See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/46404436

Volumetric Determination of the Amount of Misfit in CAD/CAM and Cast

Implant Frameworks: A Multicenter Laboratory Study

Article  in  The International journal of oral & maxillofacial implants · September 2010

Source: PubMed

CITATIONS READS

58 559

4 authors, including:

Carl Drago Daniel Domagala

Marquette University Marquette University
39 PUBLICATIONS   1,005 CITATIONS    2 PUBLICATIONS   65 CITATIONS   

SEE PROFILE SEE PROFILE

Riad Almasri
Eternim Dental Implant Center
2 PUBLICATIONS   92 CITATIONS   

SEE PROFILE

All content following this page was uploaded by Riad Almasri on 11 March 2016.

The user has requested enhancement of the downloaded file.

 Volumetric Determination of the Amount of
Misfit in CAD/CAM and Cast Implant Frameworks:
A Multicenter Laboratory Study
Carl Drago, DDS, MS1/Roxana L. Saldarriaga, DDS2/Daniel Domagala, DDS, MS3/Riad Almasri, DDS4

Purpose: The purposes of this study were (1) to evaluate the fit between implant frameworks and
implants fabricated with two types of implant framework fabrication techniques: computer-aided
design/computer-assisted machining (CAD/CAM) and conventional casting with the lost wax technique;
and (2) to describe a digital measurement system consisting of tactile scanning and computer software
programs that measured the volumetric differences between implant-supported frameworks and implant
restorative platforms fabricated with these technologies. Materials and Methods: This laboratory study
used acrylic resin models with five interforaminal implants. The models were scanned; implant-level
impressions and verification indexes were then made to construct master casts. First, a cast gold alloy
framework and a titanium milled bar fabricated with CAD/CAM technology were made to clarify the
construction processes of each. After this pilot study was completed, five cast and five CAD/CAM frame-
works were made at each of three dental schools (15 milled and 15 cast bars). Each framework was
made on a master cast from individual impressions. The implant restorative interfaces of the frameworks
were scanned, and the data were entered into a computer software program. The virtual representations
of the frameworks were fit onto digitized scans of the implant restorative platforms and used for virtual
one-screw tests on both sides of the arch. Volumetric differences between the implant restorative plat-
forms of the implant-supported frameworks and the model implants were measured to determine the
amount of misfit between the frameworks and the model implants. Results: Implant-supported frame-
works made with the CAD/CAM technology fit significantly better onto the implants than the cast implant
frameworks. There was a significant difference between the right and left one-screw tests; there were
no significant differences among the three university sites. Conclusions: The CAD/CAM frameworks
featured in this study were significantly more accurate than cast frameworks made with the lost-wax
technique. INT J ORAL MAXILLOFAC IMPLANTS 2010;25:920–929

Key words: computer-aided design, gold, implant-supported frameworks, precision of fit, volumetric
assessment

ultiple-unit implant prostheses demand the fab-

1Associate Professor, Department of Restorative and Prosthetic
Dentistry, The Ohio State University College of Dentistry; former
Director, Dental Research, Biomet 3i, Palm Beach Gardens,
Florida.
2Prosthodontic Resident, University of Minnesota School of Den-
tistry, Minneapolis, Minnesota.
3Prosthodontic Resident, Marquette University School of Den-
tistry, Milwaukee, Wisconsin.
4Resident in Postgraduate Prosthodontics, Nova Southeastern
University, College of Dental Medicine, Fort Lauderdale, Florida.
Correspondence to: Dr Carl Drago, Biomet 3i, 4555 Riverside
Drive, Palm Beach Gardens, FL 33410. Fax: +561-514-6312.
Email: carl.drago@biomet.com
Presented at the Academy of Osseointegration Annual Meeting,
Boston, MA, Feb 29, 2008.
M rication of precise screw-retained frameworks.
Even with careful, controlled clinical and laboratory
techniques, ill-fitting frameworks may still be pro-
duced. In the case of macroscopically ill-fitting frame-
works, implant-supported frameworks may need to
be sectioned, soldered, welded, or discarded and
remade. 1–4 The importance of precise fit between
implant-supported frameworks and implants has
been discussed extensively. 5–7 An imprecise fit
between implant-supported frameworks and
implants may contribute to marginal bone loss
around implants.8–10 Absolute passive fit between
implants and implant frameworks might be impossi-
ble to achieve clinically.7,11

920 Volume 25, Number 5, 2010

© 2010 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 FORM WITHOUT WRITTEN PERMISSION FROM THE PUBLISHER.

In vivo measurements of implant prostheses with
cast frameworks have been carried out by several
researchers.12,13 Jemt reported on 7 maxillary and 10
mandibular prostheses in edentulous jaws. Measure-
ments were made on master cast abutment analogs
and intraorally with a three-dimensional photogram-
metric technique. The results indicated that prosthe-
ses connected to osseointegrated implants
demonstrated distortion between the frameworks
and implants of up to several hundred microns. 12
Jemt and Book prospectively studied prosthesis mis-
fit and radiographic bone loss in 14 patients (7 maxil-
lae, 7 mandibles) for 1 year and retrospectively for the
last 4 years of a 5-year period after second-stage
surgery. 7 Their results demonstrated that, in vivo,
none of the prostheses presented a completely pas-
sive fit to the implants. Furthermore, similar misfits
were found in both groups (1 year and 5 years), indi-
cating that the implants seemed to be stable, even
after several years in function. The mean center point
misfit was 111 µm (SD 59) and 91 µm (SD 51) for the
1- and 5-year groups, respectively, with a range of 275
µm. The corresponding mean marginal bone loss was
0.5 and 0.2 mm, respectively, for the two follow-up
groups. They noted no statistical correlations (P > .05)
between changes in marginal bone levels and differ-
ent degrees of prosthesis misfit in either group. Jemt
and Book speculated that implants may possess a
certain biologic tolerance for ill-fitting frameworks.
New technologies have been developed to fabri-
cate accurate metal implant–supported frame-
works. 13–17 Titanium frameworks have been
proposed as alternatives to conventional gold alloy
castings in implant dentistry 16–19 since they have
demonstrated more precision and predictability than
conventional cast metal frameworks.13–15 Reidy et al
published the results of a laboratory study where
they compared the fit between implants and implant
frameworks fabricated with a laser-welding protocol
and castings made with the lost wax technique.15
Reidy et al reported that the laser-welded frame-
works exhibited a more precise fit than the one-piece
castings, with significant differences at four of the
five prosthodontic interfaces in the z-axis. Ortorp et
al investigated the accuracy of frameworks (n = 20)
made with a computer-aided design/computer-
assisted machining (CAD/CAM) technology called
computer numeric control (CNC) and those made via
casting conventional type IV gold alloy (n = 5).20 They
reported that the CNC frameworks demonstrated a
statistically better fit and precision of fabrication
compared to conventional castings (P < .05), and
they concluded that it was possible to fabricate CNC
implant-supported frameworks with high precision
and repeatability.
© 2010 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 FORM WITHOUT WRITTEN PERMISSION FROM THE PUBLISHER.
Drago et al
Ortorp and Jemt continued this research with a
10-year prospective study that included 126 patients
who were randomly provided with 67 CNC frame-
works and 62 gold alloy castings.21 The clinical and
radiographic 10-year data revealed few problems and
similar clinical and radiographic performance in both
groups. No implants were lost after the 5-year follow-
up visits. The cumulative survival rates for the CNC-
milled group were 95% and 95.1% for implants and
prostheses, respectively. The casting groups’ cumula-
tive survival rates were 97.9% and 98.3%, respectively.
One prosthesis was lost in each group secondary to
implant loss; one prosthesis was lost in the CNC
group because of framework fracture. Two framework
fractures were noted in each group. Ortorp and Jemt
concluded that CNC-milled titanium frameworks
were viable alternatives to gold alloy castings in
edentulous jaws during the 10 years of function, as
reported in their study.
A study by Al-Fadda et al with similar parameters
as the aforementioned study by Ortorp and Jemt was
recently published.19 This laboratory study included
18 frameworks (9 conventional castings and 9 CNC-
milled titanium frameworks). The CNC-milled frame-
works showed significantly less distortion along the
x-axis compared with conventional castings (means
of 33.7 µm and 49.2 µm, respectively) and in the hori-
zontal plane (56 µm and 85 µm, respectively). The dis-
tortions between the CNC-milled frameworks and
conventional castings were smaller for the CNC-
milled frameworks—but this was not statistically sig-
nificant—in the y-axis, z-axis, and in three
dimensions. Al-Fadda et al concluded that within the
limitations of their study, fabrication of CNC-milled
frameworks yielded a more accurate fit than currently
used casting techniques.
The aforementioned studies have described meth-
ods that identified the degree of misfit between
implant-suppor ted frameworks and implants/
abutments that commonly occur in implant dentistry.
The measurements were linear measurements of mis-
fits between implant-supported frameworks and
implants/abutments. However, clinical misfits
between implant-supported frameworks and intrao-
ral implant components actually occur in three
dimensions, as noted by the x, y, and z measurements
reported in various studies.15–21 The authors of the
present study propose that misfits between frame-
works and implant restorative components should be
expressed as volumetric measurements (cubic mil-
limeters), as this is how misfits occur clinically.
The purposes of the present laboratory study were
twofold. First, the authors sought to evaluate the
accuracy of implant-supported frameworks made
with two different techniques. One technique used a
The International Journal of Oral & Maxillofacial Implants 921
 Drago et al
Fig 1 Occlusal view of one university model with five interforami-
nal implants.
tactile-probe CAD/CAM technology, and the other
technique used the lost-wax technique and conven-
tional casting technology. The second purpose of the
study was to describe a system that included tactile
scanning with a probe and a computer software pro-
gram that recorded and measured the volumetric dif-
ferences between implant-supported frameworks
(CAD/CAM versus lost-wax technique and conven-
tional casting) and implants. The volumetric fit
between the implant-supported frameworks (implant
restorative platforms) and the implants was assessed
by using a computer software program that virtually
aligned the implant/abutment connections of the
frameworks with the implant restorative platforms of
the distal implants, as in clinical one-screw tests. The
process was performed on both sides of the jaws. The
protocol used to fabricate the frameworks was con-
sistent with a clinical protocol used to verify implant-
supported framework fit in clinical situations.
METHODS AND MATERIALS
A pilot laboratory study (n = 1) was conducted first to
identify the appropriate steps required for the defini-
tive study. This laboratory study used acrylic resin
models at each of three university testing centers. Five
internal-connection 4.1-mm-diameter implants
(IOSS410, Biomet 3i) were placed into the acrylic resin
models in a fashion consistent with a typical edentu-
lous mandible. The implants were placed interforami-
nally (Fig 1). All of the implant restorative platforms
were 2 to 3 mm supragingival. Because of time con-
straints, each university model was similar to the oth-
922 Volume 25, Number 5, 2010
© 2010 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 FORM WITHOUT WRITTEN PERMISSION FROM THE PUBLISHER.
ers, but the models were not exact duplicates. There-
fore, the frameworks made at one university would
not fit onto a model made for a different dental
school. All models were made using heat-processed
acrylic resin.
Implant-Level Impression
At each university center, 10 standardized master
casts were made of each model from 10 separate
impressions by using a controlled and reproducible
technique. Ten custom light-cured impression trays
(Triad, Denstply Int) were fabricated with a standard-
ized technique to ensure reproducible impression
volumes. Openings were provided in each of the trays
to accommodate pickup impression copings. New
pickup implant-level impression copings were used
for each impression (IIIC41, Biomet 3i). Ten individual
polyether (Impregum, 3M/ESPE) implant-level
impressions, in custom impression trays, were made
of the models. All impressions were made with an
open-tray protocol. One set of impressions was used
for the CAD/CAM frameworks; the second set was
used for the cast gold alloy frameworks. New implant
analogs were attached to the apical surfaces of the
impression copings, and primary casts were poured
in Type IV dental stone per the manufacturer’s
instructions (measured, weighed, and vacuum spatu-
lated; Prima Rock, Whip Mix Corporation). These casts
were designated as the primary casts. To account for
any differences that may have existed prior to making
the definitive impressions, each cast was assigned
randomly to one of the two groups.
Verification Indexes/Impressions/
Definitive Master Casts
Since the implants were to be splinted together
within the frameworks, it was not necessary to identify
the location of the internal-hex implant connections.
The impression technique was designed to be consis-
tent with impression techniques recognized for fabri-
cating accurate master casts. 22 Nonhexed metal
provisional cylinders (ITCS42, Biomet 3i) were placed
onto the analogs in the primary casts. Autopolymeriz-
ing acrylic resin (GC Pattern Resin, GC America) was
used to fabricate one-piece verification indexes that
splinted the provisional cylinders together on each
primary cast. After the resin polymerized, the indexes
were sectioned into individual segments (separating
disks < 1 mm thick) and continued to set for at least
24 hours.22 The individual segments were then placed
back onto the models. Care was taken to ensure that
none of the segments contacted each other, and the
segments were luted together with new autopoly-
merizing acrylic resin. The indexes set undisturbed for
15 minutes.

Fig 2 Occlusal view of five master casts from one university center.
With the indexes in place, new definitive impres-
sions were made with polyether impression material
(Impregum). The definitive impressions were made in
new custom impression trays that allowed for the
presence of the verification indexes and equal vol-
umes of impression material. The trays were made to
fit over the indexes without contact. Impression
material was injected around the verification indexes
and the indexes were picked up inside the definitive
impressions. Five new implant lab analogs were
attached to the apical surfaces of the cylinders in
each impression, and 10 definitive master casts were
poured with Type IV dental stone per the manufac-
turer’s instructions (measured, weighed, and vacuum
spatulated; Prima Rock). These casts were designated
as the master casts. To control for any differences that
may have existed in making the definitive impres-
sions and verification indexes, each master cast was
assigned randomly to one of the two groups (Fig 2).
Scanning and CAD
The models and the master casts were sent to the
dental implant manufacturer for processing. The
implant restorative platforms of the models and
implant lab analogs in the master casts were scanned
with a tactile (touch-probe) scanner (Vertex Auto-
matic Measuring System, Model 220, Florida Metrol-
ogy) (Fig 3). The scans were accomplished per the
protocol for this par ticular CAD/CAM process
(ARCHITECH PSR, Biomet 3i). The only difference
noted with this in vitro protocol was that, with this
particular scanner, clinical implants could not be
scanned. The scans of the implants were to be used in
© 2010 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 FORM WITHOUT WRITTEN PERMISSION FROM THE PUBLISHER.
Drago et al
Fig 3 Tactile probe
scanning unit.
the volumetric assessments. The scans of the master
casts were used to design and mill the frameworks.
The frameworks were designed virtually by one CAD
designer. The CAD/CAM bars were designed for use
as primary overdenture implant-supported frame-
works with the following parameters: parallel walls (6
mm buccolingual width, 4 mm vertical height), 2-mm
cantilevers measured from the distal surfaces of the
distal implants, and no attachments (Figs 4 and 5).
Milling
The bars were milled from blanks of titanium-alu-
minum-vanadium alloy within the same verification
time period on one milling machine (Fig 6). The bars
were finished as if for a clinical patient try-in appoint-
ment. All polishing and finishing procedures were
accomplished with polishing protectors in place on
the framework restorative platforms. The implant
restorative platforms of the milled bars were then
scanned with the tactile scanner used earlier. The
milled frameworks remained at the manufacturing
facility until the cast frameworks were received at the
facility.
Cast Bars
The same protocol and materials were used at each
university center to fabricate the cast frameworks. Cast
bars were designed for use as primary overdenture
implant-supported frameworks with the same para-
meters used to fabricate the CAD/CAM frameworks:
parallel walls (6 mm buccolingual width, 4 mm vertical
height), 2-mm cantilevers measured from the distal
surfaces of the distal implants, and no attachments.
The International Journal of Oral & Maxillofacial Implants 923
Drago et al
Fig 4 (above) Lingual image of the CAD design for the CAD/CAM
frameworks used at one university center.
Fig 5 (right) Occlusal view of the CAD design for the CAD/CAM
framework used at one university center.
Nonhexed UCLA cylinders (IGUCA2C, Biomet 3i) were
placed into the implant lab analogs in the definitive
master casts. Internal-hex connections were used, as
this represents the most popular implant/ abutment
connection for this particular implant company. The
abutments were manufactured with machined gold
alloy implant restorative platforms and plastic burnout
cylinders (Fig 7). Vigolo et al22 compared the accuracy
at implant interfaces of gold-machined UCLA-type
abutments and CAD/CAM titanium abutments with
external- and internal-hex connections. They reported
that there were no significant differences relative to
rotational freedom among the groups and concluded
that both types of abutments (gold-machined UCLA-
type and CAD/CAM titanium) consistently showed 1
degree of rotational freedom between implants and
abutments for external- and internal-hex connections.
For this reason, the present authors decided to per-
form the study using only internal-hex connections
and gold machined UCLA Abutments.
At each university center, the bars were sprued
with similar designs and materials (6-gauge [4.1-mm-
diameter] wax runner bars), and the framework pat-
terns were connected to the bars with 10-gauge
(2.6-mm-diameter) wax sprues 3 mm long. The bars
were waxed to the sprue bases with 6-gauge (3.3-
mm-diameter), 12-mm-long wax sprues that started
from a domed central entry point. There was no more
than 6 mm of investment from the top of the patterns
to the tops of the investment rings. Venting sprues
were used to improve gas escape during casting.
The wax patterns were invested in phosphate-
bonded (carbon-free) investment per the manufac-
turer’s instructions (GC Fujivest Powder, GC America
Inc). The rings were placed into ovens and the
burnout temperatures were set to 815°C. The alloys
924 Volume 25, Number 5, 2010
© 2010 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 FORM WITHOUT WRITTEN PERMISSION FROM THE PUBLISHER.
were cast at approximately 1,450°C (Olympia,
Jelenko). The rings were then allowed to cool to room
temperature. The frameworks were divested by
removing the bulk of the investment with 100-µm-
diameter aluminum oxide, with care taken to avoid
the implant restorative platforms. The internal sur-
faces of the screw access openings were blasted with
80-µm glass beads.
Polishing protectors were placed onto the implant
restorative interfaces prior to initiating definitive pol-
ishing and finishing procedures. The frameworks
were finished as if for clinical patient try-in appoint-
ments (Fig 8).
None of the frameworks were sectioned and sol-
dered/welded to obtain better fits between the implant
restorative platforms and the implants, since solder-
ing/welding would introduce variables in the strength
of the frameworks and the integrity of the rigid connec-
tors.23 Rather, the objective of this study was to assess
the accuracy of CAD/CAM versus cast frameworks, not
the strength or consistency of rigid connectors. All of
the cast implant-supported frameworks, master casts,
and models were sent to the implant manufacturing
facility, where the restorative platforms of the implant-
supported frameworks and implant restorative plat-
forms of the models were scanned with the tactile
(touch probe) scanner used earlier.
Evaluation
The digitized data of the frameworks’ implant restora-
tive platforms were evaluated for accuracy of fit to the
digitized data of the implants on the respective mod-
els with a computer software program (Auto desk
Inventor Professional 2009, Autodesk). The fit on the
master casts was not evaluated with scans; rather, the
analysis was conducted using a CAD principle called

Fig 6 Occlusal view of a finished CAD/CAM framework on a mas-
ter cast.
Fig 7 Laboratory image of a master cast with five nonhexed
UCLA abutments in place. These abutments were manufactured
with machined gold alloy interfaces for improved accuracy rela-
tive to the implant/abutment interface (inset).
“lofting.” Lofting is well recognized in CAD workplaces
and supports design and measurements in the auto-
motive, architectural, and aeronautical industries.24,25
The scanned implant restorative platforms and
framework abutment interfaces’ digital data were
transformed into computer-generated digitized mod-
els. The virtual frameworks and patient implants were
assembled on the computer screen to simulate sepa-
rate one-screw tests on the right and left distal
implants. For the right one-screw test, the software
program perfectly aligned the implant and restorative
interfaces for the right distal implant. The software
then evaluated the “sections” between the restorative
interfaces for each of the other four implant/restora-
tive interfaces. The volumetric spaces between the
implants and abutments were measured. The second
set of measurements was made with the same proto-
col but with the left distal implant/abutment interface
virtually aligned. The evaluations were not random-
ized because the measurements were accomplished
using the software program.
© 2010 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 FORM WITHOUT WRITTEN PERMISSION FROM THE PUBLISHER.
Drago et al
Fig 8 Occlusal view of a finished cast framework on a master
cast.
Once the frameworks and patient models were
scanned, the coordinates were exported as a text file.
This text file was converted into an Excel spreadsheet
(Microsoft) formatted for import into the CAD soft-
ware. The imported spreadsheet contained xyz points
from the Vertex tactile scanner as follows: four points
per restorative platform (three points for a plane and
one point for the center of the platform), and three
independent points from the calibration implant for
orientation. Within the software program, 10-mm
cylinders were extruded, or lofted, from the center of
the platform to create planes and center points. Next,
7.62 mm (overall length of the cylinders ± 0.0002 inch)
were subtracted from the cylinder to obtain the actual
seating surface of the cylinders. Within the software
program, the frameworks were perfectly seated onto
the patient models, center point to center point. The
sections were rotated in three dimensions to ensure
that the interfaces were perfectly aligned. The frame-
works were aligned to the models using geometric
constraints. The alignment plane was perpendicular to
The International Journal of Oral & Maxillofacial Implants 925
Drago et al
Fig 9 Virtual left one-screw test on the CAD/CAM framework. The
degree of misfit is identified by the red planes.
Fig 11 Virtual left one-screw test on the cast framework. The
degree of misfit is identified by the red planes.
the model plane. After a framework was aligned to the
distal implant seating surface on the model, a model
was created using the seating surfaces of each
restorative platform and corresponding implant. The
volumetric displacement was then extracted from the
model. Separate models were made for the right and
left one-screw tests.
The volumes (in cubic millimeters) represented the
amount of misfit between the implant-supported
framework restorative platforms and the implants on
the patient models. With the right-side virtual one-
screw test, volumetric measurements were recorded
for the four implant/abutment interfaces and aver-
aged. The process was repeated by virtually aligning
the implant/abutment interface on the left side and
repeating the volumetric measurements on the other
four implants. All measurements were done at Biomet
3i (Figs 9 to 12).
Statistical Analysis
The analysis was done using the absolute values of
the volumetric data. If absolute values had not been
used, the differences in one plane would cancel each
other out and the net value would not indicate any
926 Volume 25, Number 5, 2010
© 2010 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 FORM WITHOUT WRITTEN PERMISSION FROM THE PUBLISHER.
Fig 10 Virtual right one-screw test on the CAD/CAM framework.
The degree of misfit is identified by the red planes.
Fig 12 Virtual right one-screw test on the cast framework. The
degree of misfit is identified by the red planes.
misfit. Differences in volumes between the CAD/CAM
and cast frameworks were tested using analysis of
variance with a randomized block design. The differ-
ent sites (University of Minnesota, Marquette Univer-
sity, and Nova Southeastern University) and right
versus left one-screw tests were included as factors in
the analysis. This test was a two-sided test using a .05
significance level.
RESULTS
The volumetric measurements of the misfits between
implants and implant-supported frameworks for the
CAD/CAM and cast frameworks are noted in Fig 13
and in Tables 1 and 2.
The CAD/CAM frameworks fit significantly better
than the cast gold frameworks (P < .0001). On average,
the volumetric misfit between the implant restorative
platforms of the CAD/CAM frameworks and the
implants was 2.25 mm3 less (better) than the corre-
sponding volumetric misfit between the implant
restorative platforms of the cast frameworks and the
implants. There was a significant difference between
 Drago et al

Table 1 Volumetric Misfit Between Frameworks

and Implants on the Models During One-Screw Tests
12 CAM
CAM Site/no. of bars Left misfit (mm3) Right misfit (mm3)
olume of misfit (mm3)

Gold
10 Minnesota
8 5 4.399 3.883
6 5 3.117 4.057
Difference 1.282** 0.174*
4
Volume

Marquette
2 5 1.012 3.827
V

0 5 3.677 6.35
0 5 10 15 20 25 30 Difference 2.665* 2.523*
Implant restorative
Implant restorative interfaces
interfaces Nova
5 1.667 3.136
5 3.384 3.386
Difference 1.717* 0.25*

Fig 13 Different volumes of misfit noted for each of the implant Minnesota = University of Minnesota; Marquette = Marquette
restorative interfaces in this study. University; Nova = Nova Southeastern University.
*CAD/CAM bars more accurate than cast bars; **cast bars more
accurate than CAD/CAM bars.

Table 2 Volumetric Misfit Between Frameworks

and Implants on the Model (Pilot Study) During
One-Screw Tests
Bar Left misfit (mm3) Right misfit (mm3)

CAD/CAM milled bar 1.4369 4.9017

Cast bar 9.8960 6.2798
Difference 8.4591* 1.2781*
*The CAD/CAM bar was more accurate than the cast bar.

the right and left one-screw tests (P < .0001). There
was no significant difference among the three univer-
sity sites (P = .72).
DISCUSSION
In this study, CAD/CAM implant-supported frame-
works fit implants more accurately and with less dis-
tortion than did cast implant-supported frameworks
made with the same design parameters. These results
suggest that the scanning technology used in this
report is a viable method, in a laboratory setting, to
measure the amount of misfit between implant-sup-
ported frameworks and implant restorative compo-
nents. The scanning and measuring processes
accurately measured the small volumetric differences
that may occur in clinical practice. However, at this
time, the scanner is not applicable for clinical imaging
(tactile probing) of implants in patients’ mouths. The
fits of the cast frameworks may have been improved
with sectioning and soldering/welding.
The measurements cited in previous papers were
linear in nature, even though clinical misfits are gen-
erally three-dimensional in nature. No previous pub-
lished study reported misfits between implants and
frameworks volumetrically. Since implants and com-
ponents are three-dimensional objects, and any mis-
fits between implants and components are not all in
the same plane, it seems reasonable that volumetric
measurements will serve as more accurate and clini-
cally relevant assessments of clinical situations than
linear measurements.
The findings cited in this study are consistent with
the findings of numerous authors, in that CAD/CAM-
fabricated titanium implant-supported frameworks
have proven to be more accurate than cast implant-
supported frameworks.13–21 One interesting finding,
for which there is not yet a viable explanation, is the
statistically significant differences in the results
between the right and left one-screw tests. It was
noted that all of the authors were right handed; hand-
edness may have played a role in waxing and spruing
the wax patterns for the castings, but this would not
be an explanation in the virtual world of CAD/CAM
technology. It is important to note that there were no
differences among the university centers, which lends
credibility to the consistency of the protocol.

The International Journal of Oral & Maxillofacial Implants 927

© 2010 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 FORM WITHOUT WRITTEN PERMISSION FROM THE PUBLISHER.
 Drago et al
It should be noted that the CAD/CAM protocol
described in this study is different than the CNC-
milled protocol described by others. 14,17–21 CNC-
milled protocols differ significantly from the protocol
used in the present study in that CNC-milled proto-
cols generally involve a copy milling technique
whereby dental laboratory technicians construct cus-
tom wax or resin patterns on master casts. The pat-
terns are then scanned and the implant-supported
frameworks are milled with a CAD/CAM protocol. One
major disadvantage with the copy milling protocol
described here is that, depending on the size and
complexity of the implant frameworks being fabri-
cated, considerable time and labor are required to
fabricate the patterns. The current authors cited CNC
protocols in this paper because of their prevalence in
the dental literature. However, the CAD/CAM implant-
supported frameworks milled in this study were
designed virtually; they were not copy milled. All of
the frameworks in this study were designed by one
dental laboratory technician using a computer soft-
ware program on a computer workstation. The
CAD/CAM and cast frameworks were made as similar
as possible to each other. The authors designed the
clinical protocol based on clinical techniques that are
generally accepted by the prosthodontic community.
There were several limitations to this study. The
cast implant-supported frameworks were fabricated
by prosthodontic graduate students, each of whom
possessed different skills regarding waxing, casting,
and finishing cast frameworks. The protocol and
materials were controlled as much as possible to
standardize the process. Sectioning and soldering the
cast frameworks were not part of the protocol. Doing
so may have improved the fit of the frameworks to
the implants, but it would have introduced additional
variables regarding the consistency and strength of
the rigid connectors in the frameworks. This report
does not deny the advantages associated with sec-
tioning and soldering relative to improved accuracy
in implant frameworks, but the authors felt that this
would introduce variables that would detract from
the main focus of the study, which was determining
the accuracy of casting versus CAD/CAM technology
for implant frameworks. It should also be noted that
the frameworks were fitted onto the actual “clinical”
implants in the acrylic resin models. The authors
believe that this eliminated any potential limitations
normally associated with laboratory studies by evalu-
ating the fit of the frameworks directly on the
implants and not on implant analogs in master casts.
The biologic effects, if any, of implant frameworks that
featured a less-than-optimal fit were not evaluated.
928 Volume 25, Number 5, 2010
© 2010 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 FORM WITHOUT WRITTEN PERMISSION FROM THE PUBLISHER.
CONCLUSIONS
In this laboratory study, the computer-aided
design/computer-assisted machining (CAD/CAM) tech-
nologies featured here resulted in implant-supported
frameworks that were significantly more accurate than
cast implant-supported frameworks fabricated with
conventional casting technology and the lost-wax tech-
nique. The tactile probe system used in this study was
able to accurately identify and record the implant
restorative platforms of the models and frameworks.
The software program was able to accurately interpret
the digital data relative to the volumetric differences
between implants (clinical models) and implant restora-
tive platforms of the cast and CAD/CAM implant-sup-
ported frameworks. Additional laboratory and clinical
research is warranted to evaluate the degree of fit
obtained with other implant systems and other
CAD/CAM technologies, as well as the biologic implica-
tions of implant frameworks with a less-than-optimal fit.
ACKNOWLEDGMENTS
The authors gratefully acknowledge Tom Peterson, CDT, MDT, Presi-
dent, North Shore Laboratories, Lynn, MA, for his technical expertise
in casting and finishing the gold alloy casting in the pilot study; Dan
Rogers and Carlos Santamaria, Product Development, Biomet 3i, for
the imaging and computerized measurements; Heather J. Conrad,
DMD, MS, Assistant Professor, and Randy Harrison CDT, Division of
Prosthodontics, Department of Restorative Sciences, University of
Minnesota School of Dentistry; Barry M. Goldman, DDS, MS, Profes-
sor and Director, Postgraduate Prosthodontics and Sharon Siegel,
DDS, MS, Professor and Chair, Department of Prosthodontics, Nova
Southeastern University College of Dental Medicine; and Gerald J.
Ziebert, DDS, MS, Professor and Director of Graduate Prosthodon-
tics, Marquette University School of Dentistry.
REFERENCES
1. Binon P, Sutter F, Beaty K, et al. The role of screws in implant
systems. Int J Oral Maxillofac Implants 1994;9(suppl):48–63.
2. Jemt T, Lekholm U. Measurements of bone and framework
deformations induced by misfit of implant superstructures:
A pilot study in rabbits. Clin Oral Implants Res 1998;9:272–280.
3. Guichet D, Caputo A, Choi H, et al. Passivity of fit and marginal
opening in screw- or cement-retained implant fixed partial
denture designs. Int J Oral Maxillofac Implant 2000;15:239–246.
4. Tan K, Rubenstein J, Nicholls J, Yuodelis R. Three-dimensional
analysis of the casting accuracy of one-piece, osseointegrated
implant-retained prostheses. Int J Prosthodont 1993;6:346–363.
5. Zarb G, Schmitt A. The longitudinal clinical effectiveness of
osseointegrated dental implants: The Toronto study. Part III:
Problems and complications encountered. J Prosthet Dent
1990;64:185–194.
6. Jemt T. In vivo measurements of precision of fit involving
implant-supported prostheses in the edentulous jaw. Int J Oral
Maxillofac Implants 1996;11:151–158.

7. Jemt T, Book K. Prosthesis misfit and marginal bone loss in
edentulous implant patients. Int J Oral Maxillofac Implants
1996;11:620–625.
8. Michaels G, Carr A, Larsen P. Effect of prosthetic superstructure
accuracy on the osteointegrated implant bone interface. Oral
Surg Oral Med Oral Path Oral Radiol Endod 1997;83:198–205.
9. Jemt T, Lekholm U. Measurements of bone and framework
deformations induced by misfit of implant superstructures.
A pilot study in rabbits. Clin Oral Implants Res 1998;272–280.
10. Jemt T, Lekholm U, Johansson C. Bone response to implant-
supported frameworks with differing degrees of misfit preload:
In vivo study in rabbits. Clin Implant Dent Relat Res 2000;2:
129–137.
11. Karl M, Graef F, Heckmann S, Taylor T. A methodology to study
the effects of prosthesis misfit over time: An in vivo model. Int
J Oral Maxillofac Implants 2009;24:689–694.
12. Jemt T. Failures and complications in 391 consecutively
inserted fixed prostheses supported by Brånemark implants in
edentulous jaws. A study of treatment from the time of pros-
thesis placement to the first annual check-up. Int J Oral Max-
illofac Implants 1991;6:270–276.
13. Riedy S, Lang B, Lang B. Fit of implant frameworks fabricated
by different techniques. J Prosthet Dent 1997;78:596–604.
14. Jemt T, Back T, Petersson A. Precision of CNC-milled titanium
frameworks for implant treatment in the edentulous jaw. Int J
Prosthodont 1999;12:209–215.
15. Riedy SJ, Lang BR, Lang BE. Fit of implant frameworks fabricated
by different techniques. J Prosthet Dent 1997;78:596–604.
16. Ortorp A, Linden B, Jemt T. Clinical experiences of laser-welded
titanium frameworks supported by implants in the edentulous
mandible. A 5-year report. Int J Prosthodont 1999;12:65–72.
© 2010 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 FORM WITHOUT WRITTEN PERMISSION FROM THE PUBLISHER.
View publication stats
Drago et al
17. Bergendal B, Palmqvist S. Laser-welded titanium framework
for implant-supported fixed prostheses: A 5-year report. Int J
Oral Maxillofac Implants 1999;14:69–71.
18. Ortorp A, Jemt T. CNC-milled titanium frameworks supported
by implants in the edentulous jaw: A 10-year comparative clini-
cal study. Clin Implant Dent Relat Res 2009 Aug 17 [epub ahead
of print].
19. Al-Fadda SA, Zarb GA, Finer Y. A comparison of the accuracy of
fit of 2 methods for fabricating implant-prosthodontic frame-
works. Int J Prosthodont 2007;20:125–131.
20. Ortorp A, Jemt T. Clinical experience of CNC-milled titanium
frameworks supported by implants in the edentulous jaw:
A 3-year interim report. Clin Implant Dent Relat Res 2002;4:
104–109.
21. Ortorp A, Jemt T, Back T, Jalevik T. Comparisons of precision of fit
between cast and CNC-milled titanium implant frameworks for
the edentulous mandible. Int J Prosthodont 2003;16:194–200.
22. Vigolo P, Fonzi F, Majzoub Z, Cordioli G. An evaluation of
impression techniques for multiple internal connection
implant prostheses. J Prosthet Dent 2004;92:470–476.
23. Shehab AH, Pappas M, Burns DR, Douglas H, Moon PC. Com-
parative tensile strengths of preceramic and postceramic sol-
der connectors using high-palladium alloy. J Prosthet Dent
2005;93:148–152.
24. Barry C, Oetter T, Lane M, Mercier C. Keys to CAD/CAM in Small
Shipyards. Jersey City, NJ: SNAME, 1998.
25. Szalapaj P. 3D objects. In: Szalapaj P. CAD Principles for Archi-
tectural Design: Analytical Approaches to Computational Rep-
resentation of Architectural Form. London: Architectural Press,
2001.
The International Journal of Oral & Maxillofacial Implants 929