Вы находитесь на странице: 1из 8

Karin I.

Dahlmo, BS, CDTa


Matts Andersson, LDS, Odont Dr/PhDb
On a New Method to Assess the Martin Gellerstedt, PhDc
Accuracy of a CAD Program Stig Karlsson, LDS, Odont Dr/PhDd

Purpose: This study was initiated with the purpose of developing and evaluating a system
for measuring the magnitude of the variation between a computer-aided design (CAD)
object created on the computer screen and a replicated object produced by computer-
aided manufacturing (CAM). Materials and Methods: Controlled geometric forms, a square
and a cone, were designed in a CAD program, and measuring distances were selected.
The CAD data were transmitted to CAM data, and objects were produced. The CAD/CAM
process was the Procera system, and the holder system permitted the objects to be ground
into cross sections, similar to the visualization in the CAD program. Five objects of each
shape were produced and measured by two operators two times. Three operators
measured one of the objects 30 times. Results: Observed values were greater than the true
value. For all objects, the systematic error was at most 15.5 µm. Interoperator difference
was small. The variation because of measurement error was greater for the square object
compared to the cone. However, the variation because of object was higher for the cone
object than for the square. The total standard deviation was 7.7 µm. Thus, the total random
error caused by object variation and measurement error was in approximately 95% of all
measurements less than 15 µm. Conclusion: There are no differences in the measurement
data derived from this method and actual measurement data from an object created by the
computer-aided dental design program. The method has high validity and reliability, ie,
high accuracy. Int J Prosthodont 2001;14:276–283.

D uring the past decade, computer-aided de-


sign/manufacturing (CAD/CAM), which are well-
known processes in the industrial world, have found
the number of articles reporting on the clinical suc-
cess and the improved characteristics of restorations
fabricated by these CAD/CAM methods.16–26
their way into dentistry. Interest in this designing and CAD/CAM systems have the potential for three
manufacturing process occurred because of its po- sources of variation. The first source is the measure-
tential as an alternative to the lost-wax technique that ment accuracy of the device used in recording the con-
has been used for many decades for fabricating den- tours of the die and the transfer of the ascii data files
tal restorations. A number of reports in the literature to the computer. The second source of variation is the
in recent years detail the CAD/CAM methods that ability of the computer software to modify the data in
have been used in dentistry.1–15 Equally impressive is designing the restoration. The software must be able
to transfer to the manufacturing facility accurate in-
aManager, Procera Technical Support, Nobel Biocare, Göteborg, formation that will result in the production of a restora-
Sweden. tion that is a replica of what was created on the com-
bManager, Procera Research and Development, Nobel Biocare,
puter screen. The third source of variation involves the
Göteborg, Sweden.
c Statistician, University of Trollhättan/Uddevalla, Uddevalla,
manufacturing process itself and the precision of the
Sweden.
machinery used to fabricate the restoration.
d Professor, Department of Prosthetic Dentistry and Dental Understanding and measuring the sources of vari-
Materials Science, Göteborg University, Sweden. ation, and the effect of each on the accuracy and pre-
Reprint requests: Ms Karin Dahlmo, Nobel Biocare, PO Box 5190,
cision of the CAD/CAM system in use, is extremely
S-402 26 Göteborg, Sweden. Fax: + 46 (31) 163152. e-mail: important to the success of any dental restorations
karin.dahlmo@.nobelbiocare.se produced by these systems, and ultimately the
COPYRIGHT © 2001 BY QUINTESSENCE
PUBLISHING CO, INC. PRINTING OF THIS
DOCUMENT IS RESTRICTED TO PERSONAL
USE ONLY. NO PART OF THIS ARTICLE MAY
The International Journal of Prosthodontics 276 Volume 14, Number 3, 2001 BE REPRODUCED OR TRANSMITTED IN ANY
FORM WITHOUT WRITTEN PERMISSION
FROM THE PUBLISHER.
Dahlmo et al New Method to Assess Accuracy of CAD

healthcare of the patient. Major errors resulting from then compare these data to measurements of the ac-
any one of these error sources would eliminate this tual object produced by the CAM process.
technical approach as an option in dental treatment. A far more practical approach to determine if dif-
Variations between the contours of the original ferences exist between the computer screen mea-
die and the visualization of the data on the computer surements and the actual measurement data from
screen, and variations between the die and the man- the produced object is to use computer-generated
ufactured restoration, have been reported.15,20–22,27 forms of specific designs. A controlled geometric
However, none of the systems available have re- form will provide specific positions for measurement
ported on the variations between a design created on data from both the screen and the same locations on
the computer screen and the design produced at the the form produced by the CAM process. For such a
manufacturing facility. measurement method, the test objects created on the
In industrial applications, the CAD/CAM process computer screen must be well-defined geometric
usually involves robotic mechanisms to produce mul- forms.
tiple components of the same size and dimensions In turn, the computer software must be able to
that can easily be inspected for accuracy using a generate a cross section vertically through the long
quality-control measurement system. In dentistry, axis of the computer-generated object and display the
however, comparing the CAD object with the actual cross section as a 2-D profile on the computer screen.
object produced by a CAM protocol is very difficult. Similarly, the object produced by the CAM would
This lack of supporting documentation is under- need to be sectioned in a way that would create the
standable because all investigations dealing with the exact same vertical profile as the one viewed on the
determination of CAD/CAM variations require a sys- computer screen. From these two 2-D profiles, it
tem of measuring the variation. Such measurements would then be possible to select specific locations for
are absolutely essential to determine if the object measurements and comparisons of any measurement
produced at the manufacturing facility is exactly the differences between the screen object and the actual
same as the object designed on the computer. No object.
such measurement systems have been reported.
Therefore, this study was initiated with the purpose Creating the Computer Object
of developing and evaluating a system for measuring
the magnitude of the variation between the CAD die The Procera CADD program was used to generate the
or object created on the computer screen and a repli- test objects for this study, and simple geometric forms,
cated die or object produced by the CAM. The hy- a square and cone, were selected and created in the
pothesis for this investigation was that there are no dif- CADD program. The objects could be projected on
ferences in the measurement data derived from this the computer screen in cross sections vertically
new method and the actual measurement data from through the center of each object. A specific cross-
specific sites on an object created by the computer- section view was selected as a measuring site show-
aided dental design (CADD) software; the accuracy, ing a 180- to 0-degree profile.
ie, the validity and reliability, of this method is high. Measurement locations (height 1, 2, and 3) were
selected for computer measurements of the CADD-
Materials and Methods generated objects, and to establish the CADD value
a special CADD tool, a z reference line, was used.
This study used the CAD/CAM method of the Procera With the superior surface of the objects being zero,
system (Nobel Biocare). The recording process in the line was positioned vertically on the various se-
this system uses a mechanical scanner with a contact lected locations, and the distances, or CADD values,
probe. The collected data are transferred to the com- automatically appeared on the computer screen as
puter as ascii files. These files can be projected onto numeric figures. The CADD values for the cone were
the computer screen as two-dimensional cross sec- A = 8,200 µm (1,500 µm from top), and B = 8,500
tions through the long axis of the die. Cross sections µm (3,000 µm from top); and for the square C =
are available for every degree around the vertical 8,000 µm (3,000 µm from top) (Fig 1).
axis of the die. The prosthetic reconstruction is de-
signed using a specially developed CADD software Creating the Machined Object
program.
Each tooth form has multiple irregularities and geo- The CADD data from the square and the cone were
metric configurations that are unique for that tooth. transmitted as CAM data for the manufacturing
Therefore, it is extremely difficult to measure these process, and a Procera milling machine was used to
tooth forms as they appear on the computer screen and create the objects. Titanium was selected as the
COPYRIGHT © 2001 BY QUINTESSENCE
PUBLISHING CO, INC. PRINTING OF THIS
DOCUMENT IS RESTRICTED TO PERSONAL
USE ONLY. NO PART OF THIS ARTICLE MAY
BE REPRODUCED OR TRANSMITTED IN ANY Volume 14, Number 3, 2001 277 The International Journal of Prosthodontics
FORM WITHOUT WRITTEN PERMISSION
FROM THE PUBLISHER.
New Method to Assess Accuracy of CAD Dahlmo et al

Fig 1 Measurement locations


A, B, and C on the CADD-gener-
ated objects; the cone, left, and
A
the square, right.

B C

180° 0° 180° 0°

Fig 2 Holder with the milling


material (a titanium rod) placed in
the milling machine and adjusted
to locate the guide pin (the rota-
tional reference) to the machining
reference stop, ie, 180 degrees
from where the milling process
started.

Milling tool

Titanium rod

Holder
Reference stop

Machine milling chuck


Guide pin

180° 0°

milling material. A titanium rod was fixed in a ma- Measuring the CAM Object
chining holder device used in the normal CAM pro-
duction process. This holder was designed with a ro- A computer numeric controlled (CNC) milling ma-
tational reference and a guide pin. When placing the chine (Dynamyte 2800) equipped with the same type
holder into the milling machine, the holder was ad- of mounting device as the Procera milling machine
justed to locate the guide pin to the machining ref- was used to grind the machined objects to the cross
erence stop, 180 degrees ahead of where the milling section of 180 to 0 degrees. The position of the
process started, ie, where the first data point around mounting device was defined in the control system
the vertical axis of the rod would be produced (Fig of the CNC milling machine, and when positioning
2). This adjustment was calibrated to an accuracy of the holder into the CNC, the guide pin was adjusted
± 0.05 degrees. When the adjustment was com- to correspond to the cross section of 180 to 0 degrees
pleted, a machine tool milled the titanium blank in the CAD program. The square and the cone were
into a square and a cone form around the vertical then ground through the center of the position to the
axis of the holder. cross section of 180 to 0 degrees (Fig 3).
COPYRIGHT © 2001 BY QUINTESSENCE
PUBLISHING CO, INC. PRINTING OF THIS
DOCUMENT IS RESTRICTED TO PERSONAL
USE ONLY. NO PART OF THIS ARTICLE MAY
The International Journal of Prosthodontics 278 Volume 14, Number 3, 2001 BE REPRODUCED OR TRANSMITTED IN ANY
FORM WITHOUT WRITTEN PERMISSION
FROM THE PUBLISHER.
Dahlmo et al New Method to Assess Accuracy of CAD

Fig 3 Holder with the milled ob-


ject positioned into a mounting
device on the CNC milling ma- Object
chine and cut to the cross section
of 180 to 0 degrees.

Holder

Reference stop
Fixture Guide pin

180°

A measuring microscope (Nikon SMZ-U), a con- the measurement procedure were all performed using
nected video camera (Sony Color Videoprint a standardized protocol. An individual experienced
Mavigraph), and a coordinate table with micrometer in the operation of the CAM system performed the ma-
gauges (Mitutoyo) were used for the measurement of chining of the two objects. Individuals trained in its
the machined objects (Fig 4). An image-analysis system use operated the measurement equipment. During the
(Boeckeler VIA 100) and a fiberoptic system (Heim LQ measuring procedure, the distance coordinate value
1600) were used in the measurement procedure. On that would have normally appeared on the screen was
the coordinate table, a mounting device permitted the hidden to the operator. The reason for blinding the op-
placement of the holder with the cross-sectioned ob- erator to this value was to eliminate any bias on the
jects parallel to the table. In this way, profiles from the part of the operator, who would be performing the
square and the cone could be recorded and visualized measurements of the same object several times. In ad-
on the video screen. The mounting device allowed for dition, the numeric values of the CADD-generated ob-
alignment and fastening of the holder onto the coor- jects were unknown to the operators.
dinate table so that the vertical axis of the holder cor- In experiment 1, five samples of the square and the
responded to the Y axis on the coordinate table. To cone were milled, ground (to produce the cross-sec-
view a profile showing the selected cross section from tional view), and measured. All objects were mea-
180 to 0 degrees, the guide pin was oriented to the ref- sured independently by two operators. Each opera-
erence stop on the mounting device in a relationship tor measured each object on two different occasions.
that corresponded to the first data or milling point (0 During each measurement session the object was
degrees) in the rotation. The first step in measuring was repositioned, and thus the total measuring proce-
to determine the zero point on the video screen, the su- dure was repeated. A total of 20 measurements on
perior surface of the square and cone, respectively, for each geometric shape were performed. This experi-
the control unit in the analyzing system. From this mental protocol provided information about the dif-
zero point, the measurement distances A, B, and C ference between the measured objects and the nom-
were identified for the control unit on the video screen. inal value (CADD value), interoperator difference
(difference between operators), variability between
Study Design objects (cross-sectional views), and intraoperator
variability (differences between independent mea-
Two experiments were performed during this study. surements performed on the same object by the same
The manufacturing process, grinding process, and operator, also called measurement error).
COPYRIGHT © 2001 BY QUINTESSENCE
PUBLISHING CO, INC. PRINTING OF THIS
DOCUMENT IS RESTRICTED TO PERSONAL
USE ONLY. NO PART OF THIS ARTICLE MAY
BE REPRODUCED OR TRANSMITTED IN ANY Volume 14, Number 3, 2001 279 The International Journal of Prosthodontics
FORM WITHOUT WRITTEN PERMISSION
FROM THE PUBLISHER.
New Method to Assess Accuracy of CAD Dahlmo et al

Fig 4 Holder with the cross-sec-


tioned object placed in a mount-
Video head Video screen ing device on the measurement
coordinate table with the guide
pin oriented to the reference stop.
Profiles from the object are visu-
alized on the video screen, and
the measurement locations can
be identified for the control unit on
the video screen.
180° 0°
Measuring
microscope

Guide pin
Holder V block
Object

Coordinate table

Reference stop

In experiment 2, one cone was measured by three Results


independent operators. Each operator performed 30
independent measurements. For each measurement All mean distance values for the three locations in ex-
the object was repositioned, and thus the total mea- periment 1 were greater than the CADD values (Table
suring procedure was repeated. This experimental 1). For the square, C, the mean distance value was
protocol provided information about the difference 15.5 µm greater than the CADD value. For the cone,
between the measured object and the nominal value there was also a systematic error; the total mean
(CADD value), the difference between the three op- value for A and B was 4.2 µm and 7.0 µm greater than
erators, and the measurement error (a more precise the CADD value, respectively.
estimate than in the first experiment). In experiment 2, including only one object, the sys-
tematic error was negative; the mean for all three op-
Statistical Analysis erators was less than the CADD value (Table 2). The
mean measured for location A differed from the CADD
In the first experiment, a two-way analysis of variance value by only 7.0 µm and for location B by only 6.6 µm.
(ANOVA) was used. The objects were considered as The difference between the two operators from
a random sample from the manufacturing and grind- experiment 1 and between the three operators from
ing process. Thus, differences between objects (cross- experiment 2 was at most 7.2 µm (Tables 3 and 4).
sectional views) were considered as a random factor The estimates of standard deviation (SD) by differ-
in the model. The possible difference between oper- ent sources of variation (experiment 1) gave at most
ators was treated as a fixed factor in the model. The a total variation of 7.7 µm (Table 5). The estimates of
model also contained a random error, interpreted as SD caused by measurement error were smaller in ex-
the measurement error (intraoperator variability). The periment 2 (3.1 µm).
objective of this analysis was to separate and estimate
the different sources of variation and to estimate the Discussion
eventual difference between the two operators.
In the second experiment, a one-way ANOVA was The overall focus of this project was to develop a
used, also with the eventual differences between op- measurement system that is accurate and reliable
erators as a fixed factor. The objective in the analy- and could be used to compare on-screen data or
sis was to estimate the different sources of variation measurements of an object created by the CADD
and possible differences between operators. software to actual measurement data collected from
COPYRIGHT © 2001 BY QUINTESSENCE
PUBLISHING CO, INC. PRINTING OF THIS
DOCUMENT IS RESTRICTED TO PERSONAL
USE ONLY. NO PART OF THIS ARTICLE MAY
The International Journal of Prosthodontics 280 Volume 14, Number 3, 2001 BE REPRODUCED OR TRANSMITTED IN ANY
FORM WITHOUT WRITTEN PERMISSION
FROM THE PUBLISHER.
Dahlmo et al New Method to Assess Accuracy of CAD

Table 1 Mean A, B, and C Distance Values with Corresponding CADD Values for
Experiment 1

CADD Total Mean values for individual objects, in order


Distance value (µm) mean (µm) Minimum Maximum

A (cone) 8200 8204.2 8198.0 8201.0 8204.0 8204.3 8213.5


B (cone) 8500 8507.0 8498.3 8501.8 8507.5 8512.5 8515.0
C (square) 8000 8015.5 8010.5 8014.5 8015.5 8017.5 8019.5

Table 2 Mean A and B Distance Values with Corresponding CADD Values for
Experiment 2

CADD Total By operator


Distance value (µm) mean (µm) Minimum Median Maximum

A (cone) 8200 8193.0 8190.6 8192.8 8195.6


B (cone) 8500 8493.4 8490.6 8492.7 8496.9

specific locations on the object manufactured from Table 3 True Mean Difference Between the Two
the CADD program. Producing measurement values Operators for Experiment 1
that are extremely close to the CADD value for the Estimated 95% confidence
manufactured objects created by the CADD program Distance differences (µm) interval
will demonstrate that the measurement method is
A (cone) 6.9 3.1–10.7
accurate and that the CADD is also able to produce B (cone) 7.2 3.7–10.7
accurate on-screen measurement values. C (square) 1.0 –5.1–7.1
Experiment 1 provided information about the vari-
ation in five samples produced for each of the
shapes. Data from this experiment also provided in- Table 4 True Mean Difference Between the Operators
formation about the intraoperator variability (mea- for Experiment 2
surement error). In Table 1, all mean distance values Estimated 95% confidence
for all three locations were greater than the CADD Distance differences (µm) interval
values; the fact that the mean distance value is
A (cone)
greater for the square might indicate an error in the Operator 1 vs 2 2.2 0.6–3.8
process of grinding into the selected cross section. Operator 1 vs 3 5.0 3.4–6.6
If the cross section is incorrectly produced, the value Operator 2 vs 3 2.8 1.2–4.4
B (cone)
of distance C will increase because of its squared Operator 1 vs 2 2.1 0.9–3.3
shape. Also included in the squared range was the Operator 1 vs 3 6.3 5.1–7.4
production error, which was found in an earlier in- Operator 2 vs 3 4.1 3.0–5.3
vestigation.15 Interoperator variability was analyzed
by ANOVA in both experiments. In Tables 3 and 4,
the analyses of interoperator variability are pre- Table 5 Estimates of Standard Deviation (SD) by
sented; because the confidence interval in all cases Different Sources of Variation for Experiment 1
did not contain the zero, all differences were statis-
SD of variation
tically significant (5% level of significance). because of SD of variation
However, the difference was at most estimated to be measurement because of SD of total
7.2 µm. The estimated difference between operators Distance error (µm) objects (µm) variation (µm)
was smallest for the squared shape. Because both op- A (cone) 3.9 5.4 6.8
erators performed the same number of observations, B (cone) 3.6 6.8 7.7
a difference of 7.2 µm means that one operator’s C (square) 6.3 1.1 6.5
mean value was 3.6 µm greater than the average (in
this case 8507.0 µm according to Table 1) and the
other operator’s mean value was 3.6 µm less than the various location distances for the two objects. One
average. The small confidence intervals indicated can conclude from the data that different operators
that little measurement difference occurred between can use this method with small differences, ie, high
the two operators when they were measuring the interoperative reliability.
COPYRIGHT © 2001 BY QUINTESSENCE
PUBLISHING CO, INC. PRINTING OF THIS
DOCUMENT IS RESTRICTED TO PERSONAL
USE ONLY. NO PART OF THIS ARTICLE MAY
BE REPRODUCED OR TRANSMITTED IN ANY Volume 14, Number 3, 2001 281 The International Journal of Prosthodontics
FORM WITHOUT WRITTEN PERMISSION
FROM THE PUBLISHER.
New Method to Assess Accuracy of CAD Dahlmo et al

Both experiments were performed with replicates, variability because of differences between objects was
ie, several independent measurements on the same ob- higher for the conical shape than for the square. This
ject performed by the same operator. This provides the is because the conical shape is more complex and the
opportunity to estimate the intraoperator variability distance value A and B is dependent on the height.
by experiment. Furthermore, because the first experi- The variability because of intraoperator variability
ment also contained several different objects, it also could actually be improved. It is rather easy and not
provided the opportunity to estimate variation caused that time consuming to take replicates and use the
by objects (cross-sectional views). Estimates of varia- mean value. For instance, by using four measure-
tion in terms of SD from different sources are pre- ments instead of one, the intraoperator variability is
sented in Table 5; the total variation was nearly the reduced by 50%. The intraoperator variability and
same for both shapes. The SD for the cone was only variability because of differences between objects is
slightly higher. For the square, most of the variation was rather small and could be regarded as acceptable.
because of measurement error. For the cone, the SD Thus, the reliability is acceptable.
for the measurement error was smaller than the SD be-
cause of objects. Conclusions
The most pessimistic interpretation is that the total
systematic error on the two shapes could be 15.5 µm Under the assumption that the manufacturing process
(greatest systematic error) plus 3.6 µm (the operator was without systematic error, the measuring method
with the greatest systematic error), which gives a (ie, grinding and measuring) gave a positive system-
“worst-case systematic error” of 19.1 µm. In the total atic error, ie, observed values were greater than the
systematic difference from the CADD value, the op- true value. For all objects and geometric shapes, the
erator error was the smallest part. Nevertheless, analy- systematic error was at most 15.5 µm. The estimated
sis of the reasons for operator differences could pro- difference between operators, interoperator differ-
vide a means to calibrate operators and decrease ence, was so small that it is of no relevance. The vari-
differences. ation because of measurement error, intraoperator
The other part of the systematic difference could be variability, was slightly lower in the second experi-
because of the grinding process and/or measurement ment, and it was greater for the square object com-
process. This difference could be reduced by analy- pared to the cone. However, the variation because of
ses and refined processes. However, a total system- objects was higher for the cone object than for the
atic difference of at most 19.1 µm must be considered square. The total standard deviation was only 7.7 µm.
acceptable. Thus, the validity is acceptable. Thus, the total random error caused by object varia-
When comparing the two operators performing tion and measurement error was in approximately
measurements on the three distances, the estimated two thirds of all measurements lower than 8 µm. In ap-
difference between the two operators was at most 7.2 proximately 95% of all measurements, the random
µm. Although this difference is statistically significant, error will be lower than 15 µm. Within the limitations
it is from a practical point of view a small value. Thus, of this investigation, it can be concluded that:
the interoperator reliability, even in the worst case,
could be considered acceptable. Nevertheless, it may 1. There are no differences in the measurement data
be possible to reduce this difference. derived from this new method and actual mea-
The variability, in terms of SD, because of random- surement data from specific sites on an object cre-
ness, ie, intraoperator variability and differences be- ated by the CADD program. The method has high
tween objects, was at most 7.7 µm. A standard devia- validity and reliability, ie, high accuracy.
tion of 7.7 µm means that differences from the CADD 2. The new measurement method provides an exact
value because of randomness will be lower than 7.7 µm measure to determine the capabilities of the on-
in 68% of all cases; in 95% of all cases, the difference screen designing process of the Procera CAD/
because of randomness will be lower than 15.4 µm. CAM system. The method can be used when de-
Observe that the random difference should be added termining the clinical acceptability of dental
to the systematic difference. The size of the systematic restorations.
difference depends on which object is to be measured
and by which operator (earlier estimated to be at most References
approximately 19 µm). Regarding the squared shape,
1. Leinfelder KF, Isenberg BP, Essig ME. A new method for gener-
most of the variability was because of intraoperator
ating ceramic restorations: A CAD-CAM system. J Am Dent Assoc
error. Regarding the conical shape, variability because 1989;118:703–707.
of differences between objects was a little higher than 2. McLaughlin G. The reality of dental CAD-CAM: Hype or hope?
intraoperator variability. It seems reasonable that the Texas Dent J 1992;109:21–25.
COPYRIGHT © 2001 BY QUINTESSENCE
PUBLISHING CO, INC. PRINTING OF THIS
DOCUMENT IS RESTRICTED TO PERSONAL
USE ONLY. NO PART OF THIS ARTICLE MAY
The International Journal of Prosthodontics 282 Volume 14, Number 3, 2001 BE REPRODUCED OR TRANSMITTED IN ANY
FORM WITHOUT WRITTEN PERMISSION
FROM THE PUBLISHER.
Dahlmo et al New Method to Assess Accuracy of CAD

3. Duret F, Preston JD. CAD/CAM imaging in dentistry. Curr Opin 17. Bergman B, Bessing C, Ericson G, Lundquist P, Nilson H, Andersson
Dent 1991;1:150–154. M. A 2-year follow-up study of titanium crowns. Acta Odontol
4. Duret F. The practical dental CAD/CAM in 1993. J Can Dent Assoc Scand 1990;48:113–117.
1993;59:445–453. 18. Andersson M, Odén A. A new all-ceramic crown. A dense-sin-
5. Rekow ED. CAD/CAM in dentistry. Alpha Omegan 1991;84(4): tered, high-purity alumina coping with porcelain. Acta Odontol
41–44. Scand 1993;51:59–64.
6. Rekow D. Dental CAD-CAM systems: What is state of the art? J 19. Karlsson S. The fit of Procera titanium crowns. An in vitro and
Am Dent Assoc 1991;122:43–48. clinical study. Acta Odontol Scand 1993;51:129–134.
7. Rekow DE, Erdman AG, Riley DR, Klamecki B. CAD/CAM for 20. Siervo S, Bandettini B, Siervo P, Falleni A, Siervo R. The CELAY
dental restorations—Some of the curious challenges. IEEE Trans System: A comparison of the fit of direct and indirect fabrication
Biomed Eng 1991;38:314–318. techniques. Int J Prosthodont 1994;7:434–439.
8. Rekow ED. A review of the developments in dental CAD/CAM 21. Valderrama S, van Roekel N, Andersson M, Goodacre CJ, Munos
systems. Curr Opin Dent 1992;2:25–33. CA. A comparision of the marginal and internal adaption of tita-
9. Winkler P. Production of individual dental restorations using nium and gold-platinum-palladium metal ceramic crowns. Int J
CAD/CAM technology. IFIP Trans B (Appl Technol) 1992;2:3–8. Prosthodont 1995;8:29–37.
10. van der Zel JM. Ceramic-fused-to-metal restorations with a new 22. Thompson JY, Bayne SC, Heymann HO. Mechanical properties
CAD/CAM system. Quintessence Int 1993;24:769–778. of a new mica-based machinable glass ceramic for CAD/CAM
11. LoPresti JT, David SB. New concepts in restorative dentistry. The restorations. J Prosthet Dent 1996;76:619–623.
CEREC CAD/CAM system. N Y State Dent J 1994;60:39–42. 23. May K, Russel M, Razzoog M, Lang B. Precision of fit of the
12. Persson M, Andersson M, Bergman B. The accuracy of a high-pre- Procera AllCeram crown. J Prosthet Dent 1998;80:394–404.
cision digitizer for CAD/CAM of crowns. J Prosthet Dent 1995;74: 24. Razzoog ME, Lang BR. Research evaluations of a new all-ce-
223–229. ramic system. Pract Periodontics Aesthet Dent 1998;10(suppl
13. Russel M, Andersson M, Dahlmo K, Razzoog M, Lang B. A new 1):1–3.
computer-assisted method for fabrication of crowns and fixed 25. Andersson M, Razzoog ME, Odén A, Hegenbarth EA, Lang BR.
partial dentures. Quintessence Int 1995;26:757–763. Procera: A new way to achieve and all-ceramic crown.
14. Sohmura T, Takahashi J. Use of CAD/CAM system to fabricate Quintessence Int 1998;29:285–296.
dental prostheses. Part 1: CAD for a clinical crown restoration. 26. Odén A, Andersson M, Krystek-Ondracek I, Magnusson D. Five-
Int J Prosthodont 1995;8:252–258. year clinical evaluation of Procera AllCeram crowns. J Prosthet
15. Andersson M, Carlsson L, Persson M, Bergman B. Accuracy of Dent 1998;80:450–456.
machine milling and spark erosion with a CAD/CAM system. J 27. Persson M, Andersson M, Bergman B. The accuracy of a high-pre-
Prosthet Dent 1996;76:187–193. cision digitizer for CAD/CAM of crowns. J Prosthet Dent 1995;74:
16. Andersson M, Bergman B, Bessing C, Ericson G, Lundquist P, 223–229.
Nilson H. Clinical results with titanium crowns fabricated with
machine duplication and spark erosion. Acta Odontol Scand
1989;47:279–286.

Literature Abstract

The relationship between cigarette smoking and dental implant failure.

This article presents the results of a retrospective clinical study that shows the effect of to-
bacco on implant treatment. The records of 56 implant patients of a single surgeon were re-
viewed. A total of 187 Brånemark (Nobel Biocare) endosseous dental implants had been
placed over a 4-year period for removable and single- or multiple-unit fixed prostheses. A stan-
dard data sheet was used to collect information including patient gender, age, medical and
dental history, status regarding smoking, intraoral location of each implant, reason for failure,
and bone type at the placement site, ie, grafted or natural bone. A Chi-squared test was used
to compare implant failure rates in smokers and nonsmokers by whole mouth and by specific
location. This study demonstrated a significant association between implant failure rates and
cigarette smoking, with failure rates of 17% in smokers compared to 7% in nonsmokers.
Implant length was also shown to be a significant factor, with shorter implants (≤ 10 mm) being
more susceptible to failure in smokers. It was concluded that these results concur with many
other studies of the topic, and many surgeons have recommended a complete, if only short-
term, cessation of smoking in the perioperative period for implant placement. Smoking should
be considered a relative contraindication to implant placement when planning treatment.

Wallace RH. Eur J Prosthodont Restorative Dent 2001;8:103–106. References: 23. Reprints: Dr Robert H.
Wallace, Senior House Officer, Department of Restorative Dentistry, Newcastle Dental Hospital, Richardson
Road, Newcastle-upon-Tyne NE2 4AZ, United Kingdom. Fax: + 0191 2275149—AW

COPYRIGHT © 2001 BY QUINTESSENCE


PUBLISHING CO, INC. PRINTING OF THIS
DOCUMENT IS RESTRICTED TO PERSONAL
USE ONLY. NO PART OF THIS ARTICLE MAY
BE REPRODUCED OR TRANSMITTED IN ANY Volume 14, Number 3, 2001 283 The International Journal of Prosthodontics
FORM WITHOUT WRITTEN PERMISSION
FROM THE PUBLISHER.

Вам также может понравиться