an in vitro investigation into the accuracy of cad/cam digitizing

78
This electronic thesis or dissertation has been downloaded from the King’s Research Portal at https://kclpure.kcl.ac.uk/portal/ The copyright of this thesis rests with the author and no quotation from it or information derived from it may be published without proper acknowledgement. Take down policy If you believe that this document breaches copyright please contact [email protected] providing details, and we will remove access to the work immediately and investigate your claim. END USER LICENCE AGREEMENT This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International licence. https://creativecommons.org/licenses/by-nc-nd/4.0/ You are free to: Share: to copy, distribute and transmit the work Under the following conditions: Attribution: You must attribute the work in the manner specified by the author (but not in any way that suggests that they endorse you or your use of the work). Non Commercial: You may not use this work for commercial purposes. No Derivative Works - You may not alter, transform, or build upon this work. Any of these conditions can be waived if you receive permission from the author. Your fair dealings and other rights are in no way affected by the above. An in vitro investigation into the accuracy of CAD/CAM digitizing devices for measurement of tooth wear Austin, Rupert Sloan; Elliott, Thomas Awarding institution: King's College London Download date: 12. Apr. 2018

Upload: lamtram

Post on 13-Feb-2017

220 views

Category:

Documents


3 download

TRANSCRIPT

Page 1: An in vitro investigation into the accuracy of CAD/CAM digitizing

This electronic thesis or dissertation has been

downloaded from the King’s Research Portal at

https://kclpure.kcl.ac.uk/portal/

The copyright of this thesis rests with the author and no quotation from it or information derived from it

may be published without proper acknowledgement.

Take down policy

If you believe that this document breaches copyright please contact [email protected] providing

details, and we will remove access to the work immediately and investigate your claim.

END USER LICENCE AGREEMENT

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0

International licence. https://creativecommons.org/licenses/by-nc-nd/4.0/

You are free to:

Share: to copy, distribute and transmit the work Under the following conditions:

Attribution: You must attribute the work in the manner specified by the author (but not in any way that suggests that they endorse you or your use of the work).

Non Commercial: You may not use this work for commercial purposes.

No Derivative Works - You may not alter, transform, or build upon this work.

Any of these conditions can be waived if you receive permission from the author. Your fair dealings and

other rights are in no way affected by the above.

An in vitro investigation into the accuracy of CAD/CAM digitizing devices formeasurement of tooth wear

Austin, Rupert Sloan; Elliott, Thomas

Awarding institution:King's College London

Download date: 12. Apr. 2018

Page 2: An in vitro investigation into the accuracy of CAD/CAM digitizing

KING’S  COLLEGE  LONDON  DENTAL  INSTITUTE  

An  in  vitro  investigation  into  the  accuracy  of  CAD/CAM  digitizing  

devices  for  measurement  of  tooth  

wear  Submitted  in  Partial  Fulfilment  of  

MClinDent  Prosthodontics    

Thomas  Elliott  7/1/2014  

 

 

Page 3: An in vitro investigation into the accuracy of CAD/CAM digitizing

1    

Contents  

Acknowledgements  ..................................................................................................................  4  

Abstract  ..........................................................................................................................................  5  

Literature  Review  ......................................................................................................................  7  

1.1   Introduction  .................................................................................................................  7  

1.2   Profilometry  ................................................................................................................  8  

1.2.1   Surface  Profile  Measurement  ......................................................................  9  

1.2.2   Contacting  Profilometry  ................................................................................  9  

1.2.3   Non-­‐contacting  Profilometry  ....................................................................  10  

1.2.4   Limitations  of  Optical  Instruments  ........................................................  11  

1.3   Applications  of  Profilometry  in  Dental  Research  .....................................  12  

1.4   Accuracy  of  Profilometers  in  Dental  Research  ..........................................  15  

1.4.1   Fundamentals  of  Metrology  ......................................................................  17  

1.4.2   Accuracy  of  Profilometry  ............................................................................  17  

Measuring  Tooth  Wear  ...................................................................................................  22  

1.4.3   Qualitative  Measurement  of  Tooth  Wear  ............................................  23  

1.4.4   Traditional  Quantitative  Measurement  of  Tooth  Wear  ................  24  

1.4.5   Profilometry  and  Tribology  Studies  ......................................................  26  

Aims  and  Objectives  ..............................................................................................................  32  

2.1   Aims  .............................................................................................................................  32  

2.2   Objectives  ..................................................................................................................  32  

Page 4: An in vitro investigation into the accuracy of CAD/CAM digitizing

2    

Null  Hypotheses  ......................................................................................................................  34  

Materials  and  Methods  .........................................................................................................  35  

3.1   Test  Digitizing  Devices  .........................................................................................  36  

3.2   Test  Conditions  for  Measurement  ..................................................................  36  

3.3   Measurement  Software  ........................................................................................  36  

3.4   Specimen  Preparation  ..........................................................................................  37  

3.5   Engineering  Slip  Gauge  Specimen  ...................................................................  37  

3.6   Inlay  shaped  Specimen  and  Bridge  Shaped  Specimen  ...........................  41  

3.7   ‘Proof  of  Concept’  of  Using  the  3Shape®  Extra-­‐oral  Scanner  and  the  

3M™  True  Definition   Intra-­‐oral   Scanner   to   Follow  a  Profilometric  Tooth  

Wear  Measurement  Technique  ...................................................................................  50  

3.8   Statistics  .....................................................................................................................  53  

Results  .........................................................................................................................................  55  

4.1   Measurement  of  Engineering  Slip  Gauges  ...................................................  55  

4.2   Repeatability,  Reproducibility  and  Trueness  of  Measurement  of   the  

Inlay  Shaped  Specimen  and  the  Bridge  Shaped  Specimen  ..............................  57  

4.2.1   Repeatability  Measurement  Results  ......................................................  57  

4.2.2   Reproducibility  of  the  Measurement  ....................................................  59  

4.2.3   Trueness  of  the  Measurement  .................................................................  61  

4.3   ‘Proof  of  Concept’  of  Using  the  3Shape®  Extra-­‐oral  Scanner  and  the  

3M™  True  Definition   Intra-­‐oral   Scanner   to   Follow  a  Profilometric  Tooth  

Wear  Measurement  Technique  ...................................................................................  63  

Page 5: An in vitro investigation into the accuracy of CAD/CAM digitizing

3    

Discussion  ..................................................................................................................................  65  

Conclusion  .................................................................................................................................  70  

References  .................................................................................................................................  71  

 

   

Page 6: An in vitro investigation into the accuracy of CAD/CAM digitizing

4    

Acknowledgements  

Professor  David  Bartlett  

Dr.  Rupert  Austin  

Mr  Alberto  Alverez  and  Mr  Steven  Nelson  from  3M  ESPE  for  the  loan  of  the  

3M™  True  Definition  Scanner.  

The   laboratory   technicians   in  Orthodontics   for   their   training   in   the  use   of  

the  3Shape  scanner  

   

Page 7: An in vitro investigation into the accuracy of CAD/CAM digitizing

5    

Abstract  

Aim:  To  investigate  the  accuracy  and  precision  of  measurement  of  extra-­‐oral  

and   intra-­‐oral   digitizing   devices,   and   to   carry   out   a   ‘proof   of   concept’  

evaluation   following   a   previously   published   profilometric   tooth   wear  

measurement  technique.    

Methodology:   Engineering   slip   gauges   of   known  dimensions  were   scanned  

and   measurements   were   compared   to   the   gold   standard   measurement  

device   (a   triangulation   laser   profilometer).   Inlay-­‐   and   bridge-­‐shaped  

artifacts   were   used   to   assess   precision   (repeatability   and   reproducibility)  

and  accuracy  (trueness)  of  the  test  devices.  Each  device  recorded  sixty  scans  

and   the   variances   of   the   height   and   angle   measurements   of   the   artifacts  

were   calculated   to   assess   repeatability   and   reproducibility.   Accuracy   was  

then  calculated  by  measuring  the  difference  between  the  mean  repeatability  

measurement   of   the   reference   device   and   the   mean   repeatability  

measurement  of  the  test  device.  Finally,  tooth  wear  calibration  models  that  

had   previously   used   to   evaluate   a   method   of   profilometric   tooth   wear  

measurement  were  scanned  by  the  test  digitizing  devices  and  compared  to  

the   reference   profilometer.   In   all   instances   the   scans  were   analysed   using  

Geomagic®  Qualify  11  computer  software  and  parametric  statistical  analyses  

were  performed.    

Results.  All   digitizing   devices   demonstrated   good   correlation   in  measuring  

slip  gauge  widths  of  400  μm  and  above.  However,  below  this  value  the  test  

devices   were   unable   to   provide   accurate   readings,   while   the   reference  

scanner  was  able   to  measure   to  100  μm.  Repeatability  and   reproducibility  

Page 8: An in vitro investigation into the accuracy of CAD/CAM digitizing

6    

for   both   the   extra-­‐oral   and   the   intra-­‐oral   scanners   was   found   to   be  

statistically   significantly   increased   than   that   of   the   reference   profilometer  

when  measuring  either   the  height  or   the  angle  of  both   the   inlay   specimen  

and  the  bridge  specimen  (p<0.05).  Both  the  3Shape®  extra-­‐oral  scanner  and  

the   3M™   True   Definition   intra-­‐oral   scanner   were   able   to   follow   the  

previously   published   profilometric   tooth   wear   measurement   process,  

however  they  showed  a  mean  volume  error  of  ±  0.2  mm3.    

Conclusions.   The   optical   resolution   and   precision   of   the   test   scanners  was  

inferior  to  the  reference  profilometer  however  the  trueness  of  measurement  

between  the  test  devices  was  not  statistically  significant  (p>0.05).    

Page 9: An in vitro investigation into the accuracy of CAD/CAM digitizing

7    

Literature  Review  

1.1 Introduction    

Following   rapid   progress   in   the   use   of   computer   assisted   processing  

technology   in   other   industries,   research   and   development   in   dental  

profilometry  has  been  prevalent  since  the  1980’s  and  is  still  growing  today.  

In  2012  Millennium  Research  Group  (MRG),  the  global  authority  on  medical  

technology   market   intelligence,   predicted   that   the   worldwide   CAD/CAM  

system  market   is   set   to   reach   a   value   upwards   of   $560   million   by   2016.  

Moreover,  they  state  that  from  around  50  percent  of  the  CAD/CAM  market  

in   2010,   chairside   systems   and   intra-­‐oral   scanners   will   hold   nearly   60  

percent  of  a  much  larger  market  by  2016.  This  indicates  that  the  use  of  these  

technologies   will   be   far   more   widespread   in   dentistry   over   the   coming  

years.  

The   use   of   profilometry   in   clinical   dentistry   is   becoming   increasingly  

relevant,   especially   with   the   advent   of   optical   scanning   systems   that   can  

record  the  oral  environment  directly.  This  negates  the  need  for  impressions  

that  can   introduce  errors,  albeit  small  ones,  due  to  the   inherent  properties  

of  the  material  used  (Eames  et  al.  1979,  Williams  et  al.  1984),  operator  error  

(Carrotte  et  al.  1993,  Winstanley  et  al.  1997),  or  due  to  the  casting  process.  

Extra-­‐oral   non-­‐contact  profilometers   are   also  used   frequently   in  dentistry,  

including  the  3Shape®  scanner  under  investigation  in  this  study.    However,  

these   machines   still   rely   on   transfer   of   information   via   traditional  

impressions.  

Page 10: An in vitro investigation into the accuracy of CAD/CAM digitizing

8    

Surface   point   measurements   captured   by   profilometers   are   stored   in   the  

form  of  binary  digits  or  ‘bits’.  These  bits  are  then  combined  into  larger  units  

to   provide   more   meaningful   data.   In   practical   terms   the   advantages   of  

having  data  in  this  form  is  that  it  can  be  stored  and  then  used  to  guide  the  

milling  process  or  for  future  reference.  Another  advantage  is  that  scans  can  

be  recorded  in  one  location  and  the  information  sent  via  the  Internet  to  an  

alternate   location   where   it   can   be   used   for   various   purposes   but   usually  

prosthesis   construction.   This   has   transformed   the   scale   of   production  

previously  seen  with  traditional  techniques.    

The   use   of   dental   CAD/CAM   for   construction   of   prostheses   and   for  

laboratory  based  studies  have  been  well  documented  (Persson  et  al.  1995,  

Luthardt   et   al.   2001,   Syrek   et   al.   2010   etc.)   but   there   is   need   to   establish  

whether  this  technology  is  able  to  offer  further  clinical  benefits,  such  as  the  

quantitative  measurement  of   tooth  wear  over  time.  As  noted  by  DeLong  et  

al.  (2003),  “interactive  three-­‐dimensional  images  of  the  soft  and  hard  tissues  

of  the  dental  patient  will  provide  dentists  with  quantitative  evidence  to  aid  

dentists   in   diagnosis,   treatment   planning,   and   outcome   assessment.”  

Naturally,   in  order  for  this  to  be  applicable  we  must  show  that  any  images  

recorded  are  accurate  representations  of  the  patient’s  intra-­‐oral  tissues.  

1.2 Profilometry  

Essentially  a  profilometer  is  a  device  used  to  record  the  surface  profile  of  an  

object.  Vlaar  et  al.  (2006)  defined  a  dental  profilometer  as  “a  device  used  to  

record   the   topographical   characteristics   of   teeth,   dental   impressions   or  

Page 11: An in vitro investigation into the accuracy of CAD/CAM digitizing

9    

stone   models   by   analog   or   digital   methods   for   use   in   computer   assisted  

design  and  manufacturing  of  dental  prosthetic  restorative  devices.”    

1.2.1 Surface  Profile  Measurement  

Surface   profile   measurement   is   described   in   ‘Fundamental   Principles   of  

Engineering   Nanometrology   (2nd   Edition)’   as   the   measurement   of   a   line  

across  a  surface  that  can  be  represented  mathematically  as  a  height  function  

with  lateral  displacement,  z  (x).  By  lining  up  parallel  profiles  a  topographic  

image  can  be  produced  representing  z  (x,y).  Profilometers,  also  referred  to  

as  digitizing  devices  or  scanners,  can  be  broadly  divided  into  two  categories:  

contact  and  non-­‐contact.  Either  a   stylus   (contact)  or  an  optical   instrument  

(non-­‐contact)   is   used   to   measure   the   surface   profile.   In   non-­‐contacting  

profilometry  an  optical  instrument  is  used  to  carry  out  a  measurement,  and  

the  surface  profile  of   the  point   cloud  recorded   is   then  usually  extracted   in  

software.  

1.2.2 Contacting  Profilometry  

This   is  by  far  the  most  common  method  for  measuring  surface  topography  

within   engineering   measurement   applications.   In   contacting   profilometry  

the  stylus  is  traversed  across  a  line  on  the  surface  being  measured  at  a  given  

distance  and  contact  pressure  (typically  0.75  mN).  The  profile  measurement  

can   then   be   recorded   by  mapping   vertical   displacement   of   the   stylus   as   a  

function  of   its   position  z   (x,y).   The   accuracy  of   the  data   recorded  depends  

largely  on  the  radius  of  the  stylus  and  the  scan  speed.  The  stylus  can  be  as  

small   as   2   μm,   and   as   such   can   usually   offer   better   accuracy   than   non-­‐

contact  profilometers.  

Page 12: An in vitro investigation into the accuracy of CAD/CAM digitizing

10    

A   drawback   of   a   stylus   instrument   when   scanning   an   area   is   the   time  

required  to  record  a  measurement  as  this  can  often  be  up  to  several  hours.  

This  is  an  area  where  optical  instruments  offer  an  advantage  over  the  stylus  

instruments.  Within  dentistry  this  type  of  profilometer  only  has  laboratory  

applications.  

1.2.3 Non-­‐contacting  Profilometry  

Non-­‐contact   profilometers   operate   using   a   variety   of   techniques,   e.g.   laser  

triangulation  and  confocal  microscopy.  These  techniques  can  be  regarded  in  

two  categories:  those  that  measure  actual  surface  topography  by  scanning  a  

beam  or  using  field  of  view,  and  those  that  measure  a  statistical  parameter  

of  the  surface,  usually  by  analyzing  the  distribution  of  scattered  light.    

The   measurement   can   be   affected   by   the   need   to   stitch   optical   images  

together.   Stitching   can   be   a   significant   source   of   error   in   optical  

measurements   (Bray   2004,   Zhang   2006),   and   it   is   important   that   the  

process  is  well  defined  and  understood  for  any  given  application.  

1.2.3.1 Triangulation  Lasers  

Laser   triangulation   instruments  measure  the  relative  distance  to  a  surface.  

Light   is   projected  onto   the   surface  where   it   then   scatters.  A  detector   then  

focuses   the   scattered   light  onto  a   sensor.  Because   the   fixed  angle  between  

the   projector   and   sensor   is   known,   the   distance   to   the   object   can   be  

calculated   through   Pythagoras’   theorem,   as   one   side   and   one   angle   of   the  

triangle   are   now   known,   hence   the   name   ‘‘triangulation’’.   As   the   surface  

changes   the   focused   light   moves   across   the   sensor   enabling   the   surface  

characteristics  to  be  recorded.  The  reference  profilometer  used  in  this  study  

Page 13: An in vitro investigation into the accuracy of CAD/CAM digitizing

11    

and   the  3Shape®  extra-­‐oral   scanner  both  utilise   this   technology.  However,  

the   two   machines   operate   differently   in   that   the   former   has   a   fixed   scan  

head  whereas  the  latter  has  more  cameras  and  a  motion  system  supporting  

several  axes  for  positioning  the  scanned  object  towards  the  light  source  and  

camera,   allowing   surface   mapping   from   multiple   angles,   including  

undercuts.  

1.2.3.2 Active  Wavefront  Sampling  

This   is   the   3-­‐D   imaging   technique   employed   by   the   3M™   True   Definition  

intra-­‐oral  scanner.  Essentially  light  reflected  from  the  object  being  scanned  

is  reflected  through  a  lens  system  onto  a  sensor.  If  the  object  is  in  focus  then  

it  coincides  with  the  given  focal  length  of  the  lens  but  if  the  object  is  blurred  

then   the   distance   from   the   object   to   the   lens   is   calculated   by   a   set  

mathematical   formula.  This   information   is   then  combined  to  provide  three  

dimensional   cloud   points   required   for   surface   mapping.   Twenty   3-­‐D  

datasets  per  second  can  be  captured  with  over  10,000  data  points   in  each,  

resulting  in  over  or  24  million  data  points  for  an  accurate  scan.  

1.2.4 Limitations  of  Optical  Instruments  

Non-­‐contact   profilometers   offer   a   number   of   advantages   over   contacting  

profilometers:  they  do  not  physically  contact  the  surface  so  won’t  damage  it  

and  they  are  usually   faster  at  recording  measurements.  However  there  are  

also   some   limitations   to   the   use   of   these   optical   instruments.   In   order   to  

magnify   the   surface   features   of   the   object   being   recorded   a  microscope   is  

required.   Any   instrument   that   uses   a   microscope   objective   has   defined  

limitations:  

Page 14: An in vitro investigation into the accuracy of CAD/CAM digitizing

12    

1. Numerical  Aperture   –  The  numerical   aperture   is   a  measure   of   how  

much  light  can  be  collected  by  an  optical  system  such  as  a  microscope  

lens.  This  determines   the   largest  slope  angle  of   the  surface   that  can  

be   measured   because   it   is   related   to   the   acceptance   angle,   which  

indicates   the   size   of   a   cone   of   light   that   can   be  physically   reflected  

back  into  the  objective  lens  and  therefore  be  measured.  

2. Optical   resolution   of   the   objective   –   This   determines   the  minimum  

distance   between   two   lateral   features   on   a   surface   that   can   be  

measured.  

Another   limitation   for   non-­‐contacting   profilometry   is   optical   spot   size.   In  

scanning  instruments  like  the  ones  used  for  dentistry  this  will  determine  the  

area   of   surface  measured   as   the   instrument   scans,  much   like   the   stylus   in  

contacting  profilometry.  

As   result   of   the   aforementioned   limitations   most   optical   instruments  

experience   difficulties   when   measuring   very   high   slope   angles   or  

discontinuities.  

1.3 Applications  of  Profilometry  in  Dental  Research  

Over  the  years  the  description  of  techniques  and  methods  for  assessing  the  

accuracy   of   profilometers   have   been   varied.   Most   of   the   initial   reports  

focused  on  measuring  wear   in  dental  materials,  particularly  composites,   in  

order   to   assess   their   suitability   as   restoratives.   Initial   techniques   largely  

relied  on  before  and  after  impressions  and  casts  in  order  to  assess  the  wear,  

which,  as  previously  discussed,  carry  their  own  inaccuracies.  

Page 15: An in vitro investigation into the accuracy of CAD/CAM digitizing

13    

In  one  of   the   first  attempts  at  volumetric  measurement  using  profilometry  

Atkinson  et   al.   (1982)  used   laser  optical   interferometry   in  order   to   assess  

wear   in   dental   restorations.   Measurements   were   made   by   comparing  

suitable   impressions   of   the   restorative  being   investigated  both  before   and  

after  the  test.  Detailed  information  about  the  pre  and  post  wear  restorative  

was  then  obtained  from  contour  maps  of  the  impressions  in  order  to  find  the  

volume  of  restorative  material   lost  during  the  test.  The  authors  recognised  

that   the  whole   experiment   relies   on   the   accuracy   of   the   polyvinylsiloxane  

impression  material  so  they  investigated  this  and  found  it  to  be  “adequate.”  

The  authors  also  related  how  the  pre  and  post  wear  contour  maps  had  to  be  

repositioned  in  order  to  have  the  same  reference  plane  during  comparison  

and  provide  reproducibility.  The  contour  maps  were  compared  using  three  

different  methods:  a  fringe  displacement  method,  a  moiré  method  and,  most  

appropriate  to  the  experiment  being  presented,  by  computer-­‐aided  analysis.  

When   looking   at   the   volume   of   material   lost,   an   accuracy   of   4%   was  

described   for   the   computer-­‐aided   analysis   method,   which   the   authors  

considered   to  be   the   “most   accurate”   of   the   three  methods  described.  The  

discussion   does   state   that   the   main   drawback   was   the   length   of   time  

required   to   digitize   the   contour   maps   (30   minutes   per   pair)   and   the  

experiment   described   translates   to   the   reader   as   a   particularly   arduous  

undertaking.  

DeLong  et  al.   (1985)  describe  a  method  to  measure  and  capture  change   in  

surface  contour  of  restorations  using  a  combination  of  servohydraulics  and  

computer   graphics.  Measuring   two   depressions   of   a   known   value   in   a   flat  

aluminium   surface   assessed   the   accuracy   of   this   method.   Using   a  

Page 16: An in vitro investigation into the accuracy of CAD/CAM digitizing

14    

servohydraulic   machine,   a   stylus   was   run   across   the   surface   of   the  

depressions  and  the  coordinates   fed   into  a  microcomputer.  Each  pass  over  

the   surface   constituted   one   profile   and   numerous   profiles   could   then   be  

assembled   using   computer   graphics   in   order   to   provide   an   image   of   the  

surface.   The   device  measured   a   6   μm3   depression   as   5.3   ±   0.6   μm3   (12%  

error)  and  a  28.1  μm3  depression  as  28.8  μm3  ±  1.1  μm3  (2.5%  error),  which  

were   considered   “acceptable”   and   “very   good”   respectively.   Even   at   this  

early  stage  of  computer  mapping  the  authors  recognised  that  “an  important  

feature   of   this   method   is   the   ability   to   visualise   the   change   in   surface  

contour  and  relate  it  to  anatomy  of  the  tooth.”  

Hewlett   et   al.   (1992)   first   identified   the   need   for   a   gold   standard   against  

which   profilometers   could   be   periodically   calibrated.   They   note,  

“Descriptions   of   instrument   capability   have   rarely   distinguished   between  

precision   (the   level   of   instrument   variability)   and   accuracy   (the   degree   to  

which  a  measurement  actually  represents  what  it  is  intended  to  represent).”  

In  their  investigations  a  contact  profilometer  was  used  to  measure  points  at  

200   μm   intervals   on   a   precision   steel   sphere,   half   embedded   in   acrylic.  

These  points  were  stored  in  a  computer  memory.  ‘True’  value  for  any  point  

on  a  hemispherical  surface  can  be  measured  using  Pythagorean  Theory,  so  

by   using   the   known   value   of   the   scanner   contact   tip   (1   mm)   and   the  

precision   sphere   (6.35   mm)   it   was   therefore   possible   to   determine  

trigonometrically  the  error  of  each  point  measurement  due  to  the  slope  on  

the   surface   of   the   sphere.   Because   the   point   error  would   equal   zero   if   no  

mechanical   error   existed,   the   accuracy   of   the   device   as   a   function   of   the  

slope   of   the   measurement   point   could   be   quantitatively   measured.   The  

Page 17: An in vitro investigation into the accuracy of CAD/CAM digitizing

15    

results  indicated  a  linear  relationship  between  increased  surface  slope  angle  

and   increased   error   for   the   device   being   tested.   The   authors   note   that   in  

practical   terms   that   when   measuring   an   occlusal   surface   a   coordinate  

measuring   machine   will   undoubtedly   encounter   steep   angles   due   to   the  

cusps.  Not  only  that,  but  the  size  of  the  contact  probe  is  problematic  when  

considering   fine   detail.   Despite   this   the   authors   assert,   “The   accuracy   test  

described   here   has   validity   and   is   applicable   as   a   standard   for   three-­‐

dimensional   contour   measuring   instruments.”   The   findings   of   this   study  

were   similar   to   Rudolph   et   al.   (2007)   who   investigated   the   influence   of  

digitizing  and  surface  morphology  on  the  accuracy  in  CAD/CAM  technology.  

Three  different  non-­‐contact  profilometers  were  used  to  scan  metal  teeth  and  

the   authors   found   that   areas   of   strong   curvature   showed   the   largest  

deviation.   However   they   found   that   these   deviations   do   not   significantly  

influence  error   in   the  duplicate   and  are   to  be   expected  when  measuring  a  

free  form  surface  such  as  a  tooth.  

1.4 Accuracy  of  Profilometers  in  Dental  Research  

Producing   an   exact   three-­‐dimensional   replica   of   any   given   object   is   near  

impossible  no  matter  what  method  or  material   is  being  used.  The  areas   in  

which   error  might   occur   in   a   replication   process   are   often   the   subjects   of  

much   investigation   in   an   effort   to   be   as   accurate   and   precise   as   possible.  

Dahlmo  (2001)  notes  that  in  CAD/CAM  there  are  three  potential  sources  for  

error:  

1. Measurement   accuracy   of   the   device   used   and   in   the   transfer   of  

data  files  to  the  computer.  

Page 18: An in vitro investigation into the accuracy of CAD/CAM digitizing

16    

2. The   ability   for   computer   software   to   accurately   transfer   data,   so  

that  what  is  designed  on  the  computer  translates  accurately  in  the  

manufacture.  

3. The  manufacturing  processes  and  the  precision  of  machinery  used  

to  manufacture  a  restoration.  

It   is   not   the   remit   of   this   investigation   to   cover   the   accuracy   of   the  

manufacturing   process.   However,   the   inherent   accuracy   of   profilometers  

and  their  ability  to  provide  quantitative  data  will  be  considered.    

In   his   book   ‘Degradation   of   Dental   Polymers’   (1987)   Jean-­‐François   Roulet  

provided  a  detailed  description  for  accuracy  determination  of  profilometers  

using  impressions  and  casts  highlighting  three  key  determinants:  

1. The  inherent  accuracy  of  the  profilometer.  

2. Producing  an  accurate  replica.  

3. Repositioning  the  replica  accurately.  

Modern   profilometers   are   able   to   capture   information   directly   from   the  

intra-­‐oral  environment  and  surface-­‐mapping  software  exists  to  allow  best  fit  

matching  of  scans.  This  means  that  producing  and  repositioning  a  replica  is  

no   longer   necessary   and   the   inherent   accuracy   of   the   device   is   now   the  

major   determinant.   However,   before   considering   modern   devices   it   is  

relevant   to   look   at   some   of   the   original   techniques   employed   using  

profilometers   in   order   to   appreciate   the   progress   made   in   this   field   of  

dentistry.    

Page 19: An in vitro investigation into the accuracy of CAD/CAM digitizing

17    

1.4.1 Fundamentals  of  Metrology  

In  metrology  accuracy   is   the  degree  to  which  a  measurement  of  a  quantity  

conforms   to   a   ‘true’   value   of   the   quantity.   For   this   reason   it   is   sometimes  

termed  as  ‘trueness’.  Precision  of  measurement,  defined  by  repeatability  and  

reproducibility,   is   the   degree   to  which   a   quantity   can   be  measured   under  

unchanged   conditions   and   show   the   same   results.   A  measurement   system  

can  be  accurate  but  not  precise,  precise  but  not  accurate,  but  ideally  is  both.  

1.4.2 Accuracy  of  Profilometry  

As   stated   above,   accuracy   is   equivalent   to   ‘trueness’,   or   how  well  what   is  

being   measured   matches   a   ‘true’   object.   Most   investigations   into   the  

accuracy   of   profilometers   used   in   dentistry   pertain   to   the   marginal   fit   of  

prosthetic  restorations  (Brandestini  et  al.  1985,  May  et  al.  1998,  Denissen  et  

al.  1999,  2000,  Groten  et  al.  1997,  Tinschert  et  al.  2000).  Most  studies  report  

inaccuracy  of  fit  to  be  between  50  to  75  μm,  which  is  well  within  the  range  

of   the   100   μm   error   advocated   for   clinical   production   (McLean   1971,  

Karlsson   1993)   and   generally   considered   to   be   an   acceptable   result.  

However,  any  conclusions  drawn  for  these  studies  take  into  account  errors  

that  may   exist   in   the  manufacturing   process   and   as   such   are   not   directly  

applicable  to  the  present  study.    

Vlaar   et   al.   (2006)  note   that   an   international   standard  exists   for   assessing  

‘Digitizing  quality’  of  profilometers  but  that  the  test  methods  are   laborious  

and   do   not   take   into   account   the   geometries   and   undercut  measurements  

encountered  in  scanning  dental  surfaces.  The  authors  set  up  a  study  in  order  

to   find  a   suitable  artifact   to   serve  as  a  dental   standard.  The  method   is   the  

Page 20: An in vitro investigation into the accuracy of CAD/CAM digitizing

18    

same   as   that   described   in   section   1.4.2.2   of   this   present   study   and   the  

authors   found   the  accuracy  of   two  digitizing   systems   to  be   “adequate”  but  

there   is   no   comment   made   as   to   this   method’s   further   use   as   a   ‘dental  

standard.’  

Vlaar  et  al.  (2006)  and  Persson  et  al.  (2006)  have  both  shown  that  accuracy  

in  dental  profilometers  depends  on  the  density  of  the  point  cloud  recorded.  

This  partly  helps  to  explain  why  profilometers  encounter  differences  when  

recording   areas   with   strong   changes   in   curvature   (Hewlett   et   al.   1992,  

Rudolph   et   al.   2006).   The   other   reason   that   a   strong   change   of   curvature  

causes   problems   in   measurement   is   given   by   the   numerical   aperture,   as  

described  in  section  1.2.4.  

One   of   the   only   studies   to   consider   precision   (i.e.   repeatability   and  

reproducibility)  as  well  as  accuracy  of  dental  profilometers  was  conducted  

by  Persson  et  al.  (2008).  The  authors  investigated  the  precision  of  Procera,  

which  is  a  contact  profilometer  for  the  CAD/CAM  of  crowns  and  one  of  the  

first   commercial   systems   available.   The   aim   was   to   determine   the  

reproducibility   of   digitized   dental   stone   replicas   compared   to   the   master  

model,   and   the   reliability   of   computer-­‐aided   analysis.   The   reliability   of  

computer-­‐aided  analysis  was  expressed  in  terms  of  accuracy  and  precision.  

Accuracy  of  measurement,  which  was  the  mean  values  of  the  discrepancies,  

was  shown  to  be  between  0.2  and  0.8  μm.  The  repeatability  coefficient   for  

the  measurement  method  was  shown  to  be  between  7  and  16  μm,  whereas  

analysis  of  the  stone  replicas  showed  a  repeatability  coefficient  of  19  to  26  

μm.  However  it  must  be  noted  that  only  6  measurements  were  taken,  which  

Page 21: An in vitro investigation into the accuracy of CAD/CAM digitizing

19    

is  well  below  the  30  measurements  recommended  by  BS  EN  ISO  12836.  The  

authors   felt   that   “the   reliability   of   the   computer   evaluation   method   has  

shown  to  be  both  accurate  and  precise.”  

1.4.2.1 Reference  Profilometer  

The   reference   measurement   system   was   a   previously   calibrated   non-­‐  

contacting   laser  profilometer  (Xyris  2000TL  NCLP,  Taicaan®  Technologies-­‐  

Southampton,  UK).  This  profilometer  consists  of  a  785  nm  wavelength  laser  

triangulation  sensor  which  has  a  spot  diameter  of  30  μm  a  z  resolution  of  0.1  

μm  and  an  angular  tolerance  of  90°.  The  sensor  was  mounted  onto  a  z  stage  

that  allows  vertical  movement  with  a  focal  depth  of  15  mm.  The  object  to  be  

measured  was  placed  underneath  the  sensor  and  an  x  and  y  stage  was  used  

to   scan  backwards  and   forwards   in  a   raster  pattern  whilst   the  data  points  

are  being  captured.  The  measurement  control  software  was  used  to  capture  

a  data  point  every  50  μm  in  the  x  and  y  direction  and  each  scan  of  a   tooth  

took  45  minutes.  The  scanner  therefore  records  3-­‐D  data  points  that  form  a  

character  encoded  (ASCII)  point  cloud  data  file  which  can  be  converted  into  

a  polygon  mesh.    

The  system  has  been  previously  calibrated  (Rodriguez  et  al.  2012)  and  the  

accuracy  and   repeatability  were  1.3  μm  and  1.6  μm  respectively,   at   a   step  

over  distance  of  50  μm.  

1.4.2.2 Test  Device  1:  3Shape®  Extra-­‐oral  Scanner  

The  3Shape®   extra-­‐oral   scanner  has   a   red   laser   sensor   and   scans   a  dental  

model   as   it   rotates   on   a   360°   spinning   stage.   Two   charge-­‐coupled   device  

Page 22: An in vitro investigation into the accuracy of CAD/CAM digitizing

20    

cameras  with  1.3  Mega  Pixels  capture  the  scan  data  over  the  course  of  25  to  

80  seconds  and  the  point  cloud  data  is  then  converted  into  a  3-­‐D  data  format  

for   export   as   an   .stl   file.   The   stated   accuracy   of   the   system   is   10   μm   for  

crowns   and   bridges   (3Shape®   Dental   Systems   Brochure,   2014).   The  

3Shape®  extra-­‐oral   scanner   is  designed   to  scan  gypsum  models  poured  up  

from   impressions   or   the   impression   directly   and   does   not   require   any  

surface  modification  of  the  models  or  the  impressions.    

As  mentioned  above,  Vlaar  et  al  (2006)  set  up  a  study  in  order  to  find  a  value  

for  the  measurement  error  of  scanned  dental  surfaces  and  to  try  and  find  an  

artifact  to  serve  as  a  dental  standard  for  profilometers.  A  sphere  of  known  

radius  was  chosen  because  it  “has,  like  a  dental  preparation,  a  continuously  

changing  surface  and  is  therefore  the  perfect  object  to quantify  the  undercut  

as  a  gauge  of  how  far  a  scanner  can  measure  steep  walls  and  undercuts.” The

sphere was sprayed  with   fine  titanium  dioxide  powder  and  scanned  from  8  

different   angles   by   two   different   profilometers,   one   of   which   was   a   early  

model   of   the   3Shape®   extra-­‐oral   scanner.   The   recorded   data   was   then  

entered  into  software  in  order  to  calculate  a  ‘best  fit’  sphere  and  compared  

to   the   actual   artifact   of   known   dimension.   Both   machines   showed  

“adequate”  accuracy  (7.7  ±  0.8  μm  and  13.9  ±  1.0  μm).  

Sousa  et  al.  (2012)  set  up  a  study  to  investigate  accuracy  and  reproducibility  

of   three-­‐dimensional   digital  model  measurements.  Again,   an   earlier  model  

of  the  3Shape®  extra-­‐oral  scanner  was  used  to  scan  20  dental  casts.  Fifteen  

anatomical  dental  points  were  identified  and  11  linear  measurements  made.  

The   digital   models   were   analysed   using   software,   while   the   casts   were  

Page 23: An in vitro investigation into the accuracy of CAD/CAM digitizing

21    

analysed  with  calipers  and  then  the  two  measurement  modalities  compared.  

No   statistical   differences   were   found   between   the   two   measurement  

systems  but  the  authors  consider  linear  measurements  of  digital  models  to  

be  more  accurate  and  reproducible.    

1.4.2.3 Intra-­‐oral  Scanners  

The  3M™  True  Definition   intra-­‐oral   scanner   is   based   on   “active  wavefront  

sampling”   technology   and   can   capture   3-­‐D   images   directly   of   teeth   and  

periodontal  tissues  in  vivo  (van  der  Meer  et  al.  2012).  Light  is  emitted  from  a  

source  within  the  system  and  is  transferred  down  a  cable  to  a  wand  that  has  

similar   dimensions   to   a   dental   handpiece.   The   wand   is   inserted   into   the  

patient’s   mouth   with   a   stand-­‐off   distance   between   5   and   20   mm.   Before  

scanning   the   surface   needs   to   be   lightly   dusted   with   Lava   Powder   (3M™  

ESPE,  St.  Paul,  USA),  which  is  a  titanium  oxide  powder  required  in  order  to  

provide   an   optically   active   surface   for   the   registration   of   the   3-­‐D   patches  

obtained  during  scanning.  The  3-­‐D  positioning  of  the  surface  is  recorded  in  a  

live  video   feed   format  as   the   teeth  are  scanned   in  a   ‘zigzag’  pattern  across  

the  dental  arch.  The  scanner  captures  twenty  3-­‐D  frames  per  second.  During  

the   scanning  procedure,  patented  software   stitches   the  3-­‐D  scans   together  

in  real  time  to  allow  for  continuous  data  capture  up  to  a  total  scanning  time  

of  seven  minutes  per  arch.  A  post-­‐scanning  processing  cycle  recalculates  the  

registration  and  compensates   for  potential  errors  and   the  3-­‐D  data   is   then  

uploaded   to  3M™   for  processing   into  an   .stl   file   that   can   then  be  accessed.  

3M™  have  not  published  any  accuracy  data   for   the  digitizing  of  crown  and  

bridge   shaped   artifacts,   however   the   smallest   mean   distance   error   of   the  

Page 24: An in vitro investigation into the accuracy of CAD/CAM digitizing

22    

precursor  to  the  True  Definition  scanner,  the  Lava  COS  was  measured  using  

an  implant  bar  model  to  be  2.2  μm  (van  der  Meer  et  al.  2012).    

Giménez  et  al.  (2013)  also  investigated  the  3M™  Lava  COS  scanner  in  order  

to   assess   its   accuracy   in   scanning   dental   implants.   A   master   model  

containing  6  implants  was  scanned  with  a  contacting  profilometer  and  this  

data   served   as   ‘true’   value.   The   authors   found   that   measurements   for  

angulation  and  depth  of  implant  did  not  differ  significantly  from  ‘true’  value.  

However,  the  experience  of  the  operator  was  found  to  be  a  significant  factor  

and  training  for  digital  impression  capture  was  advocated.  

Measuring  Tooth  Wear  

In   a   review   of   tooth   wear   indices   Bardsley   (2007)   notes   there   is   “both   a  

clinical   and   a   scientific   need   to   be   able   to   measure   tooth   wear,   and   the  

literature   abounds  with  many  methods  which   can  be  broadly  divided   into  

quantitative   and   qualitative   methods.”   This   review   notes   that   qualitative  

measurements   can   be   valuable   measurement   tools   but   they   are   also  

subjective  and  require  appropriate   training.   In   the  general  practice  setting  

this  may  present  a  problem  where  patients  do  not  necessarily  get  continuity  

in  the  dentist  at  the  practice  that  they  attend.  Quantitative  assessment  relies  

on  objective  physical  measurement  and   in   the  case  of   tooth  wear   this  may  

be   something   like   the   size   of   a   wear   facet   or   the   height   of   a   crown.   The  

advantages  of  the  quantitative  measurement  are  that  it  is  less  subjective  and  

that  wear  can  be  measured  over  time,  meaning  an  overall  rate  of  wear  may  

be  determined.  The  disadvantages   in  the  past  have   largely  been  due  to  the  

fact   that   impressions  are   required   from  which   casts   are  produced  and   the  

Page 25: An in vitro investigation into the accuracy of CAD/CAM digitizing

23    

measurement   taken;   these   study   casts   then   also   require   storage.  Typically  

quantitative  measurement  has  been  recorded  by  looking  at  change  in  crown  

height,  but  using  profilometry  and  computer  software  it   is  now  possible  to  

record  a  volumetric  measurement.  

1.4.3 Qualitative  Measurement  of  Tooth  Wear  

There   exist   several   indices   that   provide   qualitative  measurement   of   tooth  

wear,   mostly   developed   or   modified   from   the   Smith   and   Knight   index  

(1984).  These  are  mostly  qualitative  and  descriptive,  and  generally  relate  to  

the   involvement  of  dentine   in   the   lesion.   In  most   instances  calibration  and  

training  in  the  use  of  these  indices  is  required.  One  of  the  most  prevalent  of  

these  is  the  Basic  Erosive  Wear  Examination  (Bartlett  et  al.  2008)  This  is  a  

partial   scoring   system   recording   the   most   severely   affected   surface   in   a  

sextant   and   the   cumulative   score   guides   the  management  of   the   condition  

for  the  practitioner.  The  four  level  score  grades  the  appearance  or  severity  

of  wear  on  the  teeth  from  no  surface  loss  (0),   initial   loss  of  enamel  surface  

texture   (1),  distinct  defect,  hard   tissue   loss   (dentine)   less   than  50%  of   the  

surface  area  (2)  or  hard  tissue  loss  more  than  50%  of  the  surface  area  (3).  

While   it   is  a  useful  screening  tool   it  should  be  noted  that  a  study  by  Ganss  

and   Lussi   (2005)   examined   teeth   both   clinically   and   histologically   for   the  

presence  of  exposed  dentine  and  found  that  the  correlation  between  the  two  

examinations   was   poor.   It   is   this   subjective   nature   of   qualitative  

measurement   that   necessitates   the   need   for   a   simple,   clinical   method   of  

measuring  tooth  tissue  loss.    

Page 26: An in vitro investigation into the accuracy of CAD/CAM digitizing

24    

1.4.4 Traditional  Quantitative  Measurement  of  Tooth  Wear  

Traditional  metrological  measurement  of  tooth  wear  has  always  required  a  

reference   point.   This   is   obviously   challenging   in   the   oral   environment  

because   it   is   living   tissue   and   is   subject   to   change.   Some   of   the   methods  

employed   to   overcome   this   problem   include   indentations   cut   into   enamel  

(Lambrechts   1984)   and   cementation   of  metal   discs   to   less   than   10%   of   a  

tooth’s   surface   (Bartlett   at   al.   1997).   Neither   method   can   be   considered  

satisfactory   for   reasons   of   ethics,   retention   of   the   metal   discs   and  

reproducibility.   As   previously   described,   Atkinson   et   al.   (1982)   made   an  

early   attempt   at   quantitative  measurement   of  wear   in   dental   restorations  

using   laser   dual   source   contouring   using   polyvinylsiloxane   impressions   to  

provide  information  about  the  topology  of  the  restorative  surfaces  pre-­‐  and  

post-­‐test   by   contour  mapping.   This   was   a   time   consuming   but   innovative  

attempt  at  employing  the  new  technology.  

Molnar  et  al.  (1983)  measured  wear  rate  by  using  casts  taken  at  intervals  of  

64   Aboriginal   children   (from   the   ages   of   seven   and   up)   during   a   growth  

study   and   then   measuring   the   cusp   heights   of   the   first   and   second  

permanent  molars.  Photos  of  casts  were  traced  using  electronic  planimetric  

methods  that  automatically  recorded  the  size  and  location  of  wear  facets.  By  

18  years  of  age  an  average  of  0.5  mm  of  molar  cusp  height  had  been  lost,  a  

wear  rate  of  0.41  μm  per  year.  In  general  they  found  wear  rates  to  be  greater  

than   Europeans   and   put   this   difference   in   wear   rate   down   to   quantity   of  

dietary  abrasives.  

Page 27: An in vitro investigation into the accuracy of CAD/CAM digitizing

25    

Glentworth  et  al.   (1984)  used  a  147Pm  β  particle  backscatter   instrument   to  

measure  changes  in  surface  profiles  of  dental  restorations  due  to  wear.  They  

found  this  instrument  to  be  capable  of  measuring  changes  of  the  order  of  15  

μm  in  the  profile  of  irregularly  shaped  surfaces.  

Lambrechts  et  al.  (1989)  carried  out  a  quantitative  in  vivo  study  of  wear  in  

human   enamel.   Twenty-­‐one   subjects   were   selected   with   the   inclusion  

criteria  being  possession  of  a  complete  dentition  and  ‘normal’  occlusion.  The  

mean  age  of  the  subjects  was  20  years  and  there  was  no  gender  segregation.  

Polyvinylsiloxane   impressions   were   taken   of   molars   and   premolars   at  

baseline,  6,  12,  18,  24,  36  and  48  months  and  then  these  were  converted  into  

accurate   positive   replicas   using   a   copper   plating   technique.   Intraoral  

photographs,   articulation   paper   and   stereomicroscopic   evaluation   of   the  

replicas   were   used   to   identify   attrition   sites   and   the   maximum   loss   of  

substance   at   these   sites   was   then   measured   using   a   computerised   three-­‐

dimensional   measuring   microscope.   Forty-­‐eight   occlusal   contact   areas   on  

premolars  and  49  occlusal  contact  areas  on  molars  were  measured  and  the  

means   and   standard   deviations   of   vertical   loss   due   to   wear   determined.  

After  four  years  the  total  enamel  wear  was  153  μm  on  molars  and  88  μm  on  

premolars.  This  translated  to  a  steady  wear  rate  of  about  29  μm  per  year  for  

molars   and   about   15   μm   per   year   for   premolars.   The   advantage   of   the  

copper   plating   technique   was   that   it   removed   the   significant   variation   in  

thickness  caused  by  applying  a  surface  coating  (e.g.  gold  leaf)  to  an  already  

formed   replica.   The   disadvantages   of   electroplating   are   the   need   for  

specialised  equipment  and  the  time  taken  for  the  electroplating  process.    

Page 28: An in vitro investigation into the accuracy of CAD/CAM digitizing

26    

1.4.5 Profilometry  and  Tribology  Studies  

In  a  study  into  using  mapping  apparatus  for  measuring  tooth  wear  over  time  

Chadwick   et   al.   (1995)   noted,   “The   subjective   ranking   scales   hitherto  

employed   in   [such]   epidemiological   studies   are   insufficiently   sensitive   to  

detect   small   amounts   of   tooth   wear   that,   in   themselves,   may   be   of   little  

significance  but  cumulatively  may  result  in  significant  loss  of  tooth  surface.  

The   ability   to   detect   such   increments   would   be   of   value   when   clinically  

following  up  patients  with  tooth  wear.”  The  present  study  aims  to  eliminate  

this   subjectivity   by   investigating   the   ability   of   dental   profilometers   to  

provide  an  accurate  volumetric  measurement  of  tooth  substance  loss.  

Indeed  the  use  of  profilometry  is  not  a  particularly  new  concept.  As  far  back  

as   the   early   1990’s   Noordmans   et   al.   (1991)   were   able   to   use   computer-­‐

aided   profilometry   to   assess   the   abrasive   wear   of   human   enamel   and  

dentine.   Freshly   extracted   incisors   were   prepared   and   scanned   using   a  

contact  profilometer   (Perthometer  C5D,  Perthen  GmbH)   to   give   a  baseline  

reading   of   surface   roughness.   These   samples   were   then   brushed   over   a  

period  of  time,  both   in  vivo  and   in  vitro.  Measurements  were  taken  every  7  

days   for  7  weeks.   In  vivo  wear  rates   in   the  experiment  were  0.2   to  3.0  μm  

per  week  for  enamel,  depending  on  the  toothpaste  used  and  the  subject.  For  

dentine   these   values   were   4.0   to   35.0   μm   per   week.   Attin   et   al.   (2001)  

performed   a   similar   in   situ   study   in   order   to   assess   the   susceptibility   of  

previously  demineralised  enamel  to  tooth  brushing  action.  Enamel  samples  

were   analysed   using   a   laser   profilometer   and   mean   (standard   deviation)  

wear   of   the   samples  was   found   to   be   between  0.66  ±   1.11  μm  and  6.78  ±  

Page 29: An in vitro investigation into the accuracy of CAD/CAM digitizing

27    

2.71  μm  depending  on  the  time  between  the  demineralization  and  the  tooth  

brushing.   The   authors   were   able   to   conclude   from   this   that   at   least   60  

minutes   should   pass   before   tooth   brushing   after   presenting   teeth  with   an  

erosive  challenge.  

In  another  study,  Bartlett  et  al.   (1997)  cemented  metal  discs   to   the  palatal  

tooth   surface   of   20   patients   (13   with   unexplained   palatal   erosion   and   7  

controls  without  evidence  of  tooth  wear)  in  order  to  act  as  reference  areas.  

Wear  was   then   estimated   by   taking   impressions   at   6  month   intervals   and  

scanning   them   with   a   contacting   profilometer.   A   statistically   significant  

difference  between   the   two  groups  of  patients  was   found,  with  patients   in  

the  tooth  wear  category  showing  a  median  36.5  μm  of  wear  over  6  months  

in   comparison   to   a  median  of   3.7   μm   in   the   control   group.  Two   follow  up  

studies  have  been  carried  out  (Azzopardi  et  al.  2001,  Sundaram  et  al.  2007)  

in  which  teeth  were  coated  with  resin-­‐based  dentine  bonding  agent  to  see  if  

it   provided   any   protection   from   erosive   and/or   abrasive   tooth   wear.  

Impressions  were   taken  and  scanned  using  a   laser  profilometer,   and  were  

able  to  show  a  statistical  difference  between  the  rate  of  wear  in  the  control  

group   and   those   protected   by   dentine   bonding   agent.   This   use   of  

profilometry  demonstrated  potential  for  dentine  bonding  agents  to  provide  

up   to   3   months   of   protection   from   tooth   wear.   In   their   study   into   the  

prevention   of   erosion   and   abrasion   Azzopardi   et   al.   (2001)   note   in   the  

discussion,  “Digital  terrain  modelling,  which  relies  on  the  premise  that  acid  

does   not   affect   the   whole   area   under   investigation   equally,   also   shows  

particular  promise.”  

Page 30: An in vitro investigation into the accuracy of CAD/CAM digitizing

28    

The  reference  profilometer  used  in  the  present  study  was  a  non-­‐contacting  

laser   profilometer   (NCLP-­‐Taicaan®   Technologies-­‐Southampton)   that   has  

been   used   in   previous   studies   by   Rodriguez   et   al.   (2009,   2010).   In   the  

former  study  the  same  profilometer  was  used  to  measure  surface  accuracy  

of   impression   materials   and   dental   stones.   The   system   works   by   using   a  

triangulation   laser   sensor   that   detects   the   deflection   of   a   laser   spot   on   a  

charge   coupled   device   camera.   This   NCLP   has   a   785nm  wavelength   laser  

triangulation  sensor  with  a  spot  diameter  of  30  μm  and  an  axis  and  sensor  

resolution   of   0.1   μm.   Its   accuracy   in   measuring   surface   roughness   was  

tested   by   scanning   an   impression   made   of   a   6   μm   roughness   standard  

(Taylor   Hobson-­‐Reference   Specimen   Type   112/1534).   The   mean   and  

standard  deviation  of  the  roughness  value  from  this  impression  of  the  6  μm  

roughness   standard  was   found   to  be  5.99  μm  ±  0.29,   a  mean  difference  of  

0.01  μm,   indicating  excellent  accuracy   in  recording  surface  roughness.  The  

authors   do  note,   “the  NCLP  was   capable   of   detecting   vertical   features   less  

than   the   laser   spot   diameter   because   the   vertical   resolution   of   the   sensor  

was  0.01  μm.”  In  a  later  study  the  accuracy  of  measurements  of  tooth  wear  

by   the   same   NCLP,   in   combination   with   surface   matching   software   was  

assessed.  Accuracy  and  repeatability  were  assessed  by  repeatedly  scanning  

a   calibrated   25   mm   engineering   steel   gauge   block.   Scanning   titanium  

frustums   of   varying   volumes   made   volumetric   assessments   possible.  

Accuracy   and   repeatability   of   the   systems   in   measuring   step   height   and  

volume  after   surface  matching  were  measured  using  a   custom  model  with  

cemented  engineering  slip  gauges  and  cemented  onlays  of  super-­‐plastically  

formed  titanium.  The  accuracy  and  repeatability  were  1.3  μm  and  1.6  μm  in  

Page 31: An in vitro investigation into the accuracy of CAD/CAM digitizing

29    

measuring  length  and  the  system  was  accurate  for  volumetric  measurement  

with   coefficients   of   variation   <5%.   It   is   this   accuracy   of   volumetric  

measurement   that   can   potentially   be   applied   to   measurement   of   tooth  

substance  loss  in  cases  of  tooth  wear.  

The   same   authors   (Rodriguez   et   al.   2012)   carried   out   an   in   vivo  

measurement  of  tooth  wear  of  63  patients.  Maximum  follow  up  time  was  12  

months  for  only  30  of  the  patients  but  all  63  had  a  maximum  follow  up  time  

of   6  months.  A  questionnaire  was   also  used   in   order   to   assess   tooth  wear  

risk   factors   in   the  participants   and   silicone   impressions  were   taken  of   the  

patients   at   baseline   and   subsequently   every   6   months.   The   impressions  

were   poured   in   type   IV   gypsum   and   scanned   using   the   same   reference  

profilometer   being   used   in   the   present   study.   Images   were   then  

superimposed   using   computer   software   in   order   to   assess   the   amount   of  

tooth  wear  in  each  subject.  Measurement  error  was  stated  as  15  μm.  Of  the  

1078  teeth  measured  only  72.2%  of  showed  wear  <15  μm  over  a  6  month  

period,  while  of  the  63  patients  participating  77.7%  showed  a  median  tooth  

wear   <15   μm.   There   was   a   statistical   trend   associated   with   gastric   risk  

factors.  

Three-­‐dimensional  measurement  of  tooth  wear  was  previously  described  by  

Mehl   et   al.   1997   where   the   authors   investigated   the   accuracy,   with   and  

without  referenced  positioning,  of  a  3-­‐D  optical  scanner  and  reference  free  

automated  superimposition  software.  Crucially  they  were  able  to  show  that  

accuracy  depends  on  surface  inclination  of  the  object  being  scanned,  as  per  

the   limitation   of   numerical   aperture.   Up   to   an   angle   of   60   degrees   the  

Page 32: An in vitro investigation into the accuracy of CAD/CAM digitizing

30    

precision   was   better   than   3   μm   and   the   accuracy   better   than   6   μm.    

However,   the   method   described   does   assume   that   the   three-­‐dimensional  

calculations   made   by   the   software   are   perfectly   accurate   which   may   not  

have  been  the  case,  and  in  addition  no  actual  volumetric  measurement  was  

described.  

Pintado  et  al.  (1997)  demonstrated  annual  tooth  wear  rates  in  a  group  of  18  

dental  students  to  be  0.04  mm3.  This  was  assessed  over  two  years  by  taking  

polyvinyl   silicone   impressions,   making   epoxy   resin   models   and   then  

scanning   them  with  a   contacting  profilometer.  They  also   found   that   for  all  

angles  less  than  60°  the  accuracy  was  7  μm  and  the  precision  was  5  μm,  in  

close   agreement   to  Mehl.   A  more   recent   study   by  Tantbirojn   et   al.   (2012)  

looked   at   12  patients  with   gastro-­‐esophageal   reflux  disease   (GERD)   and  6  

control  subjects.  Impressions  were  taken  at  baseline  and  at  6  months,  cast  in  

dental   stone   and   scanned   with   an   optical   scanner.   The   mean   (standard  

deviation)   volume   loss   per   tooth   in   the   participants   was   found   to   be  

significantly  higher  than  that  in  control  participants:  0.18  ±  0.12  mm3  versus  

0.06   ±   0.03  mm3.   This  measurement  was   calculated   by   overlapping   scans  

using   computer   software,   a   technique   that   the   authors   provide   a   mean  

(standard  deviation)  accuracy  of  0.006  ±  0.001  mm3.  

Currently   in   dental   literature   there   exist   no   studies   that   investigate  

volumetric  measurement  of  tooth  wear  using  an  intra-­‐oral  scanner.  With  the  

an   increasing   plethora   of   digital   scanners   now   available   to   the   dental  

community,   there   is  a  need  for  greater  understanding  of  the  capabilities  of  

Page 33: An in vitro investigation into the accuracy of CAD/CAM digitizing

31    

these  systems  for  both  diagnostic  and  treatment  purposes,  all  of  which  rely  

of  the  fundamental  concept  of  accuracy  of  measurement.  

Page 34: An in vitro investigation into the accuracy of CAD/CAM digitizing

32    

Aims  and  Objectives  

2.1 Aims  

The  overall  aim  was  to  investigate  the  accuracy  of  intra-­‐oral  and  laboratory  

profilometers  for  the  measurement  of  tooth  wear.    

The   investigation  was   comprised  of   three   individual   experiments  with   the  

following  specific  aims:    

1. To   determine   the   resolution   of   the   digitizing   devices   at   measuring  

objects  of  known  dimensions.  

2. To   assess   the   reproducibility,   repeatability   and   trueness   of  

measurement   of   the   digitizing   devices,   with   comparison   to   a  

reference  measurement  system.    

3. To   compare   the   performance   of   the   digitizing   devices   using   a  

previously   validated   profilometric   tooth   wear   measurement  

technique,  using  an  in  vitro  calibration  model.  

2.2 Objectives  

1. To  measure  engineering  slip  gauges  of  known  dimensions  in  order  to  

determine   the   minimum   resolution   of   the   digitizing   devices,   in  

comparison  to  a  calibrated  triangulation  laser  profilometer.  

2. To   determine   the   reproducibility,   repeatability   and   trueness   of   the  

digitizing  devices  in  the  measurement  of  an  inlay-­‐shaped  artifact  and  

a   crown-­‐and-­‐bridge-­‐shaped   artifact,   in   comparison   to   a   calibrated  

triangulation  laser  profilometer.  

Page 35: An in vitro investigation into the accuracy of CAD/CAM digitizing

33    

3. To   determine   ‘proof   of   concept’   for   the   accuracy   of   the   digitizing  

devices  in  the  measurement  of  calibration  models  previously  used  to  

evaluate   a   method   of   profilometric   tooth   wear   measurement,   in  

comparison  to  a  calibrated  triangulation  laser  profilometer.  

   

Page 36: An in vitro investigation into the accuracy of CAD/CAM digitizing

34    

Null  Hypotheses  

The   digitizing   devices   will   not   be   able   to   measure   the   engineering   slip  

gauges.  

There   is   no   difference   in   the   accuracy   of   measurement   between   the  

digitizing  devices  and  the  triangulation  laser  profilometer.  

There   is   no   difference   in   the   capabilities   of   the   digitizing   devices   and   the  

triangulation  laser  profilometer  for  the  measurement  of  calibration  models  

used  to  evaluate  a  profilometric  tooth  wear  measurement  method.  

   

Page 37: An in vitro investigation into the accuracy of CAD/CAM digitizing

35    

Materials  and  Methods  

Preliminary   discussions   with   the   Freeform   Measurement   Centre   at   the  

National   Physical   Laboratory  were   entered   into   regarding   the   procedures  

required   for   fully   following   the  measurement   protocol   specified   by  BS   EN  

ISO   12836   (2012).   However,   during   these   discussions   it   became   apparent  

that   the   cost  of   fabricating   and   calibrating   the   artifacts   required  would  be  

extremely   costly   and   moreover   the   dimensions   of   the   specified   artifacts  

were  much  greater  than  that  would  be  likely  to  be  seen  in  a  natural  human  

dentition.  

There  are  many  different  protocols  currently  existing  for  the  assessment  of  

accuracy   of   intra-­‐oral   scanners,   many   of   which   are   based   on   the  

measurement   of   a   full   arch   implant   model   with   6   implants   evenly  

distributed   across   the   arch   (van  der  Meer   et   al.   2012).   This  model   has   its  

merits   but   for   the   measurement   of   tooth   wear   more   local   accuracy   is  

required   and   therefore   the   ISO   bridge   and   inlay   specimen   (BS   EN   ISO  

12836:2012,  2012)  and   the   tooth  wear  calibration  model   (Rodriguez  et  al.  

2012)  have  more  relevance  and  so  were  employed  here.  

   

Page 38: An in vitro investigation into the accuracy of CAD/CAM digitizing

36    

 

3.1 Test  Digitizing  Devices  

The  test  digitizing  devices  investigated  in  this  study  were  the  3Shape®  extra-­‐

oral   scanner   (D700,   3Shape®   Copenhagen,   Denmark)   and   the   3M™   True  

Definition  intra-­‐oral  scanner  (3M™  True  Definition  intra-­‐oral  scanner,  3M™  

ESPE®  St.  Paul,  MN,  USA).  

3.2 Test  Conditions  for  Measurement  

The   specific   test   conditions   for   measurement   were   carefully   controlled,  

including  monitoring  of  the  change  in  temperature  during  the  test  to  ensure  

that  it  remained  within  ±1oC,  and  this  was  measured  using  an  EL-­‐USB-­‐2  data  

logger  (Lascar  Electronics®  -­‐  Salisbury,  UK).  The  ambient  room  temperature  

was   maintained   at   23   ±   2   °C   and   the   quality   of   the   data   set   in   terms   of  

missing  or   corrupted  data  were   continuously   evaluated   and   in   the   case   of  

missing  or  corrupted  data  the  test  was  repeated.    

3.3 Measurement  Software  

All   measurements   were   carried   out   using   Geomagic®   Qualify   surface  

matching  software  (Geomagic®  Qualify  11  –  Geomagic®  Incorporated,  North  

Carolina,  USA),  following  measurement  workflows  as  defined  below  for  each  

measurement  specimen.    

Page 39: An in vitro investigation into the accuracy of CAD/CAM digitizing

37    

3.4 Specimen  Preparation  

Three  types  of  artifacts  were  prepared  for  measurement  using  the  digitizing  

devices.    

3.5 Engineering  Slip  Gauge  Specimen  

Ten   engineering   slip   gauges   (Laser   Tools®,   Southampton,   UK)   of   varying  

thicknesses  from  50  μm  to  800  μm  were  selected  in  order  to  determine  the  

smallest  possible  item  that  could  be  measured.    

In  order   for   the  dental  digitizing  devices   to  be  able   to  scan  the  data   it  was  

necessary  to  mount  the  gauges  in  a  base  that  was  shaped  like  a  dental  arch,  

as  the  scanning  devices  were  designed  to  recognised  dental  jaw  shapes  and  

initial  scanning  failures  had  identified  this  as  an  issue  when  scanning  a  free  

standing  object  that  was  not  tooth  shaped.    

In  order  to  make  the  model  as  clinically  relevant  as  possible,  thirteen  human  

teeth   were   collected   according   to   research   ethics   (REC   reference:  

12/LO/1836)  and  the   teeth  were   then  set  up   in  Moonstone™  (Bracon  Ltd.,  

Etchingham,  England)  in  a  die  tray  (Sterdo  Split  model  tray,  Zirc  Company,  

Buffalo,  MN).  Condensation   cured   laboratory  putty   (Zetalabor®,   Zhermack,  

Rovigo,   Italy)  was  used   to   replicate   the  periodontal   tissues  with   the  putty  

trimmed  to  the  cemento-­‐enamel  junction  around  all  the  extracted  teeth.  This  

therefore   simulated   a   partially   dentate  mandibular   human   arch  with   only  

the  lower  right  first  molar  missing,  as  shown  in  Figure  1  below.  

Page 40: An in vitro investigation into the accuracy of CAD/CAM digitizing

38    

 

Figure  1  -­‐  Extracted  teeth  set  up  to  simulate  a  partially  dentate  mandibular  arch  

The  slip  gauges  were  then  mounted  individually  between  the  teeth  and  were  

fixed  in  place  with  laboratory  putty.  As  the  surface  of  the  slip  gauges  was  a  

shiny  metal,   in   order   for   all   scanners   to   be   able   to   scan   the   gauges   it  was  

first   necessary   to   apply   a   thin   coating   of   50   nm   Telescan   carbon   coating  

(DFS®   Diamon   GbmH   –   Riedenburg,   Germany).   In   order   to   control   film  

thickness   the  gauges  were  sprayed   from  a  distance  of  30  cm  for  no   longer  

than  2  seconds.    

Each   thickness   of   slip   gauge   was   then   scanned   with   the   three   scanners  

following  measurement  protocols  defined.  The  scan  data  was  then  imported  

into  Geomagic®  Qualify  11  for  measurement  of  the  thickness  of  the  edge  of  

the  slip  gauges.    

Page 41: An in vitro investigation into the accuracy of CAD/CAM digitizing

39    

For  the  measurement  of  the  slip  gauges  the  measurement  work  flow  was  as  

follows.  The  point  cloud  data  was  converted  into  a  polygon  mesh  of  the  area  

of  the  slip  gauge  and  the  surrounding  teeth  as  seen  in  Figure  2  below.  

 

Figure  2  -­‐  Digital  model  with  slip  gauge  in  situ  

The  scan  was  cleaned  up  such  that  only  the  scan  data  from  the  top  edge  of  

the  slip  gauge  alone  was  considered.  This  resulted  in  an  image  as  shown  in  

Figure  3  below.    

 

Figure  3  -­‐  Digital  model  of  a  slip  gauge  thickness  

Page 42: An in vitro investigation into the accuracy of CAD/CAM digitizing

40    

In  order  to  ensure  that  the  thickness  of  the  slip  gauges  was  measured  in  the  

same  plane  of  orientation   for   the   three  measurement  devices,  a   three-­‐step  

process   was   carried   out   within   Geomagic®   Qualify.   Firstly   a   best-­‐fit  

alignment   was   carried   out   in   order   to   ensure   that   the   scans   of   the   slip  

gauges  were  all  orientated  in  the  same  3-­‐D  space.  Secondly  a  cross  section  of  

the  scans  was  made  using  a  plane  that  was  identical  for  all  the  slip  gauges  as  

shown  in  Figure  4  below.  

 

Figure  4  -­‐  Geomagic®  software  being  used  to  create  a  2-­‐D  outline  of  a  slip  gauge  thickness  

This  created  a  2-­‐D  outline  of  the  identical  part  of  the  slip  gauge  for  all  three  

measurement  devices.  The  final  part  of  the  measurement  was  to  carry  out  a  

measurement   of   the   2D   dimensions   of   the   parallel-­‐sided  walls   of   the   slip  

gauge  as  shown  in  Figure  5  below.  

Page 43: An in vitro investigation into the accuracy of CAD/CAM digitizing

41    

 

Figure  5  -­‐  Measurement  of  the  2-­‐D  outline  of  a  slip  gauge  thickness  

The   mean   and   standard   deviation   of   the   30   measurements   was   then  

calculated  for  the  three  digitizing  devices.    

3.6 Inlay  shaped  Specimen  and  Bridge  Shaped  Specimen  

The   measurement   protocol   specified   in   ISO   standard   12836   publication  

‘Dentistry   —   Digitizing   devices   for   CAD/CAM   systems   for   indirect   dental  

restorations  —  Test  methods  for  assessing  accuracy’  was  used  to  measure  an  

inlay  shaped  specimen  and  a  bridge  shaped  specimen  prepared  from  human  

teeth   that   had   been   mounted   in   a   simulation   of   an   mandibular   jaw   as  

described   above.   The   natural   teeth   were   restored   thus   providing   more  

clinically   relevant  artifacts   than   those   specified   in   the   ISO  standard,  which  

were  not  of  similar  dimensions  to  natural  teeth.  The  teeth  were  prepared  to  

provide   clinically   relevant   inlay   shaped   specimens   and   a   bridge   shaped  

specimen   consisting   both   of   natural   tooth   tissue   and   artificial   restorative  

materials   thus  providing  a   clinically   relevant   range  of  materials  with   their  

varying  optical  surface  properties.    

Page 44: An in vitro investigation into the accuracy of CAD/CAM digitizing

42    

To  prepare  the  teeth,  a  truncated  cone  shaped  diamond  bur  with  a  medium  

grit   and   a   head   diameter   of   2.0   mm   (Hi-­‐Di   No.625,   Dentsply®   Ash  

Instruments,  UK)  was  inserted  into  the  shank  of  a  milling  machine  (PFG  100,  

Cendres   &  Metaux   S.A,   Bienne,   Switzerland).   The   lower   left   second  molar  

was   prepared   for   an   inlay-­‐shaped   cavity,   roughly   following   the   shape  

specified  in  the  ISO  standard.  This  is  shown  below  in  Figure  6.    

 

Figure  6  –  Photograph  of  the  inlay  shaped  preparation  

Figure  7  below  shows  a  3D  comparison  of  the  inlay  shape  before  and  after  

preparation   showing   the   amount   of   tooth   tissue   that   was   removed.   Red  

equals  the  most  tissue  removed  and  green  equals  no  tissue  removed.  

Page 45: An in vitro investigation into the accuracy of CAD/CAM digitizing

43    

 

Figure  7  –  3D  comparison  of  the  shape  of  the  inlay  shape  preparation  before  and  after  

preparation  showing  the  amount  of  tooth  tissue  that  was  removed  for  the  preparation  (red  

equals  the  most  tissue  removed  and  green  equals  no  tissue  removed)  

The   three-­‐unit   fixed-­‐fixed   bridge   preparation   was   carried   out   following  

standard  preparation  guidelines  using  the  lower  right  second  premolar  and  

the   lower   right   second  molar   (with   the   lower   right   first  molar  missing)  as  

shown  in  Figure  8  below.  

Page 46: An in vitro investigation into the accuracy of CAD/CAM digitizing

44    

 

Figure  8  -­‐  Bridge  specimen  

Again,  a  thin  coating  of  Telescan  (DFS®  Diamon  GbmH  –  Riedenburg,  

Germany)  was  applied  prior  to  scanning  and  then  the  models  were  scanned  

with  a  similar  scanning  technique  as  previously  described  to  produce  3-­‐D  

images  as  shown  in  the  Figure  9  and  Figure  10  below.  

Page 47: An in vitro investigation into the accuracy of CAD/CAM digitizing

45    

 

Figure  9  -­‐  3-­‐D  digital  model  of  inlay  specimen  

 

Figure  10  -­‐  3-­‐D  model  of  bridge  specimen  

Following  the  ISO  standard  protocol,  the  repeatability,  reproducibility  and  

trueness  of  measurement  of  the  height  b,  in  mm,  of  the  preparations,  and  the  

angle  α,  in  °,  of  the  taper  of  opposing  walls  of  the  preparation  were  

quantified  as  described  below.  The  ISO  standard  specifies  the  measurement  

protocol  to  quantify  the  height  and  taper  of  the  inlay  specimen  and  the  

bridge  specimen  and  recommends  the  use  of  surface  analysis  software  such  

Page 48: An in vitro investigation into the accuracy of CAD/CAM digitizing

46    

as  Geomagic®  Qualify  to  carry  out  the  analyses.    

3.6.1.1 Height  and  Taper  of  Inlay  Specimen  

As  shown  in  Figure  11  the  angle  α  and  a  height  b  of  the  inlay  shaped  cavity  

were  measured  using  Geomagic®  Qualify.  The  same  workflow  as  described  

above  for  measuring  the  thickness  of  the  slip  gauges  was  used  but  this  time  

the  height  and  taper  of  the  preparation  was  quantified.  

 

Figure  11  Three  primary  views  of  the  inlay-­‐shaped  artifact  as  specified  in  the  ISO  standard  (BS  

EN  ISO  12836:2012,  2012)  

 

3.6.1.2 Height  and  Taper  of  Bridge  Specimen  

As  shown  in  Figure  12,  the  angle  α  and  a  height  b  of  the  bridge  shaped  cavity  

was  measured  using  Geomagic®  Qualify.  

Page 49: An in vitro investigation into the accuracy of CAD/CAM digitizing

47    

 

Figure  12  Views  of  the  bridge-­‐shaped  artifact  as  specified  in  the  ISO  standard  (BS  EN  ISO  

12836:2012,  2012)  

Following  scanning  of   the  bridge  preparations,  height  and   taper  was  again  

measured   using   the   workflow   described   for  measuring   the   slip   gauges.   A  

screen  grab  of  this  process  can  be  seen  in  Figure  13  below.  

 

Page 50: An in vitro investigation into the accuracy of CAD/CAM digitizing

48    

 

Figure  13  -­‐  Height  b  and  angle  α  being  measured  for  digital  model  of  bridge  specimen  

In  order  to  quantify  the  accuracy  of  the  digitizing  devices  at  measuring  the  

inlay   specimen   and   the   bridge   specimen   the   repeatability,   reproducibility  

and  trueness  of  measurement  were  calculated  as  described  below.    

3.6.1.3 Repeatability    

Repeatability   is   defined   as   the   closeness   of   the   agreement   between   the  

results  of   successive  measurements  of   the   same  measurement  and  carried  

out   under   the   same   conditions   of   measurement   (Leach,   2014).   This   is   a  

qualitative  concept  that  can,  however,  be  quantified  by  calculating  the  mean  

and  standard  deviation  of  the  measurements,  and  the  standard  deviation  is  

then  expressed  as  the  repeatability  of  measurement.    

For   the   purposes   of   this   investigation   the   repeatability   conditions   were  

thirty   individual   measurements   obtained   with   the   same   measurement  

procedure   in   the   same   laboratory   by   the   same   operator   (T.E.)   using   the  

same   equipment.   For   the   laser   profilometer,   the   measurement   control  

software   was   used   to   carry   out   consecutive   measurements,   without  

Page 51: An in vitro investigation into the accuracy of CAD/CAM digitizing

49    

removing  the  specimens  from  the  digitizing  device.  For  the  3Shape®  extra-­‐

oral   scanner   the   measurements   were   carried   out   within   a   short   space   of  

time,   again  without   removing   the   specimen   from   the   device.   For   the   3M™  

True   Definition   intra-­‐oral   scanner   the   measurements   were   carried   out  

within  a  short  space  of  time  with  the  specimen  held  in  the  same  position  on  

the  bench  top.  

3.6.1.4 Reproducibility  

Reproducibility   is   defined   as   the   closeness   of   the   agreement   between   the  

results   of   measurements   of   the   same   measure   and   carried   out   under  

changed   conditions  of  measurement.  The  changed  conditions  may   include  

the   observer,   measuring   instrument,   location   and   time.   (Leach,   2014).   As  

with   repeatability,   this   is   a   qualitative   concept   that   can   be   quantified   by  

calculating   the  mean  and  standard  deviation  of   the  measurements  and   the  

standard  deviation  is  then  expressed  as  the  reproducibility  of  measurement.    

For  the  purposes  of  this  investigation  the  reproducibility  conditions  were  30  

measurements  obtained  with  the  same  measurement  procedure  in  the  same  

laboratory  by  the  same  operator  (T.E.)  using  the  same  equipment  as  above,  

however   this   for   the   reproducibility   measurement   the   specimens   were  

removed   from   the  measurement   device   and   repositioned   in   the   digitizing  

device  before  each  new  measurement.    

3.6.1.5 Trueness  

Trueness   is   defined   as   the   closeness   of   agreement   between   the   mean  

obtained   from   repeated  measurements   and  a   true  value  or   a   conventional  

true   value   (Leach,   2014).   As   with   the   concepts   above,   trueness   is   a  

Page 52: An in vitro investigation into the accuracy of CAD/CAM digitizing

50    

qualitative   concept   however   it   can   be   expressed   by   calculating   the  

systematic  error  of  a  measurement  system.    

For  the  purposes  of  this  investigation,  trueness  was  measured  by  calculating  

the  difference  between  the  mean  of  the  thirty  repeatability  measurements  of  

the   test   digitizing   devices   and   the   mean   of   the   thirty   repeatability  

measurements  of  the  reference  device.  

3.7 ‘Proof  of  Concept’  of  Using  the  3Shape®  Extra-­‐oral  Scanner  and  the  

3M™  True  Definition   Intra-­‐oral   Scanner   to   Follow   a   Profilometric  

Tooth  Wear  Measurement  Technique    

As   a   ‘proof   of   concept’   investigation,   a   previously   constructed   calibration  

model   used   to   simulate   tooth   wear   was   scanned   using   the   test   digitizing  

devices  to  create  data  sets  of  the  calibration  models.  Rodriguez  et  al.  (2012)  

fabricated  the  model  using  type  IV  dental  stone  (MoonstoneTM  Bracon  Ltd.,  

Etchingham,   England),   which   had   involved   the   use   of   superplastically  

formed  Titanium-­‐64  alloy  onlays  of  known  step  height  being  cemented  onto  

the  occlusal/palatal  surfaces  of  the  UR5,  UR6,  UL5  and  the  palatal  surface  of  

UL2.  When   in   place   these   additions   represented   a   ‘before’  wear   state   and  

another   MoonstoneTM   cast   of   the   original   model   without   these   additions  

represented  an  ‘after’  wear  state.    

For  the  scanning  with  the  two  digitizing  sytems  being  tested  the  casts  were  

again   coated   with   Telescan   carbon   powder   (DFS®   Diamon   GbmH   –  

Riedenburg,  Germany)  and  scanned  as  described  above.  The  description  of  

the  measurement  of  the  occlusal/palatal  onlays  for  profilometric  tooth  wear  

Page 53: An in vitro investigation into the accuracy of CAD/CAM digitizing

51    

measurement  technique  (step  height)  has  been  fully  described  by  Rodriguez  

et   al.   however   Figure   14   and   Figure   15   below   illustrate   how   this  

measurement  was  carried  out.  

 

Figure  14  -­‐  Geomagic®  software  being  used  to  create  a  2-­‐D  outline  of  the  calibration  model  and  

its  onlays  

 

Figure  15  -­‐  Geomagic®  software  being  used  to  analyze  calibration  model  volume  

In  order  to  quantify  the  accuracy  of  the  3Shape®  extra-­‐oral  scanner  and  the  

3M™  True  Definition  intra  oral  scanner  for  measurement  of  this  calibration  

Page 54: An in vitro investigation into the accuracy of CAD/CAM digitizing

52    

model  the  step  height  data  from  the  onlays,  as  quantified  by  the  test  devices,  

was   compared   to   the   true  known  step  height  of   the  onlay  as  described  by  

Rodriguez  et  al.  (2012).    

Page 55: An in vitro investigation into the accuracy of CAD/CAM digitizing

53    

3.8 Statistics  

Data  were  exported  to  an  Excel  spread  sheet  (Microsoft®  Office  Excel®  2010,  

Microsoft®  Corporation,  USA)  and  statistical  analyses  were  performed  using  

GraphPad   Prism   statistical   software   (GraphPad   Prism   version   6.00   for  

Windows,  GraphPad  Software,  La  Jolla  California  USA,  www.graphpad.com)  

using   the   GraphPad   Statistics   Guide   to   guide   the   choice   of   analysis  

(GraphPad  Software,  2013).    

For   the  comparison  of   the  resolution  of  measurement  a  Bland-­‐Altman  plot  

was   used   to   graphing   the   comparison   of   measurement   techniques   by  

calculating   the   difference   between   the   test   measurements   and   the   true  

values.  These  differences  were  then  described  descriptively.  

In  order   to  compare  the  reproducibility  and  repeatability  of   the  3M™  True  

Definition   intra-­‐oral  scanner  and   the  3  Shape®  extra-­‐oral  scanner  with   the  

reference   profilometer,   the   F-­‐test   of   the   equality   of   two   variances   was  

carried  out  using  Excel.  Firstly,   the  normality  of   the  data  were   tested  with  

the   D'Agostino   &   Pearson   omnibus   normality   test   using   GraphPad   Prism  

statistical   software.   The   data   were   found   to   be   normally   distributed   and  

were   therefore   expressed   as  means   and   standard   deviations.   Secondly,   in  

order  to  compare  the  variance  data  of  the  test  scanners  in  comparison  to  the  

reference  scanner,  the  F-­‐test  of  the  equality  of  two  variances  was  carried  out  

with  p<0.05  used  to  infer  any  statistically  significant  differences.  

Finally,  unpaired  t-­‐tests  were  carried  out  in  order  to  compare  the  trueness  

(height   b   and   angle   α)   of   the   3M™   True   Definition   intra-­‐oral   scanner   in  

comparison  to   the  3  Shape®  extra-­‐oral  scanner.  Statistical  significance  was  

Page 56: An in vitro investigation into the accuracy of CAD/CAM digitizing

54    

determined   with   p<0.05   and   height   and   angle   measurement   data   were  

analysed   individually,  without  assuming  a  consistent  standard  deviation  of  

the  sampled  population  data.  

As   ‘proof   of   concept’   of  wear  measurement  was   carried  out   using   a   single  

measurement  no  statistical   comparisons  were  required   for   this  part  of   the  

study.  .    

Page 57: An in vitro investigation into the accuracy of CAD/CAM digitizing

55    

Results  

4.1 Measurement  of  Engineering  Slip  Gauges  

Table  1  Mean  (SD)  results  of  the  measurement  of  slip  gauges  using  the  test  and  reference  

measurement  devices  (n=30)  

Thickness  of  slip  gauge  (µm)  

50   100   150   200   300   400   500   600   700   800  

Mean  (SD)  Reference  

Profilometer  measurement  

(n=30)  

NM   198  (34)  

172  (21)  

252  (14)  

346  (18)  

420  (5)  

513(4)  

608  (4)  

712  (5)  

818  (6)  

Mean  (SD)  3Shape®  extra-­‐oral  scanner  measurement  

(n=30)  

NM   NM   NM  660  (110)  

433  (120)  

445  (13)  

596  (26)  

635  (18)  

701  (13)  

815  (14)  

Mean  (SD)  3M™  True  Definition  

intra-­‐oral  scanner  

measurement  (n=30)  

NM   NM   NM  440  (142)  

460  (127)  

491  (43)  

540  (12)  

660  (19)  

734  (21)  

822  (16)  

 

Table  1  Mean  (SD)  results  of  the  measurement  of  slip  gauges  using  the  test  

and   reference  measurement  devices   (n=30).     It  was   clear   that  none  of   the  

measurement   devices   were   accurate   when   measuring   the   width   of   a   slip  

gauges   below   400   µm.   The   reference   profilometer  was   able   to   obtain   the  

most  readings  and  would  appear  to  be  the  most  accurate  as  well.  The  3M™  

True  Definition  intra-­‐oral  scanner  was  not  able  to  obtain  any  readings  below  

200   µm   and   was   not   particularly   accurate   below   500   µm.   The   3Shape®  

extra-­‐oral  scanner  was  not  able   to  provide  quantifiable  readings   for  either  

50  or  100  µm,  while  the  scan  data  it  provided  below  400  µm  was  not  close  to  

the  ‘true’  value.  

Page 58: An in vitro investigation into the accuracy of CAD/CAM digitizing

56    

 

Figure  16  Bland-­‐Altman  plot  of  the  agreement  between  the  three  measurement  systems  at  the  

measurement  of  engineering  slip  gauges  of  increasing  dimensions  

The   Bland-­‐Altman   plot   shows   that   there   is   a   clear   trend   for   a   smaller  

difference   between   the   methods   as   the   dimensions   of   the   slip   gauges  

increases.  As  the  thickness  of  the  slip  gauges  increased  the  agreement  also  

increased,   such   that   at   the   largest   thickness   the   limits   of   agreement  were  

very   narrow   and   from   700   µm   onwards   the  measurement  methods   were  

essentially   equivalent.   However,   the   limits   of   agreement   were   seen   to   be  

very  wide   at   the   lower   limits   of   the   dimensions   of   the   slip   gauges  with   a  

threshold  level  seemingly  around  a  dimension  of  400  µm.    

0 2 0 0 4 0 0 6 0 0 8 0 0 1 0 0 00

2 0 0

4 0 0

6 0 0

8 0 0

1 0 0 0

T ru e d im e n s io n o f s lip g u a g e (mm )

Measu

red  valu

e  (mm

)

R e fe re n ce  p ro filom e te r

3 S h a p e®  e x tra -­‐o ra l  s c a n n e r

T ru e  D e fin it io n ™  in tr a -­‐o ra l  s c a n n e r

Page 59: An in vitro investigation into the accuracy of CAD/CAM digitizing

57    

4.2 Repeatability,   Reproducibility   and   Trueness   of   Measurement   of  

the  Inlay  Shaped  Specimen  and  the  Bridge  Shaped  Specimen  

4.2.1 Repeatability  Measurement  Results    

4.2.1.1 Inlay  Specimen  (Height  and  Taper)  

Table  2  Repeatability  results  of  the  measurements  of  the  height  b  (mm)  and  angle  α  (°)  of  the  

inlay  specimen  

 

Repeatability  of  reference  data  (Reference  

profilometer)  

Repeatability  of  test  data  (3Shape®  extra-­‐

oral  scanner)  

Repeatability  of  test  data  (3M™  True  

Definition  intra-­‐oral  scanner)  

Mean  (SD)  Height  b  (mm)    

1.767  (0.009)   1.753  (0.019)*   1.775  (0.014)*  

Mean  (SD)  Angle  α  (°)  

17.484  (0.199)   17.822  (0.288)*   18.572  (0.294)*  

*  Indicates  statistically  significant  differences  between  the  SD  of  the  test  digitizing  devices  vs.  the  reference  device  (p  <0.05)  

Table  2  Repeatability  results  of  the  measurements  of  the  height  b  (mm)  and  

angle  α  (°)  of  the  inlay  specimen  (n=30).  As  specified  in  the  ISO  standard  (BS  

EN  ISO  12836:2012,  2012)  the  repeatability  of  measurement  is  expressed  as  

the  standard  deviation  therefore  the  repeatability  of  height  b  measurement  

of   the  reference  profilometer  was  9  µm,  of   the  3Shape®  extra-­‐oral  scanner  

was  19  µm  and  of   the  3M™  True  Definition   intra-­‐oral   scanner  was  14  µm.  

Repeatability   of   the   devices   with   regards   to   angle   α   measurement   was  

0.199°,  0.288°  and  0.294°  respectively.  

Statistical  testing  using  the  F-­‐test  showed  the  variance  of  the  test  data  was  

statistically   significantly   increased   in   comparison   to   the   reference   data  

(P<0.05),  which  therefore  suggested  that  the  extra-­‐  and  intra-­‐oral  scanners  

showed  poorer  repeatability  at  measuring  the  height  and  angle  of  the  inlay  

specimen  than  the  reference  profilometer.    

Page 60: An in vitro investigation into the accuracy of CAD/CAM digitizing

58    

4.2.1.2 Bridge  Specimen  (Height  and  Taper)  

Table  3  Mean  (SD)  Repeatability  results  of  the  measurements  of  the  height  b  (mm)  and  angle  α  

(°)  of  the  bridge  specimen  

 

Repeatability  of  reference  data  (Reference  

profilometer)  

Repeatability  of  test  data  (3Shape®  extra-­‐

oral  scanner)  

Repeatability  of  test  data  (3M™  True  Definition  intra-­‐

oral  scanner)  

Mean  (SD)  Height  b  (mm)    

2.659  (0.005)   2.599  (0.012)  *   2.754  (0.027)  *  

Mean  (SD)  Angle  α  (°)  

9.292  (0.170)   9.980  (0.413)  *   10.192  (0.305)  *  

*  Indicates  statistically  significant  differences  between  the  SD  of  the  test  digitizing  devices  vs.  the  reference  device  (p  <0.05)  

Table  3  Mean  (SD)  Repeatability  results  of  the  measurements  of  the  height  b  

(mm)  and  angle  α  (°)  of  the  bridge  specimen  (n=30).  Again,  the  repeatability  

of  measurement  is  expressed  as  the  standard  deviation  so  the  repeatability  

of   height   b  of   the   reference  profilometer  was  5  μm,   of   the  3Shape®  extra-­‐

oral  scanner  was  12  μm  and  of  the  3M™  True  Definition  intra-­‐oral  scanner  

was   27   μm.   Repeatability   of   the   devices   with   regards   to   angle   α  

measurement  was  0.170°,  0.413°  and  0.305°  respectively.  

Statistical  testing  using  the  F-­‐test  showed  the  variance  of  the  test  data  was  

statistically   significantly   increased   in   comparison   to   the   reference   data  

(P<0.05),  which  therefore  suggested  that  the  extra-­‐  and  intra-­‐oral  scanners  

showed  poorer  repeatability  at  measuring  the  height  and  angle  of  the  bridge  

specimen  than  the  reference  profilometer.    

Page 61: An in vitro investigation into the accuracy of CAD/CAM digitizing

59    

4.2.2 Reproducibility  of  the  Measurement    

4.2.2.1 Inlay  Specimen  (Height  and  Taper)  

Table  4  Mean  (SD)  Reproducibility  results  of  the  measurements  of  the  height  b  (mm)  and  angle  

α  (°)  of  the  inlay  specimen  

 

Reproducibility  of  reference  data  (Reference  

profilometer)  

Reproducibility  of  test  data  (3Shape®  extra-­‐

oral  scanner)  

Reproducibility  of  test  data  (3M™  True  Definition  intra-­‐

oral  scanner)  

Mean  (SD)  Height  b  (mm)    

1.755  (0.027)   1.757  (0.047)  *   1.780  (0.061)  *  

Mean  (SD)  Angle  α  (°)   17.441  (0.409)   17.947  (0.654)  *   18.006  (0.580)  *  

*  Indicates  statistically  significant  differences  between  the  SD  of  the  test  digitizing  devices  vs.  the  reference  device  (p  <0.05)  

Table   4   Mean   (SD)   Reproducibility   results   of   the   measurements   of   the  

height  b  (mm)  and  angle  α  (°)  of  the  inlay  specimen.  The  reproducibility  of  

the   reference   profilometer   with   regards   to   mean   height   b   of   was   27   μm,  

whereas  the  3Shape®  extra-­‐oral  scanner  and  the  3M™  True  Definition  intra-­‐

oral   scanner   had   a   reproducibility   of   47   μm   and   61   μm   respectively.   The  

reproducibility   of   angle   α   measurement   for   the   reference   scanner   was  

0.409°,  for  the  3Shape®  extra-­‐oral  scanner  was  0.654°  and  for  the  3M™  True  

Definition  intra-­‐oral  scanner  was  0.580°.  

Statistical  testing  using  the  F-­‐test  showed  the  variance  of  the  test  data  was  

statistically   significantly   increased   in   comparison   to   the   reference   data  

(P<0.05),  which  therefore  suggested  that  the  extra-­‐  and  intra-­‐oral  scanners  

showed   poorer   reproducibility   at   measuring   the   height   and   angle   of   the  

inlay  specimen  than  the  reference  profilometer.    

Page 62: An in vitro investigation into the accuracy of CAD/CAM digitizing

60    

4.2.2.2 Bridge  Specimen  (Height  and  Taper)  

Table  5  Mean  (SD)  Reproducibility  results  of  the  measurements  of  the  height  b  (mm)  and  angle  

α  (°)  of  the  bridge  specimen  

 

Reproducibility  of  reference  data  (Reference  

profilometer)  

Reproducibility  of  test  data  (3Shape®  extra-­‐

oral  scanner)  

Reproducibility  of  test  data  (3M™  True  Definition  intra-­‐

oral  scanner)  

Mean  (SD)  Height  b  (mm)    

2.669  (0.021)   2.657  (0.020)  *   2.768  (0.051)  *  

Mean  (SD)  Angle  α  (°)  

9.271  (0.311)   9.426  (0.582)  *   9.909  (0.403)  *  

*  indicate  statistically  significant  differences  between  the  SD  of  the  test  digitizing  devices  vs.  the  reference  device  (p  <0.05)  

Table   5   Mean   (SD)   Reproducibility   results   of   the   measurements   of   the  

height   b   (mm)   and   angle   α   (°)   of   the   bridge   specimen.   When   looking   at  

reproducibility  of  height  b  measurement  of   the   reference  profilometer   the  

value   was   21   μm,   which   was   similar   to   the   3Shape®   extra-­‐oral   scanner’s  

reproducibility   of   20   μm   but   statistically   different.   The   reproducibility   of  

height   b   measurement   the   3M™   True   Definition   intra-­‐oral   scanner   was  

recorded  as  51  μm.  The  reference  profilometer  was  again  most  reproducible  

for  measurement   of   the   angle   α   with   a   standard   deviation   of   0.311°.   The  

3Shape®   extra-­‐oral   scanner   and   3M™   True   Definition   intra-­‐oral   scanner  

showed  a  reproducibility  of  0.582°  and  0.403°  respectively.  

Statistical  testing  using  the  F-­‐test  showed  the  variance  of  the  test  data  was  

statistically   significantly   increased   in   comparison   to   the   reference   data  

(P<0.05),  which  therefore  suggested  that  the  extra-­‐  and  intra-­‐oral  scanners  

showed   poorer   reproducibility   at   measuring   the   height   and   angle   of   the  

bridge  specimen  than  the  reference  profilometer..    

Page 63: An in vitro investigation into the accuracy of CAD/CAM digitizing

61    

4.2.3 Trueness  of  the  Measurement    

Table  6  Mean  (SD)  trueness  of  the  test  data  from  the  measurement  of  the  height  b  (µm)  of  the  

inlay  specimen  and  the  bridge  specimen  using  the  3Shape®  extra-­‐oral  scanner  and  the  3M™  

True  Definition  intra-­‐oral  scanner  

 Trueness  of  test  data  from  the  3Shape®  extra-­‐oral  scanner    

Trueness  of  test  data  from  the  3M™  True  Definition  intra-­‐oral  scanner  in  comparison  to  the  

reference  profilometer  Height  b  (µm)  of  the  

inlay  specimen  -­‐14.2   7.6  

Height  b  (µm)  of  the  bridge  specimen  

-­‐68.8   94.8  

Mean  (SD)    -­‐41.5    (38.61)    ns  

51.2    (61.66)    ns  

 

Table  7  Mean  (SD)  trueness  of  the  test  data  from  the  measurement  of  the  angle  α  (°)  of  the  

inlay  specimen  and  the  bridge  specimen  using  the  3Shape®  extra-­‐oral  scanner  and  the  3M™  

True  Definition  intra-­‐oral  scanner  in  comparison  to  the  reference  profilometer  

  Trueness  of  test  data  (3Shape®  extra-­‐oral  scanner)  

Trueness  of  test  data  (3M™  True  Definition  intra-­‐oral  scanner)  

Angle  α  (°)  inlay  specimen   0.338   1.087  

Angle  α  (°)  bridge  specimen   0.688   0.900  

Mean  (SD)  0.51    (0.25)    ns  

0.99    (0.13)    ns  

 

Table  6  Mean   (SD)   trueness  of   the   test  data   from   the  measurement  of   the  

height   b   (µm)   of   the   inlay   specimen   and   the   bridge   specimen   using   the  

3Shape®  extra-­‐oral  scanner  and  the  3M™  True  Definition  intra-­‐oral  scanner  

Although  these  results  suggest  that  there  was  a  trend  for  the  3Shape®  extra-­‐

oral  scanner  to  underestimate  the  true  value  of  the  height  and  angle  of  the  

specimens   in   comparison   to   the   3M™   True   Definition   intra-­‐oral   scanner,  

which   showed   a   trend   of   overestimating   these   dimensions,   this   was   not  

Page 64: An in vitro investigation into the accuracy of CAD/CAM digitizing

62    

statistically   significant.  Further   to   this,   these   results  also  demonstrate   that  

the   variability   of   the   trueness   of   the   angle   test   data  was   less   for   the   3M™  

True   Definition   intra-­‐oral   scanner   than   the   3Shape®   extra-­‐oral   scanner,  

which  may   suggest   that   the   angular   tolerance   of   the   3M™  True  Definition  

intra-­‐oral  scanner  may  be  more  consistent.    

Page 65: An in vitro investigation into the accuracy of CAD/CAM digitizing

63    

4.3 ‘Proof  of  Concept’  of  Using  the  3Shape®  Extra-­‐oral  Scanner  and  the  

3M™   True  Definition   Intra-­‐oral   Scanner   to   Follow   a   Profilometric  

Tooth  Wear  Measurement  Technique    

 

Figure  17  –  Volume  of  ‘tooth  tissue  loss’  demonstrated  using  Geomagic®  best  fit  analysis  

As  shown  in  Figure  17  above,  a  heat  map  of  ‘tooth  tissue  loss’  was  generated  

using   Geomagic®   Qualify.   The   blue   areas   designate   where   there   has   been  

loss   height   when   comparing   the  model   without   onlays   to   the  model   with  

onlays.  The  darker  the  blue,  the  more  ‘tissue  loss’  has  occurred.  

Table  8  Results  of  the  volumetric  measurement  of  the  calibration  model  

  Reference  profilometer  

3Shape®  extra-­‐oral  scanner  

True  Definition™  intra-­‐oral  scanner  

  Volume  data  (mm3)  

Volume  data  (mm3)  

Volume  error  (mm3)  

Volume  data  (mm3)  

Volume  error  (mm3)  

UR5     8.6   8.5   -­‐0.1   8.7   0.1  

UR6     18   18.1   0.1   18.2   0.2  

UL2     12.7   12.3   -­‐0.4   12.9   0.2  

UL5     11   10.5   -­‐0.5   10.9   0.3  Mean  (SD)  volume  error  (mm3)   -­‐0.2  (0.3)     0.2  (0.1)    

 

Page 66: An in vitro investigation into the accuracy of CAD/CAM digitizing

64    

Table   8   Results   of   the   volumetric   measurement   of   the   calibration   model.  

Both  the  3Shape®  extra-­‐oral  scanner  and  the  3M™  True  Definition  intra-­‐oral  

scanner  show  a  mean  volume  variation  from  the  reference  volume  of  ±  0.2  

mm3.  However  the  3Shape®  extra-­‐oral  scanner  tended  to  underestimate  the  

volume,   while   the   3M™   True   Definition   intraoral   scanner   tended   to  

overestimate  the  volume.    

Page 67: An in vitro investigation into the accuracy of CAD/CAM digitizing

65    

Discussion  

As   can   be   seen   in   the   Bland   and   Altman   plot   (Figure   16   above),   all   three  

devices   show   good   correlation   in   their   ability   to   measure   slip   gauge  

thicknesses  of  400  μm  and  above.  The  Bland  and  Altman  plot  was  chosen  to  

display  the  data  because  it  is  the  most  appropriate  statistical  analysis  to  use  

when   needing   to   compare   two   or  more  measurements   that   each   contains  

some   errors   in   their  measurement.   Essentially   the   Bland   and   Altman   plot  

allows   these   measurements   to   be   evaluated   for   their   ‘agreement’.   The  

reference   scanner’s   sensor   has   a   30   μm   diameter   and   can   capture   data  

points   every   50   μm.   This   explains  why   it  was   unable   to   provide   readable  

data  when  measuring   the  50  μm  slip  gauge.  Accuracy  of  both   the  3Shape®  

extra-­‐oral  scanner  and  3M™  True  Definition  intra-­‐oral  scanner  in  measuring  

the  width  of   the  slip  gauges  below  400  µm  was  poor   in  comparison   to   the  

reference  scanner.  It  can  be  deduced  from  this  that  the  optical  resolution  of  

the  test  scanners  is  inferior  to  that  of  the  reference  scanner.    

As   has   previously   been   described,   optical   resolution   of   the   objective   is   a  

limiting   factor   for   non-­‐contacting   profilometers.   This   describes   the  

minimum  distance  between  two  lateral  features  that  can  be  measured,  and  

in  the  case  of  this  study  the  lateral  features  were  the  right  angle  edges  of  the  

engineering   slip   gauges.   The   intra-­‐   and   extra-­‐oral   dental   scanners   are  

designed  to  measure  tooth-­‐shaped  freeform  objects  which  may  have  limited  

the   capabilities   of   the   stitching   software   to   allow   images   to   be   reliably  

reconstructed  when  the  thickness  of   the  gauges  was  below  400  µm.  Below  

this  400  µm   threshold  value,   the   two   right   angles  presented   to   the   sensor  

Page 68: An in vitro investigation into the accuracy of CAD/CAM digitizing

66    

were   too   closely   approximated   and   so   the   profilometers   had   difficulty   in  

recording  the  width.  This  was  a  situation  that  may  not  be  clinically  relevant  

and  the  test  was  perhaps  more  applicable   for   investigating  the  capabilities  

of  the  hardware,  rather  than  the  hardware  and  software  combined.  Another  

limitation   of   this   study   that   could   have   affected   the  measurements   is   that  

accurate  scanning  depends  on  an  uninterrupted  line  of  sight  between  laser,  

surface  and  detector.  Therefore,  if  a  step  is  to  be  measured,  as  is  essentially  

the  case  when  measuring  a  slip  gauge   from  side  on,   the  sensor  must  be   in  

the  correct  orientation  so  that  the  laser  spot  is  not  hidden  by  the  edge  of  the  

step  (Zeng  et  al,  1997).  

Repeatability  and  reproducibility   for  both   the  extra-­‐oral  and   the   intra-­‐oral  

scanners   was   found   to   be   poorer   than   that   of   the   reference   profilometer  

when  measuring  either   the  height  or   the  angle  of  both   the   inlay   specimen  

and   the   bridge   specimen.   Therefore   the   reference   scanner   had   the   best  

repeatability  and  reproducibility  with  a  standard  deviation  of  5  –  9  μm.  The  

3Shape®  extra-­‐oral  scanner  had  the  next  best  repeatability  with  a  range  of  

12  –  19  μm  and  the  3M™  True  Definition  intra-­‐oral  scanner  had  a  range  of  

14  –  27  μm.  All  of  these  were  similar  to  the  coefficient  of  repeatability  range  

of   7–16   μm   found   by   Persson   et   al   (2008)   and   the   poorer   precision   of  

measurement  in  comparison  to  the  reference  profilometer  was  perhaps  not  

unexpected   considering   that   the   reference   profilometer   was   a   high  

specification  measurement  device.    

Previous  studies  have  looked  at  implant  angulation  but  as  far  as  the  authors  

are  aware  there  are  no  other  studies  that  have  investigated  the  repeatability  

Page 69: An in vitro investigation into the accuracy of CAD/CAM digitizing

67    

or  reproducibility  of  convergence  angle  α.  Despite   the  poorer  repeatability  

and   reproducibility,   the   results   suggest   that   the   precision   of   angular  

measurement  is  reasonably  good  for  all  three  scanners.  The  ability  of  intra-­‐

oral   scanners   to   accurately   measure   angulation   will   of   course   be   of  

particular   benefit   in   implant   dentistry,   where   the   clinician   is   often   faced  

with   large   span   frameworks   that   through   errors   do   not   fit   passively.   If  

implant   fixtures   at   varying   angles   can   be   measured   with   precision   and  

accuracy   then   costly   and   time-­‐consuming   chair   side   adjustment   may   be  

avoided.  

Without   a   metrological   standard,   the   reference   profilometer   scans   of   the  

inlay  and  bridge  specimens  served  as  the  reference  model.  The  accuracy  of  

this  scanner  in  measuring  length  was  shown  to  be  1.3  μm  by  Rodriguez  et  al.  

(2012)  and  it  was  felt  that  this  was  an  acceptable  standard  against  which  to  

compare   the   two   scanners  being   investigated.   It   should  be  noted  however  

that  BS  EN   ISO  12836   required   inlay   and  bridge   specimens   to   be   of   exact  

dimensions  and  made  from  a  dimensionally  stable  material  but  this  was  not  

possible  for  this  study  for  reasons  already  stated.    

With  regard  to  the  accuracy  results  of  the  two  test  scanners,  as  shown  in  the  

trend  towards  a  overestimation  of  the  trueness  for  the  3M™  True  Definition  

intra-­‐oral   scanner   in   comparison   to   the   3Shape®   extra-­‐oral   scanner   may  

have   been   as   a   result   of   the   need   for   a   light   dusting   of   titanium   dioxide  

powder,   which   is   of   unknown   thickness.   Titanium   dioxide   application   is  

described  by  various  authors  (Syrek  et  al.  2010,  Van  der  Meer  et  al.  2012)  

but  at  present  there  is  no  report  into  how  this  might  scanning  accuracy.  

Page 70: An in vitro investigation into the accuracy of CAD/CAM digitizing

68    

As   considered   by   Bartlett   (2010),   understanding   of   the   aetiology   and  

pathogenesis   of   tooth   wear   is   still   lacking.   Techniques   to   quantitatively  

measure  tooth  wear  have  proved  to  be  time  consuming  and  costly,  usually  

requiring  specialised  hardware  and  software  (Lee  et  al.  2012).  Traditionally  

in  vivo  quantitative  tooth  wear  studies  have  measured  tooth  surface  loss  in  

terms   of   height   lost,   which   can   be   subjective   to   measure   and   offers   little  

information  about  the  wear  lesion.  Conversely,  in  vitro  studies  offer  little  to  

no  information  with  regard  to  aetiology  and  no  matter  how  well  designed  do  

not   precisely   replicate   the   conditions   found   in   an   intra-­‐oral   environment,  

although  Sagakuchi  et  al.  (1986)  have  previously  reported  good  correlation.  

Both  the  3Shape®  extra-­‐oral  scanner  and  the  3M™  True  Definition  intra-­‐oral  

scanner  show  a  mean  volume  variation  from  the  reference  volume  of  ±  0.2  

mm3.    

There  are  few  studies  that  look  at  volumetric  tooth  wear  but  the  volumetric  

error  measurement  found  in  this  ‘proof  of  concept’  investigation  carried  out  

in   this   present   study   would   potentially   render   the   intra-­‐oral   scanner  

unsuitable   for  use   in  a  clinical   study   investigating  short   term  rate  of   tooth  

wear   given   that   Tantbirojn   et   al.   (2012)   described   mean   (standard  

deviation)  volume  loss  per  tooth  of  0.18  (0.12)  mm3  over  a  6  month  period  

(in  patients  with  gastro-­‐esophageal   reflux  disease).  However  Tantbirojn  et  

al.   do   note   that   they   scanned   dental   stone   models   cast   from  

polyvinylsiloxane  and  they  go  on  to  state,  “the  accuracy  [of  measurements]  

can   be   improved   further   with   advances   in   intra-­‐oral   scanning   techniques  

that   will   eliminate   potential   distortions   and   dimensional   changes   in   the  

impression  material  and  dental  stones.”    

Page 71: An in vitro investigation into the accuracy of CAD/CAM digitizing

69    

As  CAD/CAM  technology  continues  to  develop  and  become  more  prevalent  

in   dental   surgeries   it   will   not   be   surprising   to   see   its   application   to   the  

quantitative   measurement   of   tooth   wear.   However,   further   research   is  

certainly   required   in   this   area   in   order   to   determine   more   clearly   the  

effectiveness   of   intra-­‐oral   scanning   in   the   quantitative   measurement   of  

wear,   in   comparison   to  more   established   profilometric   techniques.   In   vivo  

studies  to  investigate  volume  of  tooth  substance  loss  over  time  would  be  of  

particular   benefit,   especially   in   ‘at   risk’   individuals,   however   the   time   and  

cost  implications  of  the  large  numbers  of  participants  who  would  need  to  be  

recruited  for  such  a  study  do  have  to  be  factored  in  to  the  planning  of  future  

clinical  research  in  this  area.    

Page 72: An in vitro investigation into the accuracy of CAD/CAM digitizing

70    

Conclusion  

The  digitizing  devices  under  investigation  were  found  to  be  able  to  measure  

the   engineering   slip   gauges   accurately   to  400  μm  width   and  above.  Below  

this  no   readable  data   could  be   recorded.  When  compared   to   the   reference  

profilometer   this   would   suggest   that   the   optical   resolution   of   the   test  

devices  is  inferior.  

Repeatability  and  reproducibility  for  both  of  the  test  scanners  was  found  to  

be   statistically   significantly   increased   than   that   of   the   reference  

profilometer  when  measuring  either  the  height  or  the  angle  of  both  the  inlay  

specimen   and   the   bridge   specimen   (p<0.05).   Both   the   3Shape®   extra-­‐oral  

scanner  and  the  3M™  True  Definition  intra-­‐oral  scanner  were  able  to  follow  

the   previously   published   profilometric   tooth   wear   measurement   process  

however   they  displayed  a  mean  volume  error  of  ±  0.2  mm3.  This   suggests  

that  as  dental  digitizing  devices  improve  in  their  accuracy  they  will  certainly    

have  future  application  in  the  quantitative  measurement  of  tooth  wear.    

Page 73: An in vitro investigation into the accuracy of CAD/CAM digitizing

71    

References  

2012.BS  EN   ISO  12836:2012.  Dentistry  —  Digitizing  devices   for  CAD/CAM  systems   for   indirect  dental   restorations  —  Test  methods   for  assessing  accuracy.  BSI  Standards  Limited.  

2012.  DIN  EN   ISO  10360   -­‐  VDI/VDE  2617  Part  6.2  Accuracy  of   coordinate  measuring   machines   -­‐   Characteristics   and   testing   of   characteristics   -­‐  Guideline   for   the   application   of   DIN   EN   ISO   10360   to   coordinate  measuring  machines  with  optical  distance  sensors.  

3Shape®   (2014)   Dental   System:   Easy   and   productive   workflows   [online]  Available   at:   http://www.3shapedental.com/restoration/dental-­‐lab/lab-­‐scanners/d700.  

Reuters  (2012)  Global  Dental  CAD/CAM  System  Market  Will  Grow  Strongly  to   $540   Million   by   2016   [online]   Available   at:  http://www.reuters.com/article/2012/07/26/idUS136792+26-­‐Jul-­‐2012+BW20120726.  

Atkinson,   J.T.,   Groves,   D.  &   Lalor,  M.J.,   1982.   The  measurement   of  wear   in  dental   restorations   using   laser   dual-­‐source   contouring.   Wear,   76(1),  pp.91–104.  

Attin,   T.,   Knöfel,   S.,   Buchalla,  W.  &  Tütüncü,   R.,   2001.   In   situ   evaluation   of  different   remineralization   periods   to   decrease   brushing   abrasion   of  demineralised  enamel.  Caries  Research,  35,  pp.216-­‐222.  

Azzopardi,   A.,   Bartlett,   D.W.,Watson   T.   &   Sherriff,   M.,   2001.   The  measurement   and   prevention   of   erosion   and   abrasion.   Journal   of  Dentistry,  29(6),  pp.395–400.    

Bardsley,   P.F.,   2008.   The   evolution   of   tooth   wear   indices.   Clinical   Oral  Investigations,  12(1),  pp.15–19.  

Bartlett   D.W.,   2010.   A   proposed   system   for   screening   tooth   wear.   British  Dental  Journal,  208,  pp.207–209.    

Bartlett,   D.,   Ganss,   C.   &   Lussi,   A.,   2008.   Basic   Erosive   Wear   Examination  (BEWE):  a  new  scoring  system  for  scientific  and  clinical  needs.  Clinical  Oral  Investigations,  12(1),  pp.65–68.    

Bartlett,   D.W.,   1997.   The   causes   of   dental   erosion.   Oral   Diseases,   3(4),  pp.209–211.    

Bartlett,   D.W.,   Blunt,   L.   &   Smith,   B.,   1997.   Measurement   of   tooth   wear   in  patients   with   palatal   erosion.   British   Dental   Journal,   182(5),   pp.179–184.    

Page 74: An in vitro investigation into the accuracy of CAD/CAM digitizing

72    

Bayne,  S.C.,  Felton,  D.A.  &  Kanoy,  B.E.,  1991.  Effect  of   in  vivo  crown  health  margin  discrepancies.  The  Journal  of  Prosthetic  Dentistry,  65(3),  pp.357–364.  

Benington,  P.C.M.,  Khambay,  B.S.  &  Ayoub,  A.F.,  2010.  An  overview  of  three-­‐dimensional  imaging  in  dentistry.  Dental  Update,  37(8),  pp.494–6,  499–500,  503–504.    

Brandestini,  M.,  Mörmann,  W.,  Lutz,  F.  &  Krejci,  I.,  1985.  Computer  machined  ceramic  inlays:  in  vitro  marginal  adaptation.  Journal  of  Dental  Research,  64,  pp.208.  

Bray,   M.,   2004.   Stitching   interferometry:   recent   results   and   absolute  calibration.  Proc.  SPIE,  5252,  pp.305–313.    

Carrotte,   P.,   Winstanley,   R.   &   Green,   J.,   1993.   A   study   of   the   quality   of  impressions   for   anterior   crowns   received   at   a   commercial   laboratory.  British  Dental  Journal,  174,  pp.235–240.    

Chadwick,   R.G.,   Mitchell,   H.L.,   Cameron,   I.,   Hunter,   B.   &   Tulley,   M.,   1997.  Development   of   a   novel   system   for   assessing   tooth   and   restoration  wear.  Journal  of  Dentistry,  25(1),  pp.41–47.    

Dahlmo,  K.I.,  Andersson,  M.,  Gellerstedt,  M.  &  Karlsson,   S.,   2001.  On  a  new  method   to   assess   the   accuracy   of   a   CAD   program.   The   International  Journal  of  Prosthodontics,  14(3),  pp.276–83.    

DeLong,   R.,   Pintado,  M.   &  Douglas,  W.H.,   1985.  Measurement   of   change   in  surface  contour  by  computer  graphics.  Dental  Materials,  1(1),  pp.27–30.    

DeLong,  R.,  Heizen,  M.,  Hodges  M.J.,  Ko,  C.-­‐C.  &  Douglas,  W.H.,  2003.  Accuracy  of  a  system  for  creating  3D  computer  models  of  dental  arches.  Journal  of  Dental  Research,  82(6),  pp.438–442.  

Denissen,  H.W.,  van  der  Zel,  J.M.,  van  Waas,  M.A.J.,  1999.  Measurement  of  the  margins   of   partial-­‐coverage   tooth   preparations   for   CAD/CAM.  International  Journal  of  Prosthodontics  12,  pp.395–400.    

Denissen,  H.W.,  Dozic,  A.,  van  der  Zel,  J.M.  &  van  Waas,  M.A.J.,  2000.  Marginal  fit   and   short-­‐term   clinical   performance   of   porcelain   veneered   Procera  onlays.  Journal  of  Prosthetic  Dentistry  85,  pp.506–513.  

Eames,  W.B.,  Wallace,  S.W.,  Suway,  N.B.  &  Rogers,  L.B.,  1979.  Accuracy  and  dimensional  stability  of  elastomeric  impression  materials.  The  Journal  of  Prosthetic  Dentistry,  42(2),  pp.159–162.    

Ender,  A.  &  Mehl,  A.,  2013.  Accuracy  of  complete-­‐arch  dental  impressions:  a  new   method   of   measuring   trueness   and   precision.   The   Journal   of  Prosthetic  Dentistry,  109(2),  pp.121–128.    

Page 75: An in vitro investigation into the accuracy of CAD/CAM digitizing

73    

Flügge,   T.V.,   Schlager,   S.,   Nelson,   K.,   Nahles,   S.   &   Metzger,   M.C.,   2013.  Precision   of   intraoral   digital   dental   impressions   with   iTero   and  extraoral   digitization   with   the   iTero   and   a   model   scanner.   American  Journal   of   Orthodontics   and   Dentofacial   Orthopedics,   144(3),   pp.471–478.  

Ganss,  C.  &  Lussi,  A.,  2005.  Diagnosis  of  erosive  tooth  wear.  Dental  Erosion:  From  Diagnosis  to  Therapy,  20,  pp.32-­‐43.    

Glentworth,  P.,  Harrison,  A.  &  Moores,  G.,  1984.  Measurement  of  changes  in  surface   profile   due   to   wear   using   a   147   Pm   β   particle   backscatter  technique   II:   Application   to   the   simulated  wear   of   dental.  Wear,   9(1),  pp.53–62.    

Groten,   M.,   Girthofer,   S.   &   Probster,   L.,   1997.   Marginal   fit   consistency   of  copy-­‐milled  all-­‐ceramic  crowns  during  fabrication  by  light  and  scanning  electron  microscopic  analysis  in  vitro.  Journal  of  Oral  Rehabilitation,  24,  pp.871.  

Hewlett,   E.R.,   Orro,   M.E.   &   Clark,   G.T.,   1992.   Accuracy   testing   of   three-­‐dimensional  digitizing  systems.  Dental  Materials,  8(1),  pp.49–53.    

Kaidonis,   J.A.,  Richards,  L.C.,  Townsend,  G.C.  &  Tansley  G.D.,  1998.  Wear  of  Human   Enamel:   A   Quantitative   in   vitro   Assessment.   Journal   of   Dental  Research,  77(12),  pp.1983–1990.  

Karlsson,  S.,  1993.  The  fit  of  Procera  titanium  crowns:  an  in  vitro  and  clinical  study.  Acta  Odontologica  Scandinavica,  53(3),  pp.129-­‐134.  

Kim,   S.   &   Yeo,   I.,   2013.   Accuracy   of   dies   captured   by   an   intraoral   digital  impression  system  using  parallel  confocal  imaging.  International  Journal  of  Prosthodontics,  26(2),  pp.161–163.  

Lambrechts,   P.,   Vanherle,   G.,   Vuylsteke,   M.   &   Davidson,   C.L.,   1984.  Quantitative   evaluation   of   the   wear   resistance   of   posterior   dental  restorations:  a  new  three-­‐dimensional  measuring   technique.   Journal  of  Dentistry,  12(3),  pp.252–267.    

Lambrechts,   P.,   Braem,   M.,   Vuylsteke-­‐Wauters,   M.   &   Vanherle,   G.,   1989.  Quantitative  in  vivo  wear  of  human  enamel.  Journal  of  Dental  Research,  68(12),  pp.1752–1754.    

Leach,  R.,  2014.  Fundamental  principles  of  engineering  nanometrology.  2nd  Edition,  Elsevier.    

Lee,   A.,   He,   L.H.,   Lyons,   K.   &   Swain,   M.V.,   2012.   Tooth   wear   and   wear  investigations   in   dentistry.   Journal   of   Oral   Rehabilitation,   39,   pp.217-­‐225.  

Page 76: An in vitro investigation into the accuracy of CAD/CAM digitizing

74    

Logozzo,   S.,   Zanetti,   E.M.,   Franceschini,   G.   &   Kilpelä,   A.,   2014.   Recent  advances  in  dental  optics  –  Part  I:  3D  intraoral  scanners  for  restorative  dentistry.  Optics  and  Lasers  in  Engineering,  54,  pp.203–221.    

Logozzo,  S.,  Kilpelä,  A.,  Mäkynen,  A.  &  Zanetti,  E.M.,  2014.  Recent  advances  in  dental  optics  –  Part   II:  Experimental   tests   for  a  new   intraoral   scanner.  Optics  and  Lasers  in  Engineering,  54,  pp.187–196.  

Luthardt,   R.G.,   Sandkuhl,   O.,   Herold,   V.   &  Walter,   M.H.,   2001.   Accuracy   of  mechanical   digitizing   with   a   CAD/CAM   system   for   fixed   restorations.  International  Journal  of  Prosthodontics,  14(2),  pp.146–151.  

May,  K.B.,  Russell,  M.M.,  Razzoog,  M.E.  &  Lang,  E.R.,  1998.  Precision   fit:   the  Procera   AllCeram   crown.   The   Journal   of   Prosthetic   Dentistry,   80(4),  pp.394-­‐404.  

Meer,  W.  van  der,  Andriessen,  F.,  Wismeijer,  D.  &  Ren,  Y.,  2012.  Application  of   intra-­‐oral   dental   scanners   in   the   digital   workflow   of   implantology.  PloS  one,  7:e43312.    

Mehl,  A.,   Gloger,  W.,  Kunzelmann,  K.  &  Hickel,   R.,   1997.  A  new  optical   3-­‐D  device   for   the   detection   of   wear.   Journal   of   Dental   Research,   76(11),  pp.1799–1807.    

Miyazaki,  T.,  Hotta,  Y.,  Kunii,  J.,  Kuriyama,  S.  &  Tamaki,  Y.,  2009.  A  review  of  dental  CAD/CAM:  current  status  and  future  perspectives  from  20  years  of  experience.  Dental  Materials,  28(1),  pp.44–56.    

Molnar,  S.,  McKee,  J.K.,  Molnar,  I.M.  &  Przybeck,  T.R.,  1983.  Tooth  wear  rates  among  contemporary  Australian  Aborigines.  Journal  of  Dental  Research,  62(5),  pp.562–565.    

Noordmans,  J.,  Pluim,  L.  &  Hummel,  J.,  1991.  A  new  profilometric  method  for  determination  of  enamel  and  dentinal  abrasion  in  vivo  using  computer  comparisons:  a  pilot  study.  Quintessence  International,  22,  pp.653–657.    

Persson,   A.S.K.,   Andersson,   M.,   Odén,   A.   &   Sandborgh-­‐Englund,   G.,   2008.  Computer   aided   analysis   of   digitized   dental   stone   replicas   by   dental  CAD/CAM  technology.  Dental  Materials,  24(8),  pp.1123–1130.    

Persson,   A.S.K.,   Andersson,  M.,   Odén,   A.   &   Sandborgh-­‐Englund,   G.,   2006.   A  three-­‐dimensional   evaluation   of   a   laser   scanner   and   a   touch-­‐probe  scanner.  The  Journal  of  Prosthetic  Dentistry,  95(3),  pp.194–200.  

Persson,   M.,   Andersson,   M.   &   Bergman,   B.,   1995.   The   accuracy   of   a   high-­‐precision   digitizer   for   CAD/CAM   of   crowns.   The   Journal   of   Prosthetic  Dentistry,  74(3),  pp.223–229.    

Page 77: An in vitro investigation into the accuracy of CAD/CAM digitizing

75    

Pintado,  M.R.,  Andersson,  G.C.,  DeLong,  R.  &  Douglas,  W.H.,  1997.  Variation  in   tooth   wear   in   young   adults   over   a   two-­‐year   period.   The   Journal   of  Prosthetic  Dentistry,  77(3),  pp.313–320.    

Rodriguez,  J.M.,  Austin,  R.S.  &  Bartlett,  D.W.,  2012.  In  vivo  measurements  of  tooth  wear  over  12  months.  Caries  Research,  46(1),  pp.9–15.    

Rodriguez,   J.M.,   Austin,   R.S.   &   Bartlett,   D.W.,   2012.   A   method   to   evaluate  profilometric   tooth   wear   measurements.   Dental   Materials,   28(3),  pp.245–251.    

Rodriguez,  J.,  Curtis,  R.  &  Bartlett,  D.,  2009.  Surface  roughness  of  impression  materials   and   dental   stones   scanned   by   non-­‐contacting   laser  profilometry.  Dental  Materials,  25(4),  pp.500–505.    

Roulet,  J.,  1987.  Degradation  of  dental  polymers.  Karger,  Basel.    

Rudolph,  H.,  Luthardt,  R.G.  &  Walter,  M.H.,  2007.  Computer-­‐aided  analysis  of  the   influence   of   digitizing   and   surfacing   on   the   accuracy   in   dental  CAD/CAM   technology.   Computers   in   Biology   and   Medicine,   37(5),  pp.579–587.  

Sakaguchi,  R.L.,  Douglas,  W.H.,  DeLong,  R.  &  Pintado,  M.R.,  1986  The  wear  of  a  posterior  composite  in  an  artificial  mouth:  a  clinical  correlation.  Dental  Materials,  2,  pp.235–240.  

Smith,  B.G.  &  Knight,  J.K.,  1984.  An  index  for  measuring  the  wear  of  teeth.  British  Dental  Journal,  156(12),  pp.435-­‐438.  

Sousa,  M.V.S.,  Vasconcelos,  E.C.,  Janson,  G.,  Garib,  D.  &  Pinzan,  A.,  2012  Accuracy  and  reproducibility  of  3-­‐dimensional  digital  model  measurements.  American  Journal  of  Orthodontics  and  Dentofacial  Orthopedics,  142(2),  pp.  269-­‐273.  

Sundaram,   G.,   Wilson,   R.,   Watson,   T.F.   &   Bartlett,   D.W.,   2007.   Clinical  measurement  of  palatal  tooth  wear  following  coating  by  a  resin  sealing  system.  Operative  Dentistry,  32(6),  pp.539–543.  

Syrek,   A.,   Reich,   G.,   Ranftl,   D.,   Klein,   C.,   Cerny,   B.   &   Brodesser,   G.,   2010.  Clinical   evaluation   of   all-­‐ceramic   crowns   fabricated   from   intraoral  digital  impressions  based  on  the  principle  of  active  wavefront  sampling.  Journal  of  Dentistry,  38(7),  pp.553–559.    

Tantbirojn,  D.,  Pintado,  M.  &  Versluis,  A.,  2012.  Quantitative  analysis  of  tooth  surface   loss   associated   with   gastroesophageal   reflux   disease:   a  longitudinal   clinical   study.   Journal   of   the   American   Dental   Association,  143(3),  pp.278-­‐285  

Page 78: An in vitro investigation into the accuracy of CAD/CAM digitizing

76    

Tinschert,   J.,  Natt,  G.,  Mautsch,  W.,  Speikermann,  H.  &  Anusavice,  K.J.,  2000.  Marginal   fit   of   alumina   and   zirconia   based   fixed   partial   dentures  produced   by   a   CAD/CAM   system.   Operative   Dentistry,   26(4),   pp.367-­‐364.  

Vlaar,   S.T.   &   van   der   Zel,   J.M.,   2006.   Accuracy   of   dental   digitizers.  International  Dental  Journal,  56(5),  pp.301–309.  

Williams,   P.,   Jackson,   D.   &   Bergman,  W.,   1984.   An   evaluation   of   the   time-­‐dependent   dimensional   stability   of   eleven   elastomeric   impression  materials.  The  Journal  of  Prosthetic  Dentistry,  52,  pp.120–125.  

Winstanley,  R.,  Carrotte,  P.  &   Johnson,  A.,  1997.  The  quality  of   impressions  for   crowns   and   bridges   received   at   commercial   dental   laboratories.  British  Dental  Journal,  183,  pp.209–213.  

Zeng,   L.,   Matsumoto,   H.   &   Kawachi,   K.,   1997.   Two-­‐directional   scanning  method   for   reducing   the   shadow   effects   in   laser   triangulation.  Measurement  Science  and  Technology,  8,  pp.262.  

Zhang,  R.,  2006.  Theoretical  and  experimental  study  on  the  precision  of  the  stitching  System.  Proc.  SPIE,  6150,  p.61502Y.