In Vitro Evaluation of Laser Induced Fluorescence for Aided Caries Excavation (FACE)

In Vitro Evaluation of Laser Induced

Fluorescence for Aided Caries Excavation

(FACE)

Utvärdering av in vitro laserinducerad

fluorescens som hjälpmedel vid

kariesexkavering (FACE)

Hedvig Olsson

Mattias Tagesson

Supervisor: Anders Hedenbjörk Lager, Dept. of Cariology

Master Thesis in Odontology (30 ECTS)

Malmö University

DDS Program

Faculty of Odontology

Abstract

Aim: to evaluate the reliability of fluorescence aided caries excavation (FACE), and if FACE

could be used as a supplement to existing clinical diagnostic methods for determination of the caries excavation endpoint.

Material and methods: two procedure protocols were developed. Protocol I with teeth that were

collected from a pool of extracted teeth and protocol II with freshly extracted and immediately frozen teeth. In total, 43 extracted teeth with dentin caries lesions underwent initial excavation adhering to consensus criteria on caries removal, followed by FACE. A light probe, D-light Pro, was used to illuminate the cavity, and if fluorescence was seen, further removal was indicated. Differences in excavated tooth substance were visualised by radiographic subtraction analysis utilising a developed reproducible radiograph imaging.

Results: Tooth samples were not fluorescing as expected which is why freshly extracted teeth

were used in protocol II. Only one tooth in each protocol exhibited fluorescence after initial excavation. The two teeth that underwent FACE did show difference in the subtraction analysis.

Conclusions:FACE is not reliable in determining the endpoint of excavation since

discolourations are highlighted, the interpretation of the fluorescence is not properly defined, and furthermore, FACE does not highlight all clinically apparent carious lesions. These shortcomings make FACE potentially harmful and further research is needed until FACE should be used in the clinical setting.

Sammanfattning

Syfte: att utvärdera tillförlitligheten hos laserinducerad fluorescens som hjälpmedel vid

kariesexkavering (FACE), och om FACE kan användas som ett komplement till befintliga kliniska diagnostiska metoder för att bestämma när kariesskadan är färdigexkaverad.

Material och metod: två experimentprotokoll utvecklades. Tänder till protokoll I valdes ut från

en samling extraherade tänder, och nyextraherade och direkt infrusna tänder ingick i protokoll II. Totalt 43 tänder med dentinkaries exkaverades först efter minimalinvasiv teknik enligt

konsensuskriterier för kariesexkavering, och sedan enligt FACE. Ljussonden som användes, D-light Pro, lyste upp kaviteten och vid synlig fluorescens indikerades fortsatt exkavering. Skillnader visualiserades med röntgenologisk subtraktionsanalys vilken möjliggjordes av för studien utvecklad reproducerbar röntgenteknik.

Resultat: Inkluderade tänder fluorescerade inte som förväntat, vilket är varför nyextraherade

tänder användes i protokoll II. Endast en tand från vardera protokoll fluoroscerade efter initiala exkaveringen. Skillnader syntes i subtraktionsanalysen hos de två tänderna som gick igenom FACE.

Slutsats: FACE är inte tillförlitligt för att bestämma slutpunkt för exkavering eftersom

missfärgningar markeras, tolkningen av fluorescensen är inte tydligt definierad, samt att FACE missar uppenbara karieslesioner. Dessa tillkortakommanden gör att FACE är potentiellt skadlig och vidare forskning behövs för att FACE ska kunna användas kliniskt.

Table of Contents

Abstract

Introduction

Aim 7

Hypotheses 7

Material and Methods

X-ray mount 8

Endpoint of excavation criteria 8

Protocol I 8

Protocol II

Fluorescence Aided Caries Excavation

9 10 Subtraction analysis 10

Results

Discussion

References

Introduction

Teeth consist of enamel, cementum, dentin and pulp. The enamel is formed by hydroxyapatite crystals to 96% by weight and is considered the hardest material in the human body (1). The hardness makes it brittle, and in need of support by the underlying dentin structure. Dentin consists of 70% mineral by weight, 20% protein and 10% water, and works as a cushion for the enamel. The dentin protein is arranged in a matrix consisting mainly of collagen type I. The mineral is organised along the collagen fibres, thus creating a flexible tissue. Tubuli traverse the dentin. The innermost structure of the tooth, the pulp, is a highly vascularised connective tissue with sensory nerves, progenitor cells and immune cells. Odontoblasts lining the pulp chamber continuously form dentin and have extensions into the dentinal tubuli.

Dental caries is the main disease of the hard tissues in the oral cavity. Bacteria adhere to enamel, cementum and exposed dentin. They metabolise carbohydrates from the diet and create an acidic environment through their by-products. These include acids, particularly lactic acid, which reduces pH in the plaque (2). An acidic environment around the enamel demineralises the enamel, resulting in calcium and phosphate ions diffusing away when the pH falls below the enamel’s critical level. In the initial stages of enamel demineralisation, the lesion is considered reversible. Remineralisation will occur during episodes of increased pH. The demineralisation of the enamel can be seen clinically as a “white spot lesion”. The breach of enamel and creation of a macroscopic cavity makes the lesion irreversible. When dentin is demineralised, the mineral content of the dentin is lost in a process similar to that of enamel but leaving the protein matrix exposed. In this process, proteinases are released, breaking down the protein matrix. Denaturation of the matrix gives the caries lesion its characteristic soft texture (3).

To restore the tooth to normal function and aesthetics, restorations are crucial. A preparation is necessary for retention of the filling. The edges of enamel and the soft demineralised dentin is removed, and a restoration is placed. Apart from functional and aesthetic reasons, restorations are placed to aid plaque control. Restorations with bad marginal integrity and defects can act as retention sites for bacteria and the tooth will be prone to secondary caries. Restorations are never permanent, and apart from secondary caries, a number of reasons affect longevity: fracture of the material or the restored tooth, retention loss and pulp complications. “The restorative cycle” is a concept to describe the inevitable failure of a restoration and the further loss of tooth substance with every replacement (4).

Dental caries can be viewed as a disease manifesting itself as carious lesions. It is a complex disease involving multiple factors for aetiology, prevention, treatment and risk of developing new cavities. Salivary composition, intake of fluoride, the type of bacteria in the oral microbiota, and type of diet all affect the risk of developing caries.

Brushing and use of fluoride toothpaste are of equal importance for caries control. Biological biofilm removal is needed to get full effect against caries (5). Dietary sugars have been proved to play a major part in incidence and progression of dental caries. The Vipeholm study by

Gustafsson et al. (1954) linked frequency of intake and type of food to caries (6). A high salivary flow rate clears the mouth and rinses residual food away. Saliva also acts as a barrier in defence of microorganisms. For example, it contains secretory IgA, agglutinin and mucins. These molecules bind microbes into a mass that can subsequently be swallowed and thus cleared from

the oral cavity. In addition, saliva buffers bacterial acids and remineralises the hard tissues by providing calcium and phosphate ions (7). The enamel is coated by the salivary pellicle, which is protecting the hard surface, but also providing receptors that enable bacterial adhesion (7). Fluoride prevents caries. The effect is dependent on its availability in saliva and biofilm fluid. It is effective as toothpaste (8,9), mouthrinse (10) or varnish (11). The fluoride ion binds into the enamel as a part of the demineralisation/remineralisation balance replacing a hydroxide group changing the hydroxyapatite to fluorhydroxyapatite (9,12):

Ca10(PO4)6(OH)2 +F-  Ca10(PO4)6OHF + OH

-This creates to an outer enamel layer that is more stable in lower pH, and thus making teeth more resistant to demineralisation (9). The reaction above is reversible and fluoride needs to be

administered continuously as part of a prophylactic regimen.

Bacteria that cause caries in the enamel are associated with a shift in the microbiota in the plaque to more acid producing and acid-tolerating species (13). The composition of bacteria in enamel caries differs from dentin caries. In dentin, the microbial community is diverse and contains many facultatively- and obligately anaerobic gram-positive bacteria. Examples are members of the genera Actinomyces, Lactobacillus, Bifidobacterium, and Streptococcus. Gram-negative bacteria are found in lower numbers, for example Prevotella, Porphyromonas and

Fusobacterium.

Bacteria in dentin caries produce an abundance of by-products. One group is called porphyrins (14). They are derivates of porphin and an intermediate compound for e.g. haemin. A unique property of porphyrins (mainly protoporphyrin IX in dental caries), is the ability to fluoresce when illuminated with a violet light (14,15). Fluorescence is a phenomenon where

electromagnetic radiation is absorbed by a material, the target molecule, in this case porphyrins, which undergoes excitation, and by de-excitation it emits photons of another wavelength than the initial radiation (16).

The concentration of porphyrins in healthy biological tissue, including unaffected dentin, is extremely low. Hard healthy dentin has no fluorescence in the red band. However, in the carious lesion, the level of porphyrins is elevated enough to be seen as an emitted distinct red light around 635 nm when illuminated in the region of 405 nm. Amounts of substance as low as 1 pmol of protoporphyrin IX are able to fluoresce. It is uncertain if porphyrins are the only group of molecules contributing to bacterial autofluorescence, or even the dominant group (15). Not all bacteria in carious lesions have the ability to autofluoresce red light, most notably the strongly caries-associated mutans streptococci and lactobacilli. These genera emit mostly green light instead (14,17). A few selected bacteria studied in an article by Lennon et al. (18) were found to be responsible for the dominant red fluorescence of carious lesions: Actinomyces israelii, A.

naeslundii and Prevotella intermedia. Conflicting results regarding the fluorescence of

Lactobacilli was found (17,18), and it was described more as an orange-red dominant colour. Their conclusion was that red fluorescence in a lesion was well suited for detection of bacteria causing dentinal caries.

Correlation between dentin hardness and autofluorescence was studied by Banerjee et al. (19). Histological preparations of dentin caries lesions were illuminated and compared for hardness with a Knoop indenter. It was concluded that fluorescence could be a useful marker for visualising soft dentin in need of excavation.

To discover dental caries and treatment needs the first approach is to examine the teeth

systematically, both visually and by gentle probing of predilection sites. Visual examination is atraumatic and is helped by cleaning and drying the surfaces first, since plaque usually covers initial caries or manifest lesions. Probing is done to see if there is perforated enamel where the probe gets stuck in carious dentin. Probing should be done carefully since iatrogenic breaching of initial lesions may occur (20). Examination reliability varies between site, diagnostic method, extent of cavity, breach of enamel etc. (21).

Bitewing radiographs are useful in caries diagnostics. The carious lesion with its demineralisation results in a radiolucent zone, since less photons are absorbed by the hard tissues. X-ray

examination is a way to examine mainly proximal, but also occlusal surfaces that are inaccessible to probing and visual examination. It is important that structures usually affected by caries are shown in the radiograph. Overlapping of the proximal contact point can hide possible lesions, which is why proper projection and direction of the x-ray beam is crucial in caries diagnostics (22).

Noise is the radiologic term for the random distribution of photons over an x-ray image (23). It is caused by variation in photons generated by the x-ray source, distribution over pixels in the sensor, as well as scattering and absorption by the irradiated tissue.

After diagnosis, the treatment for a caries lesion involves operative procedures when the loss of minerals has resulted in a macroscopic cavity. This is followed by a restoration. The procedure involves burs and hand excavators to remove infected dentin, with the goal to remove bacteria and demineralised tissue. Additional aims are to avoid pulpal exposure, prepare for a restoration that will withstand loading, and restore form, function and barrier (24).

Clinically, determining the endpoint of excavation relies on textural differences between probing sound and infected dentin. The hardness of the dentin varies over the lesion and is at its softest in the outermost layers of denatured dentin and high bacterial content, whereas it becomes harder towards the translucent zone (25). Traditionally, the cavity was excavated until complete hardness and until all dark stains were gone, which defined a caries free preparation, ready for restoration. However, the International Caries Consensus Collaboration published a paper with new guidelines regarding determination of the excavation endpoint. New recommendations propose different strategies depending on lesion proximity to the pulp. Radiographic lesions reaching the inner third or fourth of dentin should be approached carefully to avoid pulp lesions, leaving soft dentin over the most pulpal part of the cavity. When the lesion is shallower,

restoration quality should be a priority, and excavation to firm dentin is preferred (24).

Colouration and stains in the lesion have been proposed as an endpoint marker of excavation, but the variability between operators makes this strategy unreliable (26). Possible explanations for the brown discoloration are the maillard reaction of dentinal proteins, possibly in combination with extrinsic staining with chromophores from the diet (27). A study by Hibst et al. reaches the

conclusion that the same stains cause colour and fluorescence (15).

The perceived lack of methods to clinically determine the excavation endpoint has led to the development of caries detection dyes. During excavation these compounds are applied to the dentin with the goal of colouring the demineralised tissue, thus facilitating removal. However, these dyes seem to perform indiscriminately of caries activity and will colour all tissues with lowered degree of mineralisation. This frequently leads to over-excavation and consequently caries detection dyes are not presently advocated (28,29).

Another technique developed to help determine the excavation endpoint uses the fluorescent abilities of caries. Light probes aim to utilize the fluorescence by emitting light around 405 nm, which is exciting porphyrins, which are re-emitting red light of 655 nm. Light probes illuminate the cavity and the red fluorescence serves as a guide indicating infected dentin, and selective removal can then be performed. Absence of red patches is considered to be indicative of complete removal of carious dentin (15,17).

Examples of light probes are Facelight (W&H) and D-light Pro (GC). Facelight emits violet light with a peak frequency around 400-405 nm. This light has adverse effects on the naked eye, which is why protective goggles must be used (figure 1A). D-light Pro (figure 1B) emits violet light with the peak frequency at 405 nm (32). The instructions for use are the same as for Facelight.

Figure 1, A and B. High filter glasses and the D-Light Pro (GC product sheet (32)) Aim

The aim of this in vitro study is to evaluate the reliability of fluorescence aided caries excavation (FACE), and if it could be used as a supplement to existing clinical diagnostic methods for determining the endpoint of caries excavation in dentin.

Hypothesis

FACE based instruments are effective in highlighting residual caries during excavation, although not conforming to a minimally invasive endpoint of excavation.

A method was developed to generate reproducible radiographs with the aim to evaluate FACE, and compare it to traditional visual-tactile criteria.

X-ray mount

An x-ray mount was constructed after a carefully planned preliminary design (figure 2A). An acrylic glass (Plexiglas) board was used as a base, onto which a LEGO© _{plate was glued in the}

centre. At one end two identical LEGO© _{pieces were glued and on the other end an x-ray sensor}

mount made out of putty was fixated. Two LEGO© _{doors were trimmed to fit the x-ray tube. The}

door acted as a stop facing the centre to fixate the distance between x-ray tube, LEGO© _{plate and}

sensor. The distance between x-ray tube and sensor was 6 cm. The LEGO© _{plate in the middle}

enabled the sample-cups to be repeatedly placed in the exact same position after each removal. The sample-cups were made by removing the bottom of a plastic cup to fit a cubic LEGO© _brick

(figure 2B). The top was trimmed for optimum height. A tooth was fixated at the top of the cup together with the LEGO© _{brick at the bottom by gypsum (type III). This was done for each}

sample. The gypsum gave stability to the sample tooth, which allowed excavation and examination without altering position.

The design ensured near-to identical reproducible isometric x-ray images of tooth samples to enable a subtraction analysis.

Figure 2, A and B. The X-ray mount with a positioned A, and removed sample cup B. Excavation endpoint criteria

The endpoint of excavation used in this study followed the criteria described in the consensus report by Schwendicke et al. (24):

• Shallow or moderately deep cavitated dentinal lesions should be selectively removed to firm dentin. “Leathery” dentin pulpally and “scratchy” dentin peripherally.

• Deep cavitated dentinal lesions (reaching inner ¼ of dentin) should be selectively removed to soft dentin. “Soft” dentin pulpally and “scratchy” dentin peripherally.

Protocol I

Teeth with dentinal caries lesions were included. Teeth with pulp perforation were excluded. Samples were taken partly from a pool of extracted teeth from the department of Cariology (Faculty of Odontology, Malmö University, Sweden), and partly collected from the department of

General Dentistry (Faculty of Odontology, Malmö University, Sweden). Extracted teeth were kept in 50 % ethanol until collected and stored in a solution of Phosphate buffered saline (PBS) and 0,1 % Sodium Azide (NaN3) and kept refrigerated (8 °C) until protocol I was carried out.

Tooth samples were processed in the following order:

1. Teeth were rinsed in tap water and lightly dried, cemented with gypsum (type III) in plastic cups with a LEGO© _{brick in the bottom. Samples were fixated in the most}

favourable direction for radiographic projection, and direction towards x-ray source was marked. Each sample was assigned with an identification code number.

2. A baseline radiograph was taken and labelled with code number and “A”.

3. Each sample was photographed (Huawei mobile camera). The first photo was called “1”. 4. Samples were excavated according to visual-tactile criteria to endpoint of excavation. 5. A second radiograph was taken, labelled “B”.

6. Sample cavities were again photographed, labelled “2”, now through the high-filter glasses and illuminated by DLP.

7. The DLP device was used to indicate residual caries, and excavation was carried out until no more red areas were visible. If no red areas were visible, no further excavation was performed, nor were any radiographs or photographs taken.

8. Third radiograph was taken, labelled “C”.

9. Third photograph was taken, through high-filter glasses and illuminated by DLP and labelled “3”.

28 teeth were processed under protocol I. Operator 1 prepared 16 samples, operator 2 prepared 12 samples. The caries lesions consisted of 10 mesial or distal cavities, 9 occlusal, 4 mesio-occlusal, disto-occlusal or disto-incisal, 3 buccal, 1 disto-occluso-buccal, and 1 mesio-occluso-buccal cavities.

Protocol II

A second protocol was developed where freshly extracted teeth were used to evaluate differences in fluorescence between teeth that had been stored in a storage medium for various amounts of time before study. Samples were collected from the same departments as protocol I, with the addition of the department of Oral Surgery (Faculty of Odontology, Malmö University, Sweden) and the Public Dental Clinic in Rosengård (Malmö, Sweden). Extracted teeth were rinsed in saline solution and immediately stored dry in -8 °C until protocol II was carried out. Tooth samples in protocol II were processed in following order:

1. The teeth were taken out of storage, defrosted in room temperature, and fixated in gypsum.

2. Protocol I was then applied from point 2-9.

15 teeth were processed under protocol II. Operator 1 prepared 10 samples, operator 2 prepared 5 samples. The caries lesion consisted of 3 occlusal cavities, 5 mesio-occlusal or disto-occlusal, 5 mesial or distal, and 2 disto-occluso-buccal cavities.

The undermined enamel was removed by high-speed hand piece with diamond bur, followed by excavation of dentin by low speed hand piece with rose bur and hand excavator until visual-tactile criteria for excavation were met. Air/water syringe and examination probe were used during excavation. After visual-tactile excavation the FACE started. The operator illuminated the cavity with the light probe, and assessed it through high filter glasses. If the dentin fluoresced, further excavation was performed. This was repeated until no more fluorescence was seen.

Subtraction analysis

To analyse the relative difference between the B and C radiographs, a software program was developed to subtract two digital images and produce a superimposed image. A digital radiograph is made up from many pixels. When a radiograph is taken, every pixel is assigned a value

depending on how much x-ray energy that has hit the corresponding part of the digital sensor. Values range from 0-256 (33). The value of the pixel is translated to a shade of grey in a scale with 0 being white and 256 being black. Together the pixels make up a radiograph image. When superimposing the B and C radiographs, each pixel is compared between the two images. For example, the value of the top left pixel in radiograph B is compared to the value of the top left pixel in radiograph C. If little has changed in radiopacity between the radiographs, the values in the compared pixels are approximately the same, and when subtracting both values, the

absolute sum is close to 0, which results in a black or dark grey pixel in the superimposed image. If the two radiographs differ in radiopacity, the values of the compared pixels will differ. The more they differ when subtracted, the higher the resulting absolute sum will be, and the lighter the pixel in the superimposed image will be. The superimposed image will therefore highlight areas where radiograph B and C differ and be black in areas that are similar.

An ethical application was not deemed necessary since no personal data was obtained and teeth were anonymously donated to a pool of extracted teeth (23).

Results

Protocol I

In 8 teeth the pulp was perforated and thus excluded from protocol I.

1 tooth was fluorescing after the first excavation, which resulted in one radiograph C and photograph 3 (presented in figure 3, A and B).

Protocol II

In 9 teeth the pulp was perforated, or dislodged from gypsum, and thus excluded from protocol II. 1 tooth was fluorescing after the first excavation, which resulted in 1 radiograph C and 1

photograph 3.

Subtraction analysis

Stable projections were obtained using the x-ray mount and reproducibility reached an expected level. B and C radiographs differed a few pixels in projection and removed tooth substance was approximately 35 pixels wide.

Figure 3, A and B.

Discussion

The samples in protocol I were not checked for fluorescence until after the initial excavation. The reason for this was to prevent the operators from being influenced regarding the endpoint of excavation and possible extent of the cavity. However, suspicions arose during protocol I that very few teeth exhibited fluorescence at all. Efforts were made in protocol II to control storage factors thought to influence the fluorescence of samples.

In protocol II, operator 2 checked a few samples for fluorescence before initial excavation was completed, contrary to the order described in protocol I. However, operator 1 was kept blinded to if the teeth were fluorescing at all and if so, the extent of fluorescence in the cavity. These

additional findings indicated that a major proportion of the samples in protocol II did in fact not exhibit any fluorescent properties before the first excavation, despite clinically apparent carious lesions. In retrospect, it would have been interesting to study the total amount of samples that

Underwent initial excavation: 28 Radiograph A Examined under light probe: 20 Radiograph B Required further excavation: 1 Radiograph C Not indicated for further excavation: 19 Excluded from trial (pulpal involvement): 8 Protocol III Underwent initial excavation: 15 Radiograph A Examined under light probe: 6 Radiograph B Required further excavation: 1 Radiograph C Not indicated for further excavation: 5 Excluded from trial (pulpal involvement, dislodgement): 9 Protocol

exhibited fluorescence before initial excavation.

According to Lennon et al. (18) certain bacterial species are fluorescing in red, and those are said to be representative of dentin caries lesions. Nevertheless, despite obvious carious lesions, not all samples included in protocol I and II exhibited fluorescence in the present study. An explanation could be that red fluorescing bacterial species were not present in these lesions. Another

explanation could be a lack of porphyrins. Since one sample in each protocol did show fluorescence, it can be concluded that the instrument used was not faulty.

The samples in protocol I was extracted and stored in room temperature in PBS solution for an unknown amount of time, and later on refrigerated in 8 °C. When protocol I was executed only one tooth was fluorescing after the first excavation. Furthermore, not all the samples exhibited caries lesions with typical clinical texture. Probing revealed a rubbery texture, instead of the clinically known leathery texture and when excavated, the caries lesions detached from the cavity floor in one piece. Thus, the method of storage was suspected to compromise the fluorescing abilities of the caries lesions in the in vitro model, which led to the development of protocol II. One published study assessed different storage methods for extracted teeth and concluded that frozen samples do not change their fluorescence response over time (34). Protocol II was designed to control factors affecting storage. Teeth were freshly extracted and frozen immediately. The primary reason for this approach was to halt the bacterial and biological activity in the lesion and to enable dry storage, preventing diffusion of bacteria and by-products into the storage medium. Protocol II included fewer tooth samples due to a more extensive collection procedure, although the teeth were in better condition after defrosting than after PBS solution storage.

The series of studies by Lennon et al. on FACE do not address the question of cavities not exhibiting fluorescence (35-37). In their studies, different storage solutions were used (thymol 0.01%) than in this study (PBS and 0.1 % Sodium Azide). Teeth were also stored under different conditions (refrigerated versus refrigerated in protocol I and frozen in protocol II), as well as using an experimental setting with violet light produced by a xenon discharge lamp (Inspektor Research Systems Bv), with a peak excitation band between 370-420 nm, as compared to DLP, using 405 nm.

In the manufacturer’s instructions for use, both Facelight and D-light Pro, the fluorescence is described as “red” when it is indicating tissue removal (30-32). In the present study, bright red fluorescence was indicated for removal, while orange was not (figure 4, A and B). Teeth showing orange areas under illumination were almost always discolourations, and they were visible

without FACE. Another device utilising the autofluorescence of cariogenic bacteria is the DIAGNOdent Pen (KaVo). A study by Neves et al. found that discolouration in residual hard dentin led to higher measurements indicating caries. Since discolouration provides a poor guide for excavation (26), it was decided that dentin emitting orange fluorescence should not be excavated.

Since the excavation is dependent on the interpretation of the colour of fluorescence, inter operator differences do probably have clinical implications. Currently there are neither a description of what fluorescence looks like, nor any clinical examples in the instructions of use

that offers calibration. This probably leads to a highly individual interpretation affecting the amount of removed tooth substance, not even considering colour vision impairment in individuals.

Figure 4, A and B: Two samples illuminated with violet light, photographed through filter. Sample 9 (A) interpreted as bright red fluorescence. Sample 10 (B) interpreted as orange. It was clinically visible as discoloration. Further excavation was not indicated.

LEGO© _{bricks as a part of mounting teeth in in vitro studies have been used previously (38) and}

was suitable for mounting teeth in this study. The LEGO© _{bricks were well cemented in the}

gypsum and no failures were recorded. The x-ray tube and sensor were only placed one time per protocol and the risk for changes in beam direction was minimal. However, teeth were removed and placed repeatedly in the x-ray mount. Extreme care had to be taken in positioning the tooth sample to avoid miniscule changes in beam projection. If a distortion is present and the

radiograph depicts clearly demarcated lines, these structures will be highlighted in the

superimposed image. A few radiographs showed that placement of samples was inadequate, but that was easily corrected by placing the sample again in the mount and making sure the LEGO©

brick was placed all the way down. A new correctly centred radiograph could be taken and the previous one deleted.

The one sample from protocol II that exhibited red fluorescence was excavated under illumination. In the caries removal process the preparation extended into the pulp, visible on radiograph B, and C (figure 7 and 8). In a clinical setting, a pulp perforation aggravates the prognosis drastically and endodontic treatment would be required. This sample shows how FACE can differ from conventional caries removal, and a consequence of not conforming to a minimally invasive endpoint of excavation.

Figure 5, A-C. B and C radiograph, followed by their superimposed and subtracted image. Red fluorescence in the cavity guided the operator to remove tooth substance pulpally, which is visualised as a crescent in the subtracted image.

The subtraction analysis of the B and C radiographs was not a method to quantify the volume of removed tooth substance, but rather a way to visualise small changes in radiopacity (figure 5C). In the two samples with B and C radiographs, projections differ a few pixels. A slight rotation can also be seen in one sample. However, this error is relatively large in comparison to the changes in radiopacity that was studied, which are approximately 35 pixels in lesion depth. A reliable evaluation of the subtraction analysis and reproducibility of radiographs was not possible with only two samples. To fully assess this method, further investigations have to be performed. The noise in the x-ray image is disturbing the interpretation by the operator. Noise can be reduced by increasing the amount of photons (23). Therefore, it would have been beneficial to increase the exposure time drastically, to further enhance the quality of the superimposed image. It appears that noise is increasing in the superimposed images, possibly due to doubled noise input from the two radiographs. Increasing exposure time would have no adverse effects since no living subjects were used.

A common reason for exclusion of teeth from this study was caries lesions that extended into the pulp, which clinically would have required an endodontic treatment. This indicates that many of the sample teeth were not adequate for the study. Preferably the included cavities would be approximately the same kind of lesion, to facilitate comparison. This was difficult to obtain, and the cavities differed widely in extent, site and depth. The number of sample teeth was not

adequate to make a statistical analysis, but since the study showed a lack of fluorescence in most samples a conclusion about the effectiveness of FACE could still be made.

The samples were cemented in gypsum in order to fixate them. In protocol I this method worked well and provided a stable position during excavation and x-ray procedures. In contrast, in protocol II, a frequent mode of failure leading to exclusion was dislodgement of the tooth from

the gypsum. A possible explanation could be that teeth kept expelling moisture after defrosting, thus affecting the setting of the gypsum.

Clinical relevance

Products on the market are often hyped and described as a simple solution to a complex issue. The issue pertinent to this study is finding the objective endpoint of caries removal. FACE is an example of this. It proposes a simple solution that would have been excellent if it worked, which is why it needs scientific evaluation. This in vitro study compared the traditional method to determining the excavation endpoint with FACE. The outcome measure was radiographic differences between the two methods. Previous in vitro studies (35-37) have utilised histologic analysis and presence of bacteria as outcome measures. Zhang uses micro-CT and calculating volume changes as outcome measure (39). In the clinical setting, the clinician usually only has radiographs available, which is why radiographic changes were used as an outcome measure in this study.

Further research

Looking forward, an in vivo study is requested to determine if all teeth show fluorescence, thus removing storage factors that might interfere with the analysis. Care should be taken when using FACE in in vivo studies, since it would be unethical to use a suspected harmful diagnostic device. The question arises if it is appropriate to use FACE to determine the excavation endpoint. Lennon

et al. (35-37) are not adhering to endpoint criteria that was defined in the consensus report (24),

and approaches the problem from a classic viewpoint with total absence of bacteria as a definition.

Another question is if the assessment of fluorescence is showing inter-operator differences. The instructions of the manufacturer are to excavate red fluorescence. This study interpreted these instructions as bright red, but other studies are using “orange-red”. Our suggestion is that further research should be focused on providing guidelines assisting in evaluating the fluorescence itself. Such guidelines should also be based on minimal caries removal and the current consensus on endpoint of excavation.

It could be advocated that use of products based on FACE should be discontinued until these research questions are answered.

Conclusion

Our conclusion from this in-vitro study is that FACE not effective or reliable in determining the endpoint of excavation since our study reports that discolourations are highlighted, the

interpretation of the fluorescence is not properly defined, and furthermore, FACE does not highlight all clinically apparent carious lesions.

These shortcomings make FACE potentially harmful and further research is needed until FACE should be used in the clinical setting.

Acknowledgements

The authors want to thank the department of General Dentistry and Oral Surgery, Malmö University, as well as the Public Dental Clinic in Rosengård, Malmö for assisting in sample collection.

Special thanks to Dr. Anders Hoszek at the Department of Cariology, Malmö University, for valuable expertise and developing the custom image subtraction software.

References

(1) Nanci A. Ten Cate's Oral Histology. 8th red. Canada: Elsevier Health Sciences; 2014. (2) Struzycka I. The oral microbiome in dental caries. Pol.J.Microbiol. 2014;63(2):127-135. (3) Fejerskov O., Nyvad B., Kidd E. Dental Caries: The Disease and Its Clinical Management. 3rd red. Oxford: Wiley Blackwell; 2015.

(4) Brantley C. F., Bader J. D., Shugars D. A., Nesbit S. P. Does the cycle of rerestoration lead to larger restorations? J.Am.Dent.Assoc. 1995 Oct;126(10):1407-1413.

(5) Dijkman A., Huizinga E., Ruben J., Arends J. Remineralization of human enamel in situ after 3 months: the effect of not brushing versus the effect of an F dentifrice and an F-free dentifrice. Caries Res. 1990;24(4):263-266.

(6) GUSTAFSSON B. E., QUENSEL C. E., LANKE L. S., LUNDQVIST C., GRAHNEN H., BONOW B. E., et al. The Vipeholm dental caries study; the effect of different levels of carbohydrate intake on caries activity in 436 individuals observed for five years. Acta Odontol.Scand. 1954 Sep;11(3-4):232-264.

(7) Meyle J., Dommisch H., Groeger S., Giacaman R. A., Costalonga M., Herzberg M. The innate host response in caries and periodontitis. J.Clin.Periodontol. 2017 Dec;44(12):1215-1225. (8) Twetman S., Axelsson S., Dahlgren H., Holm A. K., Kallestal C., Lagerlof F., et al. Caries-preventive effect of fluoride toothpaste: a systematic review. Acta Odontol.Scand. 2003 Dec;61(6):347-355.

(9) Tenuta L. M., Cury J. A. Laboratory and human studies to estimate anticaries efficacy of fluoride toothpastes. Monogr.Oral Sci. 2013;23:108-124.

(10) Marinho V. C., Chong L. Y., Worthington H. V., Walsh T. Fluoride mouthrinses for preventing dental caries in children and adolescents. Cochrane Database Syst.Rev. 2016 Jul 29;7:CD002284.

(11) Petersson L. G., Twetman S., Dahlgren H., Norlund A., Holm A. K., Nordenram G., et al. Professional fluoride varnish treatment for caries control: a systematic review of clinical trials. Acta Odontol.Scand. 2004 Jun;62(3):170-176.

(12) Buzalaf M. A., Pessan J. P., Honorio H. M., ten Cate J. M. Mechanisms of action of fluoride for caries control. Monogr.Oral Sci. 2011;22:97-114.

(13) Takahashi N., Nyvad B. The role of bacteria in the caries process: ecological perspectives. J.Dent.Res. 2011 Mar;90(3):294-303.

(14) Konig K., Flemming G., Hibst R. Laser-induced autofluorescence spectroscopy of dental caries. Cell.Mol.Biol.(Noisy-le-grand) 1998 Dec;44(8):1293-1300.

(15) Hibst R., Paulus R., Lussi A. Detection of Occlusal Caries by Laser Fluorescence: Basic and Clinical Investigations. Medical Laser Application 2001 2001;16(3):205-213.

(16) The Editors of Encyclopaedia Britannica. 2017; Available:

https://www.britannica.com/science/fluorescence. Read: 04/03, 2019.

(17) Koenig K., Schneckenburger H. Laser-induced autofluorescence for medical diagnosis. J.Fluoresc. 1994 Mar;4(1):17-40.

(18) Lennon A. M., Buchalla W., Brune L., Zimmermann O., Gross U., Attin T. The ability of selected oral microorganisms to emit red fluorescence. Caries Res. 2006;40(1):2-5.

(19) Banerjee A., Cook R., Kellow S., Shah K., Festy F., Sherriff M., et al. A confocal micro-endoscopic investigation of the relationship between the microhardness of carious dentine and its autofluorescence. Eur.J.Oral Sci. 2010 Feb;118(1):75-79.

(20) Kuhnisch J., Dietz W., Stosser L., Hickel R., Heinrich-Weltzien R. Effects of dental probing on occlusal surfaces--a scanning electron microscopy evaluation. Caries Res. 2007;41(1):43-48. (21) Bader J. D., Shugars D. A., Bonito A. J. A systematic review of the performance of methods for identifying carious lesions. J.Public Health Dent. 2002 Fall;62(4):201-213.

(22) White S., Pharoah M. Oral Radiology - Principles and Interpretation. 7th red. St Louis, Missouri: Elsevier Mosby; 2014.

(23) Huda Walter, Abrahams R. B. Radiographic Techniques, Contrast, and Noise in X-Ray Imaging. Am.J.Roentgenol. 2015 02/01; 2018/10;204(2):W126-W131.

(24) Schwendicke F., Frencken J. E., Bjorndal L., Maltz M., Manton D. J., Ricketts D., et al. Managing Carious Lesions: Consensus Recommendations on Carious Tissue Removal. Adv.Dent.Res. 2016 May;28(2):58-67.

(25) Banerjee A., Sherriff M., Kidd E. A., Watson T. F. A confocal microscopic study relating the autofluorescence of carious dentine to its microhardness. Br.Dent.J. 1999 Aug 28;187(4):206-210.

(26) Kidd E. A., Ricketts D. N., Beighton D. Criteria for caries removal at the enamel-dentine junction: a clinical and microbiological study. Br.Dent.J. 1996 Apr 20;180(8):287-291. (27) Kleter G. A. Discoloration of dental carious lesions (a review). Arch.Oral Biol. 1998 Aug;43(8):629-632.

(28) Kidd E. A., Joyston-Bechal S., Beighton D. The use of a caries detector dye during cavity preparation: a microbiological assessment. Br.Dent.J. 1993 Apr 10;174(7):245-248.

(29) Yip H. K., Stevenson A. G., Beeley J. A. The specificity of caries detector dyes in cavity preparation. Br.Dent.J. 1994 Jun 11;176(11):417-421.

(30) W&H. Facelight, Technical data. 2018; Available:

https://www.wh.com/en_global/dental-products/restoration-prosthetics/caries-detection-devices/facelight/#facelight. Read 10/11, 2018.

(31) W&H. Reports & studies, Facelight part 2. 2018; Available:

https://www.wh.com/en_global/dental-newsroom/reportsandstudies/new-article/00492/. Read

10/15, 2018.

(32) GC. D-Light® Pro from GC Dual wavelength LED curing light. 2016; Available:

https://www.gceurope.com/products/dlightpro/. Read 10/17, 2018.

(33) Kortesniemi M., Ekholm M., Kauppinen T. The Handling of Digital Images. Tandläkartidningen 2014;106(2):82.

(34) Francescut P., Zimmerli B., Lussi A. Influence of different storage methods on laser fluorescence values: a two-year study. Caries Res. 2006;40(3):181-185.

(35) Lennon A. M. Fluorescence-aided caries excavation (FACE) compared to conventional method. Oper.Dent. 2003 Jul-Aug;28(4):341-345.

(36) Lennon A. M., Attin T., Martens S., Buchalla W. Fluorescence-aided caries excavation (FACE), caries detector, and conventional caries excavation in primary teeth. Pediatr.Dent. 2009 Jul-Aug;31(4):316-319.

(37) Lennon A. M., Buchalla W., Rassner B., Becker K., Attin T. Efficiency of 4 caries excavation methods compared. Oper.Dent. 2006 Sep-Oct;31(5):551-555.

(38) Abogazalah N., Eckert G. J., Ando M. In vitro performance of near infrared light

transillumination at 780-nm and digital radiography for detection of non-cavitated approximal caries. J.Dent. 2017 Aug;63:44-50.

(39) Zhang X., Tu R., Yin W., Zhou X., Li X., Hu D. Micro-computerized tomography

assessment of fluorescence aided caries excavation (FACE) technology: comparison with three other caries removal techniques. Aust.Dent.J. 2013 Dec;58(4):461-467.