Esthetic Dentistry Stain removal efficacy: An in vitro evaluation using quantitative image analysis Daranee Tanibiro|n, DDS. MS^^/Ching-Chang Ko, DDS. PhD**/ William H. Dotiglas. BDS, MSc. PhD*** Objective: This .study developed a computer image analysis technique as a quantitative means to measure changes in dental slain after brushing with various dentifrices. Method and materials: Enamel specimens with naturally occurring mature stain were cut from bovine incisors. The specimens were subjected to in vitro toothbrushing with one of the four tested groups, consisting oftn'o dentifrices that make claims of stain removal (Aquafresh Whitening and Rembrandt Sensitive), a regular dentifrice. (Aquafresh Triple Protection), and water. Digital images of stain specimens were recorded under standardized lighting conditions and analyzed with an image analysis software. The area-intensity stain deter- minant, which accounted for the reflected inten.sity and the con-e.sponding areas of.itain, was computed. Stain removal efficacy was calculated ba.sed on the difference in area-intensity stain determinant before and afler brushing. Results: Brushing with any of the tested dentifrices removed more stain than did brushing with wcuer alone. The finding that brushing with a regular dentifrice resulted in higher stain removal efficacy than brushing with water seems to indicate a role for abrasivity. Aquafresh Whitening had a higher .stain removal eßicacy than did Rembrandt for the removal of mature calcified stain u.\ed in this study. However, there were certain stains that none of the dentifrices removed. Conclusion: Coniputer image analyst.': provides an objective and c/iiaiiiitaiive mea.sureinent to di.\imguish in vitm .stain removal eß'icacy ofdentijrice.s. (Quintessence Inl \^)')^:2'):2i^-?il) Key words: arca-intcnsity stum deicrminant, tomputer imugc analysis lechnique, dental stain, dentifrice. i,Ljin remiivai efficacy Clinical relevance The computer image analysis technique is a poten- tially powerful evaluating tool for both laboratory and clinical investigations on the stain removal effi- cacy of dentifrices. All dentifrices tested could remove some mature stains, and one dentifrice tar- geted toward stain removal was more effective than the other. iil Or:il .Sticiitx'. Sdwil iil" ipolis, MÍIIHL'M.)1ÍI. • GnidiiLilo SUidoni, Di^pj ünrvi;r.',ii)' iil' MinnesiiU *"• Assislanl Professur, Dtpjrlmi^nl (iiOr:il .SL-ÍOIUC. SCIIW] UI" Denlislry. University of Minnesolii. MiniuMpoli.s. MinHL'MiUi. •""'•Prcirtssiir, DcpLirlmcnl ol Or;il SciüniíL-. .Sfii.jol ..f DL-iUiMry. University iil" MiniiL-siilii. Mirnrapolis. MiiiiifMHii. Rtprinl requïsls: Dr Djriintc Tiitiltiirojn, Dc|iiirlnii;nt i>l' Or:il .SLÍi]ni.o. Scliool or Denlislry. tJniversily ol" Minni:sol;i. I(.-2I5 Moos Tower. SIS Di:iüw:irL-SlrLVlSE. Minne:ipiiiis, Minn^soki SS4SS. Rix: iil2 l>26 I4M4. E xtrinsic dental siains are pigtnented deposits, firmly attached to the tooth surface and essentially external to the enamel, that may concern patients, primarily for esthetic rea.sons.' The association between extrinsic slain and other tooth surface deposits, such as dental plaque and calculus, has been shown.--* Other forms of dental stain are more closely associated with enamel and may even be within the enamel. Stain removal is clearly important, not only for cosmetic reasons but also for prevention of disease and promotion of oral health in general. Professional prophylaxis can retnove [he more obvi- ous extrinsic stains, and good oral hygiene can prevent the recurrence."' Individuals who do not use abrasive dentifrices have been observed to develop pigmented pellicles within 2 weeks.'' However, eontroversy exists as to whether the abrasivity is directly related to stain removal ability.'"" High abrasivity does not assure better cleansing properties but ratber causes hard tissue damage.'" Recently, .several dentitrices containing active iriTe- dients thai claim a special property in stain removal have 28 Volume 29, Number i, 199g
Transcript
Esthetic Dentistry
Stain removal efficacy: An in vitro evaluation using quantitativeimage analysisDaranee Tanibiro|n, DDS. MS^^/Ching-Chang Ko, DDS. PhD**/William H. Dotiglas. BDS, MSc. PhD***
Objective: This .study developed a computer image analysis technique as a quantitative means to measurechanges in dental slain after brushing with various dentifrices.Method and materials: Enamel specimens with naturally occurring mature stain were cut from bovineincisors. The specimens were subjected to in vitro toothbrushing with one of the four tested groups, consistingoftn'o dentifrices that make claims of stain removal (Aquafresh Whitening and Rembrandt Sensitive), a regulardentifrice. (Aquafresh Triple Protection), and water. Digital images of stain specimens were recorded understandardized lighting conditions and analyzed with an image analysis software. The area-intensity stain deter-minant, which accounted for the reflected inten.sity and the con-e.sponding areas of.itain, was computed. Stainremoval efficacy was calculated ba.sed on the difference in area-intensity stain determinant before and aflerbrushing.Results: Brushing with any of the tested dentifrices removed more stain than did brushing with wcuer alone.The finding that brushing with a regular dentifrice resulted in higher stain removal efficacy than brushing withwater seems to indicate a role for abrasivity. Aquafresh Whitening had a higher .stain removal eßicacy than didRembrandt for the removal of mature calcified stain u.\ed in this study. However, there were certain stains thatnone of the dentifrices removed.Conclusion: Coniputer image analyst.': provides an objective and c/iiaiiiitaiive mea.sureinent to di.\imguish invitm .stain removal eß'icacy ofdentijrice.s. (Quintessence Inl \^)')^:2'):2i^-?il)
The computer image analysis technique is a poten-tially powerful evaluating tool for both laboratoryand clinical investigations on the stain removal effi-cacy of dentifrices. All dentifrices tested couldremove some mature stains, and one dentifrice tar-geted toward stain removal was more effective thanthe other.
R t p r i n l r equ ï s l s : Dr Djriintc Tiitiltiirojn, Dc|iiirlnii;nt i>l' Or:il .SLÍi]ni.o.Scliool or Denlislry. tJniversily ol" Minni:sol;i. I(.-2I5 Moos Tower. SISDi:iüw:irL-SlrLVlSE. Minne:ipiiiis, Minn^soki SS4SS. Rix: iil2 l>26 I4M4.
Extrinsic dental siains are pigtnented deposits, firmlyattached to the tooth surface and essentially external
to the enamel, that may concern patients, primarily foresthetic rea.sons.' The association between extrinsicslain and other tooth surface deposits, such as dentalplaque and calculus, has been shown.--* Other forms ofdental stain are more closely associated with enameland may even be within the enamel. Stain removal isclearly important, not only for cosmetic reasons but alsofor prevention of disease and promotion of oral healthin general.
Professional prophylaxis can retnove [he more obvi-ous extrinsic stains, and good oral hygiene can preventthe recurrence."' Individuals who do not use abrasivedentifrices have been observed to develop pigmentedpellicles within 2 weeks.'' However, eontroversy existsas to whether the abrasivity is directly related tostain removal ability.'"" High abrasivity does not assurebetter cleansing properties but ratber causes hard tissuedamage.'"
Recently, .several dentitrices containing active iriTe-dients thai claim a special property in stain removal have
28 Volume 29, Number i, 199g
Tantbirofn et al
Fig 1 Bovine incisors witti naturally occurring mature si
been introduced commercially.•*" '- The evaluation ofstain removal efficacy of these products will require adifterentiating means of measurement. An important stepto evaluate stain is tbe measurement of discoloration. Avariety of techniques, which can be generally dividedinto subjective and objective metbods, bave been devel-oped. Subjective metbods are pursued by visual gradingof Slain, according to a nonparametric ranking of thearea and/or the intensity of discoloration.''-^ ' ' A variationon the subjective method, wbicb still depends on chair-side observation, is tbe use of standard color chips todetermine tbe closest color matcb.
Objective methods follow an instrumental approachto color measurement. Addy et al'' '"• used a spec-trophotomecer to measure tbe optical density of stainformed on clear acrylic resin blocks. Anotber instru-ment for measuring the color of stain, as well as thecolor of teeth, is a Cbromameter (Minolta), which hasbeen used in several studies.'-'^ Cotnputer-assistedimage analysis has been recently applied as a quantita-tive tool in the study of stain.'" "
The primary purpose of this study wa.s to develop aquantitative method to measure stain of any kind witha computerized image analysis technique. A secondarypurpose was to use tbis quantitative technique to eval-uate tbe efficacy of stain removal of different denti-frice formulations.
Method and materials
Tbe experimental design in tbís study was to comparethe appearance of stain by integrating tbe area of cov-erage and stain intensity. Measurements of stain inten-sity and area were performed before and afler brushingwitb different dentifrices.
The four testing groups were Aquafresh Whitening;AW, SmithKline Beecham) and Rembrandt Sensitive
(RB, Den-Mat) as dentifrices tbat claim stain removalabilities; Aquafresb Triple Protection, (AT, SmithKlineBeecham) as a control dentifrice witb an abrasivecomparable to tbat of AW, but witbout a specific ingre-dient to remove stain: and water, as a nonabrasive con-trol. All dentifrices were mixed with deionized waterto make 1:1 by weight slurry.
Specimen preparation
Bovine incisors with naturally occurring mature stainson the labial or proximal surfaces were selected (Fig1). Although variation in stain was observed, stainsselected for tbe present study had common features.The stain ranged from light brown to black, and thetexture appeared as either continuous witbin tbe cov-ered areas or as isolated speckles of stain. Stains usedin this study were "mature" stains, which were not .softlike "young" stains newly formed within a few days toweeks.
Most of the teetb bad stain areas large enougb toyield four comparable specimens, witb tbe exceptionof five teetb tbat provided only two specimens eacb.The specimens were approximately rectangular blocks,about 3 x 4 mm", of tooth enamel with attached stainin their natural anatomic location.
The stain surface was protected with Teflon tape(3M); then the specimen was mounted witb posteriorresin composite ¡P-50, 3M Dental) on a speciallydesigned ring. Tbe composite was polisbed witb Sof-Lex disks (3M Dental) to provide a leveled, flat sur-face surrounding the specimen. The Teflon tape wa,sremoved from the stain surface, and the specimen wascleaned with deionized water in an ultrasonic cleanerfor 5 minutes to remove any debris and attain a com-parable starting point for all specimens.
[Quintessence International 29
Tantbirojn étal
¡muge recording
The before image (ie, enamel with attached stain in thenatural anatomic location) was taken under standardizedconditions. The equipment setup was composed of astereomicroscope (Olympus SZHIO) with a charge-coupled device camera (Sony DXC-15IA) attached. Thespecimen was placed under the microscope and illumi-nated with the adjustable coaxial light from the leus.Specular reflectance was removed by using a Polaroidfilter (Olympus ILLC2). An electronic database repre-senting the image was captured, sent to an IBM-couipat-ible personal computer via a frame grabber (Imascan,Imagraph), and stored in a digital format. The imageanalysis software used was Óptimas 5.2 (Óptimas).
As mentioned earlier, the experiment was designed tocompare the before aud after images to measurechanges in area and intensity that represent the degreeof stain removal. For valid comparison, .several parame-ters were controlled, including magnification, aperture,and light source. The camera auto settings, .such asauto-white balance, autogain, and autoiris, were dis-abled. The setup was in such a way that the cameraelectronic shulter .speed function (iris) did not operate.In addition, the video output level was fixed so that the.system did not adjusi its iris and the output level accord-ing to the amount of the incoming liglit.
Under such conditions, the light standardization wasperformed for each image. To standardize liie lighiintcn.sity before a stain innige was taken, the imagebrightness of a perfectly diffusing white alumina tilewas conirolled by adjusting lighi (luipui lo meet a par-ticular luminance value. The light source was notaltered until each specimen image-capluring procedurewas iinishcd. This subsequently guaranteed a constanilight incideni on any recorded object. The full whiteimage would have the X-bit himinancc \alue of 255,
Tdoilibrushiiiii sinuiiuiion
The specimens were subjected to in vitro toothbrushingin an oral hygiene replicator (OHRl developed at iheMinnesota Dental Research for Biomaterials andBiomechanics." The OHR is based on a closed-loopservo-control and is able to deliver an isometric (load-control) force through ihc toothbrush handle with anisotonic (stroke-control) reciprocating stroke at a righiangle. An average pressing force of 2.5 N was appliedwith slraighl reciprocaling strokes, l-Hz frequency, andlO-nim stroke length. The toothbrushes used were Oral-B 40 (Oiiil-B Liiboralorics). Specimens from the sametooth were brushed with one of the lour treatmentgroups, AW, RB, AT, or waicr. After 3 minutes of brush-
ing, the image of the specimen was recorded under esame conditions as described in the previous section.
Qiiiiiililulivi' analysis
Two complementary parameters of stain, area and inten-sity, are itnportant for the evaluation process. Óptimasimage analysi.s software offers a function that is capableof separating pixels with a designated intensity from therest of the image and then computing the spatial area ofthe defined pixels. In this manner, the intensity of stainand the area of coverage can be combined inti) one com-prehensive stain determinant.
The first step in the analysis process was to convertthe color images into iheir gray-scale equivalent. Thenthree levels of intensity range were selected as represen-tative of mild, moderate, and severe stain, and this ismore fully explained in the next paragraph. Figure 2shows the original color image (A), the converted gray-scale image (B), and the gray-scale images with yellowpixels depicting three levels of stain intensity (C. D. andE). To further clarify the process, an image (F) wasadded lo show the area of tooth surface without stain,which was not included in the calculation. The softwarecomputed the spatial area from the nutnbers of pixels,which was then stored in a spreadsheet. Identical proce-dures were carried out for the before and after images.Therefore the total difference in spatial area of the threestain intensity levels defined the change in stain, ie.slain removal.
Most clinical stain indices assign arbitrary numbers,such as 1,2, and 3. to grade stain.''"'' Although relevan!to human visual perception, this grading is arbitrary. Inihe present study, loss reflectance factors have beenintroduced as an allcrniitive to numerical grading.According lo the equipment setup, reflectance (Rl wasmeasured as a ratio between the gray-scale intensity oftlie image and that of the full white object. The fullwhite image will have a reflectance of ! (equivalent to uluminance value of 255), whereas the totally blackimage possesses the value of (). Stains appear as darkareas because they absorb light, ie, some reflectance islost. Based on ihis assLimption. loss rellcelance factorswere dcluied as 1.00-R.
Three intensity ranges were selected as the best repre-senlatives of mild, moderate, and severe stain. The medi-an values of the reflectance of these intensity rangeswere calculated: then the loss relleciance factors (1.00 -Rl for mild, moderate, and severe stain were calculatedto be 0.53. Ü.7.\ and 0.91, respectively. Figure 3 showsthe concept of loss reflectance factors and its relation-ship to différent grading intervals along a gray-sculeintensily space. One assumption of this technique is ^\^.¿^
30 Volume 29, Number 1
Tantbiroin et al
Fig 2 Sleps m quantitative image analysis. (A) original color image: (B) same ¡mage afler gray-scale conversion (C),;D). and (E] same image when three intensity levels are selected—yellow pixels correspohd to the area of severe (C>,moderate ID), anrt Tiild (El sta'-r {Fl s^ne .mage with vellow pixels öenotirg the area of tooth si.prtace without stain
R12AW bW " -^ ^
. 'S*
B
Ouintessence International31
Tantbiro)n et al
Severe Moderatestain : slain
Background tootti surface. 355 Lumina
WtiileI Rellcctan
0 09 0 27 0 45 Median rellectances (R) ol three slaminlensity levels
0 91 0.73 0.55 Loss reflectance tactor^!.DO-R
Fig 3 Diagram ol a gray-scale intensity space showing the con-cept of loss reflectance factors. Reflectance ¡B] of the tull whiteimage = 1 (equivalent tc luminance value ol 255], and R of thetotally black image = 0. The median values of reflectances ofthree intensify ranges representing mild, moderate, and severestain are 0.45, 0.27, and 0 09, respectively The loss reflectancefactors ( 1.00 - R) are 0.55, 0.73, and 0,91, respectively. Note mereflectance of the background tooth sudace, which is not includedin the measurement
the backgrt)und tooth surface has reflective; intensitylevel beyond the intensity range of mild stain; thus thebackground tooth color (loss reflectance factor < 0.46) isexcluded from the measurement.
When the area of each specified stain inicn.sity levelwas multiplied by the corresponding loss reflectance fac-tor and added together, both the amount (area) and .severi-ty (intensity) of stain were expressed in a single value.This comprehensive value was called the arca-inleiisin'slain deienninani (ASD]. The change in ASD of theimages before and after hrushing reflected the amount ofstain removed. Suiin removal efficacy (SRE) was definedas the ratio of stain removed relative t:o the original stain.The formulas used to calculate ASD and SRE were:
ASD - (area of mild stain x 0.55) + (area of moderatestain X 0.73) +(area of severe stain xO.9l)
SRE - (ASD of before image) - (ASD of after image)ASD of before image
In short, SRE is the fraction of stain that wa.s re-moved; for example, an SRE value of 0.67 means that67% of the existing stain was removed. To assist in theintuitive interpretation of SRE, Fig 4 illustrates thebefore and after images of specimens with SRE valuesofO-Ol,0-27,0.55,0.71,and 0.91.
The SRE of the four testing groups of the experi-mental design was compared statistically with one-way analysis of variance followed by a Student-Neuman-Keuls post-hoc test (SuperANOVA, AbacusConcepts).
Results
One of the significant observations found after in vitri)loothbrushing was that the selected stain specimenswere either removable (n = 12) or nonremovable (n =6) by the dentifrices used. This finding strongly sug-gested the separation of the specimens into two popula-tion groups for statistical analysis.
Table I shows the SRE for six sets of specimenswith nonremovable stains. The SRE for these speci-mens was very low. and no statistical difference wasfound among the four testing groups iP > .05).
Table 2 shows SRE for 12 sets of specimens withremovable stains. Statistical analysis revealed that brush-ing with water had .significantly lower stain removal effi-cacy than did brushing wtth any of the tested dentifrices(P < .05). Among the dentifrices tested, the SRB of AWwas significantly higher than the SRE of RB and AT (P <,05). The SRE values of RB and AT were not significant-ly different (P > .05) from each other.
Discussion
Extrinsic dental stains, by definition, are confined to thedeposiis that adhere to the tooth surface and are essen-tially external to the enamel. However, extrinsic stainmay become more closely attached to the surface enam-el, producing a more generalized discoloration. At thispoint, the stain is probably an interphase between thetrue extrinsic stain and the modified enamel surface.Stain can become internalized via enamel defects, andthus can be found within the superficial tooth surface.
Mechanical cleaning techniques arc capable ofremoving extrinsic slain. However, such techniques arele.ss successful in removing stain in the interphaseregion or stain that is completely internalized.'" As isevident in this study, the stain specimens from six teethwere nonremovable after brushing with the tested denti-frices, although they had the appearance of extrinsicstain. This appears to be a salutary lesson not to assumea microstructure from chairside observations.
Some of these specimens were cross sectionedthrough the stain area and examined under light micro-scope for further information. Figure 5 shows cross sec-tions of some nonremovable stains, revealing the darkbrown stain within the superficial enamel. This findingsupports the separation of the specimens into two popu-lation groups, namely removable and nonremovablestain, which in general terms is a safer grouping thanextrinsic and intrinsic stain. However, even though thisstain had penetrated the superficial enamel, it i.s differen-tiated from other intrinsic stains that lie deeper within theenamel and dentin.
32 Volume 29, Number
Tantbiroin et al
Figs 4a to 4e Typicai images ofspecimens before (ieft) and after(iigfit) in vitro toothbrushing
Fig 4a SRE = 0 01
Fig 4b SRE = 0 27
Fig 4c SRE = 0 55
Fig4d SRE = 0 71
Fig4e SRE = 0.91 R12AW aft
33
Tantbirain et al
TABLE 1
Tooth
NIN2N3N4N5N6
Stain removal efficacy of six sets of specimens with nonremovable stains
AW
0.020.050.000.07
-o.ot0.05
Mean ± SD 0.03 * 0.03"
AW = Aquafresh Whilening; RB = Rembrandt Sensiliv•Means are rot significantly different (P > .05).
Stain removal efficacy
BB
0.04o.oe0.020 030.030.02
0.03
, AT = Agij
AT
0.000.050.010.07
0.02- 0,03 ±0.03-
afresh Triple Proleclion.
Water
0 01
0.000.000.00
0.00 ±0.00'
TABLE 2 Stain removal efficacy of 12 sets of specimens with removable stains
Fig 5 Cross-sectional specimens with nonremovable stainsuperficial enamel.
N6AW Xs Note (he dark brown stain withm the
34 Volume 29, Number 1.
Tant b I roi n et ai
The stain specimens used in this study were naturally
occurring mature slains from bovine teeth and in this
respect should be differentiated from the methodology
of Stookey et a l . " who developed a reproducible syn-
thetic stain on bovine teeth. Ongoing work using elec-
tron probe microanalysis in the laboratories of the pre-
sent authors revealed that the bovine dental stains used
have elemental compositions similar to human dental
stains (unpublished data, 1997), The compositional ele-
ments pre.seni in stains are calcium, phosphorus, carbon,
o\ygen, sulfur, iron, and copper. Calcium and phospho-
rus indicate the calcification: ie. the.se stains are not soft.
Calcif icat ion of stain has been reported in several
publications.--'-" The calcilled characteristic of naturally
occurring mature stain used in this study is unlike the
widely used in vitro model referred to earlier, which
created stain film on enamel specimens within a 4-day
period by using a mixture of muein, coffee, tea. and
chromogenic microorganisms.' ' The use of naturally
occurr ing mature stain is also expected to he more
challenging than most short-range clinical studies, in
which stains are chemical ly induced wi th in 2 to 8
w e e k s . " ' - " ' - ' Other studies have examined naturally
occurring stains"-- and proposed these to be the most
appropriate stains for evaluating the cleaning proper-
ties of dentifrices. This conclusion, supported by the
studies on electron probe microanalysis mentioned ear-
lier, determined the choice of the natural bovine stain
for study in the present report.
Removal of dental stain ha.s been widely accepted as
an important funct ion of a denti fr ice. ' Some studies
indicate that dentifrice abrasivity is directly related to
stain removal ability, but the necessity of high abrasive
content is controversial.'^" In the present study, a regular
dentifrice (AT) did demonstrate better stain removal
efficacy than brushing with water alone, as shown in
Table 2. The other parameters that might affect the stain
removal ability, such as brushing force, toothbrushes
used, and the level (height) of the specimens, were con-
trolled. This particular result suggests that the abrasives
in a dentifrice assist the stain removal ability.
However, chemical ingredients also play an impor-
tant role, especially in dentifrices formulated directly to
remove stain and not merely plaque. It was found that
AW had a significantly higher SRE than AT in this in
vitro measurement, although both dentifrices contain
similar abrasives. The result indicates that the active
ingredient in AW enhanced the removal of naturally
occurring mature stain.
Another interesting finding of the present in vi tro
study was that AW demonstrated the better abil ity io
remove natural bovine stain than RB. Both dentifrices
are formulated for improved stain removal, but the
active ingredients, and probably the abrasives, arc dif-
ferent. The fact that RB was .shown to reduce chlorhexi-
dine stain in clinical studies" -' but performed similar to
a regular dentifrice in the present study, is hypothesized
to be related to the mechanism of stain removal and
because mature stain is likely to be more bighly calci-
fied. One of the stain removal mechanisms exerted by
AW has been propo.sed to be the sequestration of calci-
um f rom stain complex (unpubl ished data. 1996).
whereas RB contains a high-molecular weight ingredi-
ent (papain) that is believed to react with the protein
constituent of extrinsic stain. Therefore. RB was effec-
tive in clinical studies, presumably because the sub-
strates were highly organic chlorhexidine stains that
were extrinsic and accessible. In the present study,
which uti l ized more mature calcif ied stains as sub-
strates, AW achieved a higher stain remo\al efficacy
than RB.
in this study, stain removal was assessed by the
whitening effect on the gray-scale equivalent of both
AW and RB. The stain removal mechanism of AW may
be mu l t i f ac to r i a l because its ingred ient , sodium
tripolyphosphate. is a powerful surfactant in addition
to being a calcium séquestrant.-^
Several methods have been developed, by various
authors, for the quantitative evaluation of dental stain.
The optical density of stain on an acrylic resin block
was measured with a spectrophotometer." '-* The model
was simple and accurately assessed the stain removal
potential of dentifrices. However, the optical density
had to be obtained through a elear substrate. Tbe
Chromameter is another instrument utilized to assess
the color of stain on tooth surface.'- ' ' The measurement
is quantitative and sensitive to changes in stain level,
but confined to a surface area of 3 mm in diameter.
Computer-assisted image analysis is a promising
approach to evaluating discoloration because of the
lower subjectivity and the higher discriminating power.
When the technique was first introduced, the optical
density histograms of the images were qualitatively
compared, eg, a shift of a histogram to the left indicat-
ed an increase in darkness." A later study calculated
stain intensity based on the gray levels of digital image
.scanning from an intraoral photograph.'" In the present
study, a digital image of stain obtained with a stereo-
microscope was analyzed with software that provides
nutneric information of the visualized object.
Tbis quantitative technique is able to comhine the
area and intensity of stain into a single parameter,
namely area-intensity stain determinant. In the widely
used Lobene index.^ ihe severity of stain (ie. intensity)
and the extent of stain (ie, area) are scored separately.
Alternatively, the color and area can be subjectively
Quintessence International 35
Tantbiroin ef ai
graded together.'* One study proposed to calculate stainas 1.5 X intensity + amount. The rationale for the 1.5 fac-tor was that dark stain has a higher visual impact onesthetics than light stain.--' Another study"' multiplied thearea of light, medium, and dark stain with I, 2, and 3.respectively, to correlate the results with the Lobeneindex." Although the scales are arbitrât^, these tech-niques arc in broad agreement with visual perception.
The loss reflectance factor used in this study was anadditional attempt to relate the results to present andfuture clinical studies based on chairside observations;ie. tbe intensity level that best diseriminated stain colorfrom tooth color was depicted and divided into tbteeequivalent intervals (see Fig 3), Then the same intensitylevels were used with every specimen in the entire study.Therefore the loss reflectance factors are similar to non-patatnetrie grading in terms of subjective visual percep-tion yet possess a more scientillc basis. The whole spec-trum of stain color was divided into three intervals tocorrespond to the traditional clinical observations ofmild, moderate, and severe stain. Of course, the sensitiv-ity of the test is expected to be bigher if mure levels ofthe interval are retained.
Computer image analysis bas .several advantages overthe traditional nonparametric grading of stain. The pro-cedure is less subjective and more quantitative. The sen-sitivily of quantitative image analysis is expected to behigher tban thai of the nonparametric grading technique,in which the scores are di.sctete rankings. Factors thaican affect the accuracy of image analysis, especiallywhen dealing wilh opiical density comparist)ns. arelighting condition and specimen orientation. Lightingconditions were controlled and adjusted during theprocess of image capture by using a white alumina tileas rei'erence. However, specimen orientation stilldepended on ihe examiner, wbich may also bave affectedtbe result.
In the present sludy. the effect of ihe backgroundtooth rctlectance on stain color was minimized in twoways. First, the background tooth substrate badreflectance in the range that was not included in the cal-culation of the stain removal efficacy (loss rcllectancefactor less than 0.46). Second, all experiments weredone on a comparative basis. Stain removal efficacywas calculated from ihe difference between ihe beforeand after images of the same specimen. The four exper-imental groups in each run were performed on speci-mens of comparable stains that were cui from tbe samebovine ttiotb.
Another variable in Ihe rclleclance t)f specimens fromdifferent areas of the tooth arises from the differentdirections and concentrations of enamel rods, A Polaroidlens was used to remove the specular reflectance.
Therefore, only the diffused rellectanee that affected theobject's color wa.s measured. The diffused reflectance isindependent of tbe angulation of incident light andobservation direction.
This study bas introduced a quantitative image analy-sis procedure for the measurement of dental stuin and theextent of its removal. The technique revealed ihe differ-ence in stain removal efficacy among the evaluated den-tifriees. it may be possible to extend tbis promisingmethodology by combining the itnage analysis with suchequipfiient as an intraoral camera. This would be a greatadvance on tbe current chairside observations used toassess stain removal. At tbis stage, laboratory proceduresare intended to support and refine clinical evaluationsand to supply vital information on the nature of stain andthe mechanisms of stain removal, which may requirethat the teeth be retrieved for in-deptb analysis.
Conclusion
The stain removal efficacy of dentifrices was calculatedfrom changes in the intensity and area of stain on tbespecimen after brushing in a format useful for clinicalcomparisons. Computer image analysis measured boththe intensity and area of stain from a digital image andcombined them into a single parameter. The techniqueis less subjective and more quantitative and reveals thedifferent performances of tbe dentifrices tested.
Abrasives in dentifrices assist stain removal, as indi-cated by the fact that a regular dentifrice (AquafreshTriple Protection) had better stain-removing ability thandid brushing with water aione. Tbe effect of cbemicalingredients in stain removal was shown by tbe highstain removal efficacy of Aquafresh Whitening for theparticular kind of slain tested.
Acknowledgments
This study WLIS suppiined in part by Mititiesiitii Dental Rc-;ciirctiCeincr W» Biiiiiiiiteiials ;iiid Bioiiiäi.-Ii;inii.s (MDRCBB) LiiidSmithKline BeccliLini. Dt T:mlliini|n is suppwled by LI stlioUiíshipIriiiii the AiiLL[i(l;ii™iinJoi I-II|]IUILUÍ()II, Ttuiiliind.
References
I CiMiiiii/;! FA. NewniLin MG. Ciinii\i! PL-iiodoiitolo^>. cd K.
Phii:idelpliia: S;iliiiders, iy96;15M.
2. Vilbel RI. Iniriiisic ;iiid eMiiiisii; üisciiloriiliiin oí Ihe deiiuiioH ij
liteniliiio review). J Oiul Med lyTSiiOi^y-KH.
.Í. Theilade J. P:\\Y¿ KfW. Sfiiiiniii!; election iiiii;msi;(ipy nf hiaciisliiiii on luimuii perniaiieiil teeth. SeLin MÍCIOM- | 1)^7 | .iyn3-i'iii<)
4. BiiliiiKi BW. Stiii/eiibeigcr OP. Vick V, Gins.'iniLiii E. Reduction111 L•;liclllLl̂ ;i[id PciideiL slain willi Tar I ai-co mm i Crest. J ClinDem i'W.S:(>:IH.'i-IK7,
36 Volume 29. Number 1
Tantbiroin et al
tiriksen HM, Nordtw H. E\tri[is¡c discolonition of ttelh. J ClinPoriodonti.l 1978;5:229-236.
f> Manly RS. A structureless recurrent deposit on tectb, J Dent Ri;s
Anierifíin Dental Associ;ition HLMÍIII Kiiundutjon RescarthInslitutc, Clink;ii methods for determining dentitritcí-cleaning•iMhy J Am Dent Assiic 1984:109:759-762,Lobene RR Efíect i¡r dentifrice,', on tooth stains with L-otitroiledhru^liiiiy. J Am DentAssoc I%8;77:849-B55.
Addv M.GoodtieldS.TheuseotutryliL-tDcompLLrLMlie Libi util-ity and stain remov;ii properties ottt1l«hpa^tes. Clin Miller 1991-7:219-225.
Sangnes G, Traumati/ation of teeth Lind jingivu related tu hubitj-^iltooth cleaning procedures. J Clin PeriDdiintol 197ñ;3: 94-10.1.Emling RC. Levin S. Shi X. Weinberg S. Yankell SL. Rembrandttoothpaste stain prevention with und without the use of Peridex, JClin Dent l992;3:.'i9-65.
Yankell SL. Emling RC, Prencipe M. Rustogi K. Voipe AR. Lab-oratory and clinical slain removal evaluations of twu tartar con-trol dentifriœ.s. J Clin Dent 199.S;6;207-JI0.Stookey GK. Burkhard TA, Scbemehorn BR. tn vitro removal ofstain with dentifrices. J Dent Res i 9K2;61:1236-1239..'\ddy M. Pra>itno S, Taylor L, Cadogan S. An in vitro study oftbe Rile of dietary factors in the aetiology of tooth staming asso-ciated wilb the use of chlorhexidine. J Pcriodunt Res 1979;14:4O3-H0.
Joiner A. Jones NM. Raven SJ. Investigation of factors influerjc-ing stain formation utilizing an in situ model. Adv Dent Res1995:9:471^76.Mtlnnes C. Johnson B, Emling RC. Yanketl SL. Clinical andcomputcr-a.ssisted evaluations of tbe stain removal ability of theSonicare electronic toothbriisb.J Clin Dent 1994:5:13-18,Addy M. Prayitno SW, Light microscopic and color televisionimage analysis of tbe development of staining on chlorhesidine-treated surfaces, J Periodontol 1980:51:39^.1.Anderson JC. Douglas WH. Pintado MR. DeLong R, In vitriireplication of clinical toothbrush ing [abstract 504], J Dent Res1994:73:164,
Addy M. Moran J. Mechanisms of stain fürmation on leeth. inparticular associated with metal ions and antiseptics, Adv DentRes 1995:9:450-456.
Reid JS, Beeley JA, MacDonald DC. Investigations into blackes;trin,',ic tootb stiiin, J Dent Res 1977:56:895-899,Emling RC. Shi X. Yankell SL. Rembrandt toothpaste: Stainremoval following tbe use of Peridex. J Clin Dent 1992:3:66-69.Lobeni: RR. Clinical studies of the cleaning functions ol" denti-frices, J Am Dent Assoc 1982:105:798-802,Grayson M. Kirk-Otbmer Encyclopedia of Chemical Tech-nology, ed 3. New York: Wiley-Interscience. 1978:22:396.Tilliss TSt. Stach DJ. Cross-Poline GN. Use of tootbpicks forcblorbe>:idine staining, J Clin Periodontol 1992:19:398^00.
Quintessence International
CAD/CIM in AestheticDentistry: CEREC10-Year AnniversarySymposiumEdited by W. H. Mormann
