, Cao Y, Jin ZM.Mapping porous microstructures to yield desiredmechanical properties for application in 3D prin-ted bone scaffolds and orthopaedic implants.Mater Des. 2017;133:62-68.

24. International Organization for Standardization.ISO 13314:2011. Mechanical testing of metals eDuctility testing e Compression test for porousand cellular metals. Available at: https://www.iso.org/standard/53669.html. Accessed September 19,2017.

25. Shirazi-Adl A, Ahmed AM, Shrivastava SC. Me-chanical response of a lumbar motion segment inaxial torque alone and combined with compres-sion. Spine. 1986;11:914-927.

26. Zhong ZC, Wei SH, Wang JP, Feng CK, Chen CS,Yu CH. Finite element analysis of the lumbarspine with a new cage using a topology optimi-zation method. Med Eng Phys. 2006;28:90-98.

27. Schmidt H, Heuer F, Drumm J, Klezl Z, Claes L,Wilke HJ. Application of a calibration methodprovides more realistic results for a finite elementmodel of a lumbar spinal segment. Clin Biomech.2007;22:377-384.

28. Dreischarf M, Zander T, Shirazi-Adl A,Puttlitz CM, Adam CJ, Chen CS, et al. Comparisonof eight published static finite element models ofthe intact lumbar spine: predictive power ofmodels improves when combined together.J Biomech. 2014;47:1757-1766.

29. Ayturk UM, Puttlitz CM. Parametric convergencesensitivity and validation of a finite element modelof the human lumbar spine. Comput Method Bio-mech. 2011;14:695-705.

s://doi.org/10.1016/j.wneu.2017.12.127

ORIGINAL ARTICLE

ZHENJUN ZHANG ET AL. BIOMECHANICAL ANALYSIS OF POROUS LUMBAR FUSION CAGES

30. Faizan A, Kiapour A, Kiapour AM, Goel VK.Biomechanical analysis of various footprints oftransforaminal lumbar interbody fusion devices.J Spinal Disord Tech. 2014;27:E118-E127.

31. Xiao ZT, Wang LY, Gong H, Zhu D. Biomechan-ical evaluation of three surgical scenarios of pos-terior lumbar interbody fusion by finite elementanalysis. Biomed Eng Online. 2012;11:31.

32. Vadapalli S, Sairyo K, Goel VK, Robon M,Biyani A, Khandha A, et al. Biomechanical ratio-nale for using polyetheretherketone (PEEK)spacers for lumbar interbody fusion-a finiteelement study. Spine. 2006;31:992-998.

33. Chosa E, Goto K, Totoribe K, Tajima N. Analysisof the effect of lumbar spine fusion on the supe-rior adjacent intervertebral disk in the presence ofdisk degeneration, using the three-dimensionalfinite element method. J Spinal Disord Tech. 2004;17:134-139.

34. Shirazi-Adl A. Biomechanics of the lumbar spinein sagittal/lateral moments. Spine. 1994;19:2407-2414.

35. Schmidt H, Galbusera F, Rohlmann A, Zander T,Wilke HJ. Effect of multilevel lumbar disc arthro-plasty on spine kinematics and facet joint loads inflexion and extension: a finite element analysis.Eur Spine J. 2012;21:S663-S674.

WORLD NEUROSURGERY 111: e581-e59

36. Berkson MH, Nachemson A, Schultz AB. Me-chanical properties of human lumbar spine mo-tion segments—Part II: responses in compressionand shear; influence of gross morphology.J Biomech Eng. 1979;101:53-57.

37. Brinckmann P, Grootenboer H. Change of discheight, radial disc bulge, and intradiscal pressurefrom discectomy. An in vitro investigation onhuman lumbar discs. Spine. 1991;16:641-646.

38. Rohlmann A, Neller S, Claes L, Bergmann G,Wilke HJ. Influence of a follower load on intradiscalpressure and intersegmental rotation of the lumbarspine. Spine. 2001;26:E557-E561.

39. Choi J, Shin D, Kim S. Biomechanical effects ofthe geometry of ball-and-socket artificial disc onlumbar spine: a finite element study. Spine. 2017;42:E332-E339.

40. Beaubien BP, Derincek A, Lew WD, Wood KB.In vitro, biomechanical comparison of an anteriorlumbar interbody fusion with an anteriorly placed,low-profile lumbar plate and posteriorly placedpedicle screws or translaminar screws. Spine.2005;30:1846-1851.

41. Yan N, Hou TS. Porosity of biomaterials and boneingrowth. Chin J Biomed Eng. 2008;27:611-614.

1, MARCH 2018 www

42. Kim KT, Lee SH, Suk KS, Lee JH, Jeong B.Biomechanical changes of the lumbar segmentafter total disc replacement: charite(r), prodisc(r)and maverick(r) using finite element model study.J Korean Neurosurg Soc. 2010;47:446-453.

43. Grauer JN, Biyani A, Faizan A, Kiapour A,Sairyo K, Ivanov A, et al. Biomechanics of two-level Charité artificial disc placement in compar-ison to fusion plus single-level disc placementcombination. Spine J. 2006;6:659-666.

Conflict of interest statement: This study was supported bythe Industry Public Technology Service Platform Project ofShenzhen (SMJKPT20140417010001) and the Science andTechnology Plan Basic Research Project of Shenzhen(JCYJ20151030160526024).

Received 13 November 2017; accepted 19 December 2017

Citation: World Neurosurg. (2018) 111:e581-e591.https://doi.org/10.1016/j.wneu.2017.12.127

Journal homepage: www.WORLDNEUROSURGERY.org

Available online: www.sciencedirect.com

1878-8750/$ - see front matter ª 2017 Elsevier Inc. Allrights reserved.

.WORLDNEUROSURGERY.org e591