New protocol for in-house management of computer assisted orthognathic surgery

BritishJournalofOralandMaxillofacialSurgery58(2020)e265–e271

Availableonlineatwww.sciencedirect.com

ScienceDirect

New

protocol

for

in-house

management

of

computer

assisted

orthognathic

surgery

Giacomo

De

Riu

,

Luigi

Angelo

Vaira

,

Enrica

Ligas

,

Veronica

Vaittinen

,

Giovanni

Spano

,

Giovanni

Salzano

,

Pasquale

Piombino

a_{UniversityofSassariHospital,MaxillofacialSurgeryOperativeUnit,VialeSanPietro43B,07100,Sassari,Italy} b_{UniversityofSassariHospital,DentalSchool,VialeSanPietro43B,07100,Sassari,Italy}

c_{UniversityofNaples“FedericoII”Hospital,MaxillofacialSurgeryOperativeUnit,ViaPansini5,80131,Napoli,Italy}

Accepted16July2020 Availableonline25July2020

Abstract

Theaimofthisstudywastoevaluateretrospectivelytheaccuracyofaprotocolforcompletelyin-house,computer-assisted,orthognathic surgery,generatingresinprintedintermediatesurgicalguides.Aretrospective,observationalstudywasmadeonacohortof15patientstreated consecutivelyfromSeptember2017toMay2019,whounderwentbimaxillaryorthognathicsurgeryplannedwiththesame3-dimensional programandwhosesurgicalintermediatesplintswereobtainedwiththesameall-in-houseprotocol.Virtualplannedsurgicalmovements werecomparedwiththerealsurgicaloutcome.Thedifferenceswerenotsignificantforeightofthe12variablesconsidered.Thepvalues, calculatedwiththeWilcoxonsignedranktest,wereevenlydistributedandrangedfromp=0.001top=0.820.Significantdifferenceswere reportedinfourmeasurements:anglebetweensella-nasionplaneandalineconnectingtheincisaledgeandtheapexoftherootofthemost prominentincisor(U1-SN)(p=0.001),anglebetweenFrankfortplaneandalineconnectingtheincisaledgeandtherootapexofthemost prominentupperincisor(p=0.008),dentalmidlinediscrepancies(p=0.006),andocclusalplanetilt(U1-FH)(p=0.001),basicallydueto intraoperativesettings.The3-dimensionalresinprintedsurgicalguideswereshowntobeareliablealternativetothecommercialonesand showedhighrateofaccuracyformostofthevariablesassessed.Fouroutof12oftheseshowedsignificanterrors,buttwoofthemwereonly minimaldiscrepancieswithnoclinicalimplications.

©2020TheBritishAssociationofOralandMaxillofacialSurgeons.PublishedbyElsevierLtd.Allrightsreserved.

Keywords:Orthognathicsurgery;Virtualsurgicalplanning;Computer-assistedorthognathicsurgery;Virtualplanningaccuracy

Introduction

Intraoperative surgical guides are crucial to achieve pre-dictable results in orthognathic surgery, especially when bimaxillary complex movements are made in the three planes.1Classicorthognathicsurgicalplanningrequiresthe combinationsof the data obtained from different sources, suchasaestheticexaminationoftheface,frontalandlateral

∗_{Correspondingauthorat:VialeSanPietro43B,07100,Sassari,Italy.}

Tel.:+393401846168;fax:+39079229002.

E-mailaddress:luigi.vaira@gmail.com(L.A.Vaira).

1 _{Co-firstauthor}

cephalometricanalysis,andplastercastsmountedona semi-adjustablearticulatorwithfacebowregistration.2Thisallows thesurgeontoplanandsimulatethemovementsthatthebony segmentswillundergoduringsurgery.Unfortunately,the out-comeoforthognathicsurgerymaydiffersignificantlyfrom theplanning.3–7

In thelastdecade,advances inimagingtechnologyand 3-dimensional computerised planning have improved the precision of surgical corrections of complex dentofacial deformities. These advances in 3-dimensional technology haveresultedinnewcomputerisedtoolsforusein preopera-tiveplanningandthemanufactureofsurgicalsplints.8–12In theseprotocols,computerisedcompositeskullmodelsofthe

https://doi.org/10.1016/j.bjoms.2020.07.022

e266 G.DeRiuetal./BritishJournalofOralandMaxillofacialSurgery58(2020)e265–e271

patientaregeneratedtoaccuratelyrepresenttheskeleton,the dentition,andthefacialsofttissue,10,11,13,14Themovefrom 2-dimensionalto3-dimensionalimagingprovidessurgeons andpatientswithextrainformationthatcannotbeobtained fromcephalogramsalone.Thesoftwareprogramsenablethe surgeontointeractwiththe3-dimensionalimages:itis possi-bletosimulateseveraldifferentsurgicalplans,includingLe FortIosteotomy,mandibularramusosteotomies,and genio-plasties.Itisalsopossibletovirtuallyrepositionosteotomised bonystructures,controllingintercuspationandthe interfer-ences among osteotomised bonystructures and regions at thebaseoftheskull,thereforesimulatingthepostoperative resultsonhardandsofttissueinthreedimensionsona com-puterscreen.Also,theseprogramsmakeitpossibletoobtain intermediatesurgicalsplintsusingCAD/CAMtechnologyto performcomputer-assistedorthognathicsurgery,whichhas proved tobe areliable,innovative,andprecise method to transfertheorthognathicplantotheactualsurgery.Using dig-italplanning,thereisslightchance ofincorporatingerrors, andthefewthatpersistaregenerallyduetoimproper acqui-sitionofcomputedtomographic(CT)scans,oftenlinkedto poorpatientcompliance.

TheaccuracyoftheCAD/CAMsurgicalguidestoperform orthognathicsurgeryrapidlygained widespreadpopularity andmanystudiesconfirmed theease andaccuracyof this method.8,13,15–19

Ontheotherhand,mostof thesoftwarecurrently avail-ableonthemarketstillrequireswebconferencecollaboration withbioengineers,oftenindifferentcountries,tohelpinthe planning phases, andneed for assistance of outer compa-nies to finalise the planning process andmanufacture the surgicalsplints.Theseinconveniencesaretime-consuming, expensive,andlimittheautonomyoftheoperators.

Wepresent acompletely in-house and low-budget pro-tocol for surgical planning and splint manufacturing. It is possible withcommercially available programs that allow thesurgeontocompletebyhimselfallthestepsofthe vir-tualplanninguptothegenerationofasurgicalguidethatis directlyprintedwithanin-office3-dimensionalresinprinter, alreadyadoptedinprostheticdentistryandjewellery.Such kindofprotocol,whichprovideshomemademanufacturing ofthesurgicalsplint,hasnotyetbeenwidelypublished.20

Theaimofthisstudywastoretrospectivelyevaluatethe accuracy of 3-dimensional software for computer-assisted orthognathicsurgerygeneratingin-houseresinprinted inter-mediatesurgicalguidesinasampleof15subjectswhohad undergonebimaxillaryorthognathicsurgery.

Materialandmethods

A retrospective, observational study was performed on a groupofpatientstreatedconsecutivelyfromSeptember2017 toMay2019,whounderwentorthognathicsurgery,planned withthesame3-dimensionalprogram,andwhosesurgical splintswere obtained withthe sameall-in-house protocol.

Forinclusion,thepatientshadtohaveundergone bimaxil-laryorthognathicsurgeryforabnormalgrowthofthejaws. Thepatientshadpreviouslyhadorthodonticremovalofdental compensationsandwereretainedwithpassivearchwires.

Patientswereexcludedfromthestudyiftheywere diag-nosedwithatemporomandibularjointdisorderorpresented ahistoryoffacialbonetrauma,tumours,cleftliporpalate, orpreviousorthognathicsurgery.

Allparticipantsprovidedwritteninformedconsentbefore enrollment,andthestudyhasbeenconductedinaccordance withtheHelsinkiDeclarationof1973asrevisedin1983.As aretrospectivestudyitdidnotrequiredinstitutionalreview boardapproval.

All the surgeries were planned withthe same software (Dolphin 3D, Version 11.95 Dolphin). This program was selected after differentexperiences because, inadditionto greatversatilityandabundanceofdiagnostictools,ithasthe peculiaritytoprovide digitalsurgicalguidesin.stlformat, thatcanbeprintedbystandard3-dimensionalprinters.

A faithful representation of the facial skeletonand the soft tissueswas obtainedusing DigitalImagingand Com-munications in Medicine (DICOM) data from cone-beam computed tomography (CT). The following settings, were attended toobtain the scans:fieldof view20cm×20cm, 110kV,radiationexposure59mSV,andresolution0.4mm.

High resolutionvirtualdentalcasts in.stl.format,both in preoperative and postoperative occlusion, were needed as well. In our cases, theywere achievedstarting from a cone-beamCTscanof theplastermodelsat0.25mm.The DICOM data were then processed with open source soft-ware (Invesalius version 33.1.1, Centro de Tecnologia da Informac¸ãoRenatoArcher)toprocessthedentalmodelsand savethemin.stlformat.

According to the Dolphin 3-dimensisonal protocol, the DICOMdataoftheskullandthe.stldataofthedentalmodels werethenacquiredtobuildtheaugmentedmodel,readyfor planningandvirtualsurgery.

Thesameoperatorplannedalltheorthognathicsurgeries, guidedbyreal-time3-dimensionalcephalometricand clini-calmeasurements,movingthevirtualbonysegmentsuntilthe desiredoutcomewasachieved(Fig.1).Thereafter,the inter-mediatesplintwasgeneratedas.stlfileformat,thenitwas processedwithaspecificprogram(CreationWorkshop, Wan-hao)toobtainapropermodeltobeprintedin.cwsformat. Thisfilewasthenprintedonadesktop3-dimensionaldigital lightprocessingresinprinter(DLP).DLPresinprint technol-ogy isatypeofvatpolymerisation3-dimensionalprinting technology that usesaliquidphotopolymer resinwhichis abletosolidifyunderalightsource.WeadoptedtheWanhao Duplicator7plus(Wanhao3DPrinter,Jinhua).The interme-diatewaferwasprintedwithclearresin(Wanaho,Zhejiang). Thesettingswere:thickness(resolution):50microns,5 bot-tomlayers,curetime:40,000ms.

In thisway, the surgical planning and simulation were transferred tothe patientatthe timeof surgeryusing that computer-generated intermediate surgical splint (Fig. 2).

Fig.1.Virtualsurgicalplanning(publishedwiththepatient’sconsent).

Fig.2.Three-dimensionalprintedsurgicalsplint.

The same surgeon who planned the surgeries also did all the bimaxillary orthognathic procedures, according to the authors’standard osteotomy protocol(LeFort I osteotomy andEpker’s bilateral sagittal split ramusosteotomy).21–23

Thechinsegmentwasrepositionedwithsimple intraopera-tiveclinicalmeasurementsandevaluationoftheappropriate lower third vertical dimension and symmetry. The maxil-laryosteosynthesiswasdonewithfourL-shaped2.0titanium miniplatesand,themandibularramisagittalsplitswerefixed usingoneortwo,fourhole,thickened,straight,2.0titanium miniplates,bilaterally.

Theaccuracyofthe virtualplanningmethodwas deter-minedcomparing plannedmovements of theosteotomised jawstotheactualsurgicalmovements.Thiswasdoneonthe assumptionthatthesimplesuperimposition8,13,15ofthe sim-ulationandthefinalcephalometricresultsdoesnotdefineany

correlationbetweenthepositionalerrorandsurgical move-ments.Aslightpositionalerrorcanbeirrelevantwhendoing largemovementsbutwouldbeunacceptableifsurgical move-mentswere minor21 (for example,0.5mmdifference ona 10mmmaxillary advancement means5% deviationwhich wouldbequiteaminorerror,whilea0.5mmdifferenceona 2mmimpactionmeans25%deviation–thatisquiteahuge error).

Cephalometric analysis in the lateral view was made accordingtoRoth-Jarabakanalysis;cephalometricanalysis in the frontal view was made according to Ricketts anal-ysis using Dolphin® software. Virtual frontal and lateral cephalometriesofthesimulatedsurgerywereextractedfrom the 3-dimensional planning and compared to the postop-erative cephalometries,immediatelyafter surgery(3rd–5th postoperative days), before possible positional adaptive relapse, or TMJresorptionor remodelling could be estab-lished. The resulting data for the simulated and actual movementsofthejawslandmarkswerethencompared.The sameoperatoranalysed the preoperativeandpostoperative cephalometry.

Twelve angularandlinearmeasureswereconsideredin describing jawmovements (Table1).Themean difference betweenplannedandactualmovementswasusedasa mea-sureofaccuracy.

Statisticalanalysis

Statistical analysisof the differences was made withIBM SPSSStatisticsforWindows(version24.0,IBMCorp)using theWilcoxonsignedranktest.

e268 G.DeRiuetal./BritishJournalofOralandMaxillofacialSurgery58(2020)e265–e271

Table1

Linearandangularmeasurementsconsidered.

Lateralprojection Abbreviation Definition

Linearmeasurement

Anteriorfacialheight AntFacH Distancebetweenanteriornasionandmenton Posteriorfacialheight PostFacH Distancebetweensellaandgonion

Ramusheight RamusH Distancebetweenarticularandgonion

Angularmeasurement

Gonialangle GonAng Anglebetweenarticular,gonion,andmentonpoints

SNA SNA Anglebetweensella,nasion,andApoints

SNB SNB Anglebetweensella,nasion,andBpoints

ANB ANB AnglebetweenA,N,andBpoints

Firstupperincisor–SN U1-SN Anglebetweensella-nasionplaneandalineconnectingtheincisal edgeandtherootapexofthemostprominentupperincisor Firstupperincisor–FH U1-FH AnglebetweenFrankfortplaneandalineconnectingtheincisal

edgeandtherootapexofthemostprominentupperincisor Linearmeasurement

Dentalmidlinediscrepancy DentMid Discrepancybetweenupperandlowerdentalmidlines Maxillo-mandibularmidlinediscrepancy MxMdMid Discrepancybetweenmaxillaryandmandibularmidlines Occlusalplanetilt OcclTilt Differenceinheightfromtheocclusalplanetothe

zygomatico-frontalsutureleft–zygomatico-frontalsutureright

Table2

Preoperativediagnosisandsurgicalprocedures.

Diagnosis No.

ClassII 2

ClassIII 10

Facialasymmetry 2

Lateraldeviationduetocondylar hyperplasia 1 Total 15 Surgicalprocedure: Maxilla Maxillaryadvancement 8 Maxillary advancement+impaction 3 Maxillarysetback 1

Maxillaryset-back+impaction 1

Maxillaryimpaction 2 Maxillarysegmentalisation 6 Mandible Mandibularadvancement 2 Mandibularsetback 10 Mandibularcentration 3 Chin Advancement 3 Retrusion 9 Centration 3 Results

TheorthognathicpatientstreatedattheMaxillofacialSurgery UnitoftheUniversityHospitalofSassaribetweenSeptember 2017andMay2019wereintotal47.Thisstudy wasthen performedonacohortof15consecutivebimaxillarypatients (9maleand6female,meanage36years)whorespectedthe inclusionandexclusioncriteria.TwopresentedaclassII mal-occlusion,10aclassIIImalocclusion,twofacialasymmetry, and one lateral deviation due to hemimandibular elonga-tionfor condylar hyperplasia(Table2).Ofthe 15patients thatwereacceptedforthisstudy,allunderwentrepositional

genioplastyandsixmaxillarysegmentalisationintwoorthree segments (Table 3). No complications occurred during or aftersurgery.

An exhaustive summary of the measures, both angu-lar and linear, planned and actually obtained for each patient,isshowninTable3.Measurementofthedifferences amongplannedandactualmovementswasdeemedaccurate (Table4).Thedifferenceswerenotsignificantforeightof12 variablesconsidered. Thep valueswereevenlydistributed and rangedfrom p=0.001to p=0.820. Significant differ-enceswerereportedinfourmeasurements:U1SN(p=0.001), U1FH(p=0.008),dentalmidlinediscrepancies(p=0.006), andocclusalplanetilt(p=0.001)(Table1).

Discussion

The results of this study show good accuracy of computer-assisted orthognathicsurgery planned with Dol-phin 3-dimensional software and guided by homemade resin-printedintermediatesplints.Eachstepoftheprotocol hastobeexecutedwithprecision,because theaccuracyof eachstepisbuiltontheaccuracyofthepreviousone.13

Mostofthevariablesanalysedturnedouttobenot signif-icant.Thismeans thatthe differencebetweenthevirtually performed orthognathic surgery andthe actual postopera-tiveoutcomedidnotdifferthatmuch.Thisresultconfirms whatisalreadyknownfrompublishedreports:digital plan-ningismuch moreaccuratethanthe conventionalsurgical planning.3–6

Inthedigitalsurgicalplanning,alltheanatomical struc-turesarefaithfullyreproducedinanaugmentedmodelofthe face,whilstintheconventionalplanning,theonlystructures reproduced in threedimensions are the plaster cast dental models.18 Computer-aidedsurgical simulationrepresentsa paradigm shiftinsurgical planning forpatients with

cran-De Riu et al. / British Journal of Or al and Maxillofacial Sur g ery 58 (2020) e265–e271 e269 Table3

Simulatedandactualmovementsforeachpatient.

Caseno. Angulardistances(◦) Lineardistances(mm)

GonAng SNA SNB ANB U1-SN U1-FH OcclTilt RamusH AntFacH PostFacH DentMid MxMdMid

Plan Real Plan Real Plan Real Plan Real Plan Real Plan Real Plan Real Plan Real Plan Real Plan Real Plan Real Plan Real

1 125.8 123.4 86 84.4 84.5 84 1.5 0.4 97.9 107.8 110.7 116.4 1 0.6 48.3 33.6 113.9 83.8 56.5 55.5 0.1 0.2 1.1 −1.7 2 125.1 138.8 83.6 86.1 76.5 83.3 7.1 2.8 94.3 107.4 106.8 109.9 0.2 0.1 42.3 41.4 120.7 109.6 68.8 66.3 2.1 −2.3 1.9 1.8 3 142.4 142.2 73.8 80 71.4 75.8 2.4 4.2 93.5 101.7 112.9 116.6 0.8 0.6 37.8 41.2 131.2 111.5 61.6 67.8 0.7 0.4 0.2 5.7 4 125.4 139.8 89.2 88.6 84.5 84 4.7 4.5 101.3 101.4 110.7 116.2 −0.3 −0.3 35.2 42.3 97.6 102.3 58.3 63.1 0.4 0.1 0.5 −1.8 5 128.2 132.6 86 86.9 80.5 82 5.5 4.9 86.8 99.6 100.7 104.9 0.5 0.1 48.5 47.8 119.4 126.4 74.4 77.5 0.8 −0.2 0.3 1.8 6 125.9 136.6 91.2 83.7 86 77 5.2 6.7 94.6 97.2 104.9 111.2 0.3 −0.6 48.4 45 117.2 108.1 77.6 67.1 0.1 −0.2 0.2 −1.5 7 126.1 128.4 86.3 87.1 83.2 82.3 3.1 4.8 89.7 100.9 100.4 105.5 1.8 0.8 42 45.1 113 115.5 75 76.6 0.7 0.6 1.1 −1.3 8 126.7 141.2 85.2 93.1 81.3 89.2 3.9 3.9 88.7 105.4 105.6 113.7 0.5 0 51.5 42.3 130.7 110.5 75.7 68.8 0 −0.1 1.8 −3.4 9 133.4 138 92.7 86.5 88.1 83.9 4.6 2.5 95.8 97.2 108.8 111.1 1.8 −0.5 45.8 42.5 102.9 119.4 61.8 64.6 0 0.9 2.6 −2.3 10 129.5 121.1 83.5 90.5 81.7 88.1 1.8 2.4 88 97.1 102.6 106.2 1.7 0.9 42.4 36.9 109 116.2 64.7 68.7 0 −3.3 1.5 5.3 11 137.2 133.2 84.4 80.3 82.3 79.5 2.1 0.9 98 102.8 108.8 118.8 1 0.6 47.9 48.5 122.7 113.5 71.7 70.2 0.6 −0.2 0.2 0.1 12 120.7 130.9 72.1 80.3 67.9 72.4 4.2 8 74.3 84.6 93.8 96.5 0.7 0.7 38.6 38.5 119.9 119.8 74.5 69.1 1.4 −0.9 1.2 −1.4 13 125.2 122.7 93.2 94.9 92.1 89.2 1.1 5.7 101.5 104.7 109.7 110.7 3.6 0 47.2 41.1 97.9 96.3 69.4 65.6 0.3 0.1 0.6 0.4 14 129.9 133.7 90.5 86 87.5 82.1 3 3.9 99.5 98 116.2 107.8 1 −0.7 46.8 44.4 108.8 120.1 62.2 67.5 0.7 −0.7 0.5 −2.5 15 130.7 134.1 88.5 88.9 86.2 86.3 2.3 2.6 90.8 106.2 107.5 112.9 1.8 1.1 39.2 42 116 118 65.7 70.2 0.4 0.2 0.5 −0.4

e270 G.DeRiuetal./BritishJournalofOralandMaxillofacialSurgery58(2020)e265–e271

Table4

Comparisonofsimulatedandactualmovements.Dataaremean(SD).

Plannedmovement Realmovement Difference Z pvalue

Angulardistances(◦) GonAng 128.81(5.45) 132.21(6.07) −4.3(7.15) −1.846 NS SNA 85.74(6.07) 86.49(4.41) −0.75(5.03) −0.682 NS SNB 82.24(6.32) 82.61(4.85) −0.37(4.89) −0.227 NS ANB 3.50(1.71) 3.88(2.04) −0.38(2.22) −0.682 NS U1-SN 92.95(6.99) 100.80(5.82) −7.85(5.71) −3.237 0.001 U1-FH 106.67(5.61) 111.23(6.26) −4.55(4.72) −2.669 0.008 OcclTilt 1.09(0.94) 0.22(0.57) 0.87(0.98) −3.182 0.001 Lineardistances(mm) RamusH 44.13(4.84) 42.10(3.91) 1.96(5.49) −1.363 NS AntFacH 114.73(10.24) 111.40(10.72) 3.34(13.01) −0.682 NS PostFacH 67.86(6.84) 67.78(5.19) −0.05(5.05) −0.284 NS DentMid 0.55(0.58) −0.36(1.10) 0.91(1.41) −2.731 0.006 MxMdMid 0.95(0.74) −0.08(2.70) 1.03(2.93) −1.591 NS

iomaxillofacial deformities. Allthe information needed to perform the surgical simulation are storedvirtually inthe patientfileandwill automaticallybecombinedinasingle 3-dimensionalimage. Itispossibletohaveasimultaneous representationinthe3-dimensionalspaceofallthestructures thatwillbeaffectedbythesurgery.

In the present paper we compared the planned move-mentswithactualsurgicalmovementsin15bimaxillarycases operatedonwithanewprotocolofcomputer-assisted orthog-nathicsurgerywithin-house3-dimensionalprintedsurgical guides.Wereportedahighrateof accuracyinmostofthe linear andangularcephalometricmeasures assessed. Only fourof 12variablesweresignificantlydifferent.Forallof thepatients,U1-SNandU1-FHangles(Table1)increased inthepostoperativeconditioncomparedtothesurgical sim-ulation.Interestingly,thepostoperativeSNAangleremained substantiallyadherenttotheprevision.

Apossible explanationis that both U1-SN andU1-FH valuesaremostlyincreasedinthepatientsthatunderwent seg-mentalisationofthemaxilla:theanteriorfragmentacquireda morepronouncedtorqueduringsurgerythanitwasexpected inthesimulation,thecentreofrotationofthepremaxillary fragmentintheAlandmark.Itiseasytounderstandthatthe torqueofthemaxillaryfragmentsdiffersbetweensimulated andrealconditionsbecauseoftheactualandunpredictable bonycontactsalongtheosteotomiesofthepalatalvault.

Asecond,lessconvincing,explanationfortheseincreased valuesisthatduringthe3-dimensionalcephalometric anal-ysisphase,thelandmarksoftheupperincisors’rootshave beenplacedontothebonysurfaceofthepremaxillarybone inaslightlymoreadvancedpositionthantheywereplacedon the2-dimensionallateral postoperativecephalograms.This way,therewouldbesomediscrepancybetweenthesimulated andtheactualpostsurgicalincisor’sproclination:thislatter wouldbegreater thanexpectedbecausetheupperincisor’s axiswasmistakenlyplacedinamoreverticalpositionduring theplanningsteps.ThiserrorcanaffecttheangleamongU1 axisandboththelinesSNandFH.

The Dental-Midlinediscrepancywasalsofoundtobea significant variable, although the differences between the virtual (0.55 (0.58)) and the postoperative (−0.36 (1.10)) mean values are so slight that could havebeen causedby the smallest variation inpositioning the landmarks on the cephalometries.

Finally,theocclusalplanetiltcomparisongaveapvalueof 0.001,whichisconsideredtobesignificantinstatistics.The differencebetweenthevirtualandthepostoperativemeantilt isapproximately0.90◦,ameasuresosmalltobeconsidered anyway insignificantfrom a clinical point of view. Facial asymmetry is in fact usually noticed only when occlusal cantingismorethan3–4◦.24

Themajorlimitationofthisstudyisthequitesmallnumber ofpatientsundergoingevaluation.Forthisreason,asubgroup analysisofthecaseserieshasnotbeenmade.Thistypeof analysiscouldbeusefultoestablishiftherearedifferences intheaccuracyoftheprocedureaccordingtothetypeofjaw movements.Furtherstudiesonlargercaseswillbeneeded.

In our opinion, the most relevant aspect in this study is the reliabilityof the intermediatedigital splint in-house manufacturing. Digital planning software for PC-assisted orthognathic surgery is nowadays quite diffused. The one adopted inthisstudy (Dolphin3D,Version11.95)permits in-houseprototypingatacostaffordableforanyinstitution (aroundD14,000dependingonthelocalmarketcondition). Afterthesurgicalguidewasvirtuallyprocessed,thefilewas suddenlytransferredtoa3-dimensionalresinprinterinthe sameoffice.Thisrecenttypeoflowcost3-dimensional print-ersadoptsDirectUV-lightPrinting(DUP)technology.

Normal UVlight is the source of energy that provides tostartthefreeradicalpolymerisationneededtochangethe liquidImagepacresinintohardenedobjects.

Inourexperience,theseprintersarereliableandaccurate, theyareeasytouseanddonotrequireadedicatedroom.The costsofthehardwarearequiteminimal:3-dimensionalresin printer described inthiswork(WanhaoDuplicator7 plus, Wanhao3DPrinter)wassold,atthattime,ataroundD500. Wearenowtestingdifferentmodelsofresinprinters,already

adoptedinjewelleryanddentistry,thatcostevenless(around D200–250)andseemtoperformwell(forexample,Photon, Anycubic).Resinconsumptionisnegligibleandeachsplint requiredlessthanaD ofmaterials.Ultimately,theupsideof 3-dimensionalresinprintersisthattheymakeitpossibleto finalisetheplanningprocessincompleteautonomy,without theneedtodiscussthecaseswithnon-medicalpersonnel,to uploaddatatooutercompanies,waitinglongdaysfor manu-facturingandshipment,oftenfromforeigncompanies.Inour experiencewithtwodifferentmedicalcompanies,inaddition tothecostofthesoftware(aroundD9–10,000),thepriceof eachsplint wasaroundD300 tobe chargedtothepatient, thatisnowthepriceofawholeLCD-resinprinter.

Conclusions

Anewprotocolforvirtualsurgicalplanningandcompletely in-housesurgicalguidesmanufacturingisreported.These 3-dimensionalresinprintedsurgicalguidesweredemonstrated tobeareliablealternativetothecommercialonesandshowed high rate of accuracy for most of the variables assessed. Fourout of12 of theseshowedsignificant errors,buttwo of themwere onlyminimal discrepancieswithnoclinical implications.

To confirm thisstudy’s results,isnecessary toperform otherresearchesonlargerandmorehomogeneoussamples.

Conflictofinterest

Wehavenoconflictsofinterest.

Ethicsstatement/confirmationofpatients’permission

Asaretrospectivestudyitdidnotrequiredethicsapproval. Patients’permissionwasobtained.

Fundingsources

Thisresearchdidnotreceiveanyspecificgrantfromfunding agenciesinthepublic,commercial,ornot-for-profitsectors.

References

1.EllisIIIE.Bimaxillarysurgeryusinganintermediatesplinttoposition themaxilla.J.OralMaxillofac.Surg1999;57:53–6.

2.StokbroK, Aagaard E,Torkov P, etal.Surgicalaccuracy of three-dimensionalvirtualplanning:apilotstudyofbimaxillaryorthognathic proceduresincludingmaxillarysegmentation.Int.J.OralMaxillofac. Surg2016;45:8–18.

3.EllisIIIE,TharanonW,GambrellK.Accuracyofface-bowtransfer: effectonsurgicalpredictionandpostsurgicalresults.J.OralMaxillofac. Surg1992;50:562–7.

4.GatenoJ,ForrestKK,CampB.Acomparisonof3methodsof face-bowtransferrecording:implicationsfororthognathicsurgery.J.Oral Maxillofac.Surg2001;59:635–40.

5.SharifiA,JonesR,AyoubA,etal.Howaccurateismodelplanningfor othognathicsurgery?Int.J.OralMaxillofac.Surg2008;37:1089–93.

6.ZizelmannC,HammerB,GellrichNC,etal.Anevaluationofface-bow transferfortheplanningoforthognathicsurgery.J.OralMaxillofac.Surg

2012;70:1944–50.

7.BarbenelJC,PaulPE,KhambayBS,etal.Errorsinorthognathicsurgery planning:theeffectofinaccuratestudymodelorientation.Int.J.Oral Maxillofac.Surg2010;39:1103–8.

8.XiaJJ,GatenoJ,TeichgraeberJF,etal.Accuracyofthecomputer-aided surgicalsimulation(CASS)systeminthetreatmentofpatientswith com-plexcraniomaxillofacialdeformity:apilotstudy.J.OralMaxillofac.Surg

2007;65:248–54.

9.MetzgerMC,Hohlweg-MajertB,SchwarzU,etal.Manufacturingsplints fororthognathicsurgeryusingathree-dimensionalprinter.OralSurg. OralMed.OralPathol.OralRadiol.Endod2008;105:e1–7.

10.SwennenGRJ,MollemansW,DeClercqC,etal.Acone-beamcomputed tomographytriplescan proceduretoobtainathree-dimensional aug-mentedvirtualskullmodelappropriateforothognathicsurgeryplanning.

J.Craniofac.Surg2009;20:297–307.

11.SwennenGRJ,MollemansW,SchutyserF.Three-dimensionaltreatment planningoforthognathicsurgeryintheeraofvirtualimaging.J.Oral Maxillofac.Surg2009;67:2080–92.

12.SwennenGR,MommaertsMY,AbeloosJ,etal.Acone-beamCTbased techniquetoaugmentthe3Dvirtualskullmodelwithadetaileddental surface.Int.J.OralMaxillofac.Surg2009;38:48–57.

13.HsuSS,GatenoJ,BellRB,etal.Accuracyofacomputer-aidedsurgical simulationprotocolfororthognathicsurgery:aprospectivemulticenter study.J.OralMaxillofac.Surg2013;71:128–42.

14.DeRiuG,MeloniSM,BajA,etal.Computer-assistedorthognathic surgeryforcorrectionoffacialasymmetry:resultsofarandomised con-trolledclinicaltrial.Br.J.OralMaxillofac.Surg2014;52:251–7.

15.MarchettiC,BianchiA,BassiM,etal.Mathematicalmodelingand numericalsimulationinmaxilla-facialvirtualsurgery(VISU).J. Cran-iofac.Surg2006;17:661–8.

16.MazzoniS,BadialiG,Lancellotti L, etal.Simulation-guided navi-gation: a newapproach to improve intraoperative three-dimensional reproducibility during orthognathic surgery. J. Craniofac. Surg

2010;21:1698–705.

17.TuckerS,Soares-CevidanesLH,StynerM,etal.Comparisonofactual surgicaloutcomesand3-dimensionalsurgicalsimulations.J.Oral Max-illofac.Surg2010;68:2412–21.

18.Aboul-HosnCenteneroS,Hernàndez-AlfaroF.3Dplanningin orthog-nathicsurgery:CAD/CAMsurgicalsplintsandpredictionofthesoftand hardtissuesresults–ourexperiencein16cases.J.Craniomaxillofac. Surg2012;40:162–8.

19.ZinserMJ,MischkowskiRA,SailerHF,etal.Computer-assisted orthog-nathic surgery: feasibility study using multipleCAD/CAM surgical splints.OralSurgOralMedOralPatholOralRadiol2012;113:673–87.

20.LinHH,LonicD,LoLJ.3Dprintinginorthognathicsurgery–aliterature review.J.Formos.Med.Assoc2018;117:547–58.

21.DeRiuG,VirdisPI,MeloniSM,etal.Accuracyofcomputer-assisted orthognathicsurgery.J.Craniomaxillofac.Surg2018;46:293–8.

22.FrisciaM,SbordoneC,PetrocelliM,etal.Complicationsafter orthog-nathic surgery: ourexperience on 423 cases. Oral Maxillofac Surg

2017;21:171–7.

23.VairaLA,MassarelliO,MeloniSM,etal.Alveolarnerveimpairment followingbilateralsagittalsplitramusosteotomyandgenioplasty.JOral MaxillofacSurgMedPathol2017;29:203–9.

24.Choi KY. Analysis of facial asymmetry. Arch Craniofac Surg