Detection of intraoral lesions using a fluorescence camera

Michael Thoms

Dürr Dental GmbH & Co KG, Höpfigheimer Str. 17, D-74321 Bietigheim-Bissingen

University of Erlangen-Nürnberg, Institute of Material Science VI, Martensstr. 7, D-91058 Erlangen

ABSTRACT

Optical methods for the detection of carious lesions, calculus and plaque have the advantage of being minimally

invasive. The use of endogeneous fluorescence markers like porphyrins could simplify the application of fluorescence

techniques in the dental practice. It is known that porphyrins are produced by some of the bacterial species that are

present in the oral cavity. Since porphyrins have an excitation band at about 400nm they have the potential to be used as

fluorescent markers of locations in the oral cavity where the production of bacteria is out of the limits of healthy

regions. Further, modern and efficient GaN-based semiconductor diodes emit light in this spectral range and thus make

the implementation of fluorescence sensors with excitation at this wavelength easy.

Carious lesions, calculus and plaque have been measured using a self build fluorescence camera using GaN-diodes for

illumination at 405nm. Further, emission spectra under this excitation were recorded. For the latter purpose freshly

extracted teeth were used. It has been found that already in the case of an initial carious lesion red porphyrinfluorescence

is emitted whereas it is absent in healthy enamel. In already brown coloured carious lesions the emission

bands of porphyrin are present but the observed overall fluorescence intensity is lower, probably due to the absorption

of the fluorescence by the carious defect itself. In dental calculus, dental plaque and subgingival concrements porphyrin

originated luminescence was found as well. Since in these cases the emission spectra differ slightly it can be concluded

that they originate from different types of porphyrins and thus also from different bacteria. These results show that this

fluorescence technique can be a promising method to diagnose carious lesions, calculus and plaque.

Keyword List: fluorescence, porphyrin, bacteria, caries, calculus, plaque, camera, imaging, emission spectrum

1. INTRODUCTION

In dentistry non invasive methods that can diagnose early stages of diseases like caries or periodontitis are desirable.

Such diseases are started by the infection of the oral cavity with specific bacteria. The next step is that these bacteria

accumulate and grow on specific intraoral areas in the form of plaque. If the plaque is not removed secondary effects

like demineralization or bone loss can follow. There are methods like i.e. x-ray radiography that are only sensitive to

these secondary effects and not to the primary effect of the bacteriological equilibrium being out of order at specific

intraoral areas. However, methods that are instantaneously able to detect the bacteriological concentration and show up

sites where the concentrations are out of healthy limits would allow a much earlier diagnoses of the coming intraoral

diseases. With this information these sites could be treated and secondary effects caused by bacteria minimized. Some

methods like i.e. the strip mutans test allow already to get an information about the global average concentration of such

bacteria. However, this test gives no spatial resolution to identify specific areas with increased accumulation of these

bacteria. Other methods are the growth of bacterial cultures or DNA sequencation and amplification after the acquisition

of bacterial agglomerates from specific areas using a paper tip. Especially the latter methods has the advantage of being

able to identify the bacterial species. However, they can not be used to inspect routinely all intraoral areas. The

traditional method of optical inspection has the advantage of high spatial resolution but is i.e. lacking the sensitivity to

detect very early carious lesions, because it visualizes only the late secondary effect of demineralization. Further there is

the possibility to use staining solutions to visualize plaque with the slight disadvantage that the stains remain for some

time.

Obviously a method of choice should have the following characteristics. It should be instantaneous, work with sufficient

spatial resolution, give signals that correlate with the concentrations of pathogen key bacteria, allow an easy

documentation of the situation, should be sufficiently low in price and fast in use in order to be employed routinely.

In principal a camera based fluorescence method

would have the potential to fulfill the requirements

above if there would exist intrinsic fluorescent

molecules that are produced by the relevant bacteria.

Such molecules can be the porphyrins, which are

known to be produced at least by some bacteria. It

can be excited efficiently in the so called Soret band

at about 400nm and with lower efficiency in four Qbands

having their maxima between about 500 and

630nm. These porphyrin molecules fluoresce mainly

in the red and near IR. The molecular structure of

the porphyrin molecules is shown in Fig. 1.

Depending on the side groups of the various

porphyrin derivates the maxima of the excitation

and luminescence bands can be shifted. Further it is

known that the exact positions of the maxima are

shifted with the type of solvent. In the following

thus such a fluorescent camera based method to identify plaque, calculus, subgingival concrements and carious lesions

under the excitation in the Soret band of the porphyrins is investigated. Further, emission spectra under the excitation in

the Q-band at 405nm are reported for these sites and compared to the emission spectra of some bacteria in vitro under

the same excitation.

2. EXPERIMENTAL

For the studies an intraoral camera of the type Vistacam (Dürr Dental, Germany) has been modified by exchanging the

white LEDs of the camera by blue LEDs emitting at 405nm an optical power of 60mW. Further an optical long pass

filter has been introduced into the beam path in front of the CCD-sensor to cut down the excitation light below 495nm.

To acquire images by the camera software of the type DBSWIN (Dürr Dental, Germany) has been used. As a result of

the digitization of the video signal the images were composed of 720x576 pixel with 3x8bit intensities of the RGBchannels.

Some of the images where recorded with automatic white balance off (Fig. 2 and 5) the others with white

balance on in the video processing. In order to collect conventional intraoral images an unmodified intraoral camera of

type Vistacam (Dürr Dental, Germany) has been used.

The local emission spectra have been investigated using a spectrometer of the type HR4000 (Ocean Optics,

Netherlands) that has been coupled by a fiberoptical cable to the site. To cut down the excitation light below 495nm an

optical long pass filter has been introduced into the beam path in front of the spectrometer.

3. RESULTS

3.1. Caries lesions

Typical images of uncavitated caries defects taken with the intraoral camera and the fluorescence camera are shown in

Fig. 2. The beginning demineralization of the enamel can be observed in the intraoral image by surfaces resembling

unglazed china or chalk (arrow). It is well known that this demineralization is caused by plaque that lowered the pH for

some time under a critical value of about 5.5 and etched the enamel surface. The resulting surfaces areas are called

white spot lesions. They can be remineralized in order to convert them back to sound enamel. The fluorescence image of

the same intraoral region shows at the position of the white spot lesion red luminescence. As a consequence of the green

luminescence of sound enamel the “red” white spot lesion can be much easier observed in the fluorescence image than

in the normal intraoral image on the left. Therefore the dentist has a higher sensitivity to observe the white spot lesion in

the fluorescence image compared to the standard intraoral image. Hence a dentist could treat and remineralize the

affected sites earlier if he uses the fluorescence camera. Further it should be noted, that the fluorescence image shows

red luminescence also on the right side of the white spot of the ordinary intraoral image, where no white spot has yet

developed. One can speculate that the demineralization is not strong enough in this region to become visible in the

ordinary intraoral image.

It is interesting to investigate the fluorescence spectra of the enamel and the white spot lesion. They are shown in Fig. 7.

The fluorescence of the intact enamel decreases monotonically in intensity from the green to the red spectral region with

increasing wavelength. The sharp edge on the short wavelength side at about 500nm is caused by the cut off of the

Fig. 1: Molecular structure of porphyrins. R and R´ are side groups

that are specific for the specific type of porphyrin. They are i.e. for

Protoporphyrin R=─CH═CH2, R´=─ CH2─ CH2─COOH

Hematoporphyrin R=─CHOH─CH3, R´=─ CH2─ CH2─COOH

Coproporphyrin R= R´=─COOH

Uroporphyrin R= R´=─COOH and instead of ─CH3 groups also

─COOH groups

optical long pass filter. The fluorescence spectrum at the white spot lesion (Caries 1) shows a similar decrease of the

intensity from the green to the red spectral region superimposed by two dominant emission peaks at about 640 and

700nm. These peaks in the emission spectrum are very similar to the emission spectra of porphyrins. Therefore it is very

probable that the fluorophore that is excited at 405nm is a porphyrin that has been produced by bacteria and

accumulated at the tooth surface. It is interesting to investigate, whether all white spot lesions have the same emission

spectrum. Therefore a caries lesion of another patient was examined. The fluorescence spectrum of this lesion is shown

as ´Caries 2´ in Fig. 7. In this spectrum also porphyrin fluorescence is observed. However, compared to the spectrum

´Caries 1´ the maxima of the emission are shifted about 10nm to shorter wavelength. It is known that different types of

porphyrins have slightly different emission spectra. Thus one can speculate that this shift is caused by different bacteria

with different metabolisms that have produced different types of porphyrin. But since it is known that the emission

spectrum of a specific porphyrin changes with the type of solvent the shift could also be interpreted as being due to

differences in the local environments of the molecule (pH-value, composition of saliva and bacterial film etc.).

A similar situation can be found in a single tooth having caries lesions of different progression as shown in Fig. 3. Here

the initial caries lesion shows emission peaks shifted to longer wavelength compared to the peaks of the much more

progressed caries lesion. Thus also in this case one can speculate that in the differently progressed caries lesions

different bacteria are dominant that produce different types of porphyrin. Further it should be noted that the absolute

luminescence intensity in the more progressed caries lesion has been significantly lower than in the initial caries lesion.

This can be explained by a significant absorption of the porphyrin luminescence in the brown coloured lesion. Therefore

the absolute porphyrin luminescence intensity is not a good measure for the progression of carious defects. However,

the normalized emission spectra in Fig. 3 show that the emission spectrum of the much more progressed caries lesion

has larger intensity contributions in the red range of the visible spectrum than the initial lesion. Hence the ratio of the

luminescence intensities in the red and the green part of the visible spectrum is a much better indicator of the activity of

the defect than the amount of the red luminescence intensity itself.

3.2. Supragingival plaque

Since the development of caries starts with the accumulation of plaque on the tooth surfaces the fluorescent emission of

plaque has also been investigated. In Fig. 4 the emission spectra of plaque and a typical image taken with the

fluorescence camera are shown. It can be seen that the emission spectra of the plaque of Patient 1 is quite similar to the

spectrum of the initial caries as shown in Fig. 3. Again emission bands similar to the emission of porphyrins are

observed. But the emission spectrum of the plaque of patient 2 differs from the one of patient 1 and shows similarities

with the emission spectrum of initial caries in Fig. 3. However, in the case of the plaque of patient 2 an additional band

at about 590nm is present. This large spectral shift can not be explained by a different local environment. It is more

likely that there is a different bacterium present in the plaque of patient 2 that produces in its metabolism a different

fluorophore that is probably another type of porphyrin. Further measurements of the emission spectra for other patients

showed still different emission spectra of the plaque than for patient 1 or patient 2. Accordingly the composition of the

fluorophores in the plaque differs from patient to patient. This reflects a large variety in the composition of plaque by

different bacterial species producing different fluorophores or porphyrins.

3.3. Dental calculus

As shown in Fig. 5 strong fluorescence emission in the red spectral range can also be observed from dental calculus

under the excitation at 405nm. The emission spectrum, as shown in Fig. 7, is very similar to the emission spectra of

caries lesions. Accordingly it is also in the case of dental calculus very probable that the fluorophor is a porphyrin. This

porphyrin has been produced by bacteria that have been present in or on the surface of the calculus and has been

accumulated there.

3.4. Subgingival concrements

Like dental calculus also subgingival concrements on root surfaces emit enhanced red luminescence when excited with

light of a wavelength of 405nm as shown in Fig. 6. Again the emission spectrum shows the fingerprint of porphyrins

with a band at about 640nm and a another broader band at about 700nm. Also the emission band that has been observed

for the plaque of Patient 2 as shown in Fig. 4 at about 590nm is present. Similar to the emission spectrum of enamel as

shown in Fig. 4 and 7 the emission spectrum of the root surface without concrement decreases with increasing

wavelength above 520nm. The cut off below 500nm is caused by the optical long pass filter in front of the spectrometer.

Similar spectra but with missing band at 590nm have been obtained on the concrements of other root surfaces. This

indicates that different bacterial species with different emission spectra accumulate on and in the concrements.

3.5. Dental bacteria in vitro

There is a large variety of bacteria in the oral cavity. Some specific types are known to have a high correlation with a

dental disease. I.e. Streptococcus Mutans, Streptococcus Sobrinus and Lactobacillus have a high correlation i.e. with

dental caries. For i.e. the bacteria Actinobacillus Actinomycetemcomitans, Porphyromonas Gingivalis, Bacteroides

Forsythus, Prevotella Intermedia and Fusobacterium Nucleatum a correlation to the periodontal disease exist. Therefore

some specific bacteria have been grown on agar, removed from their substrate and spectroscopically investigated upon

excitation with light of 405nm on a glass slide. The emission spectra are shown in Fig. 7 on the right side. It can be

seen, that the porphyrin related luminescence with a band at about 640nm and another broader band at about 700nm can

be clearly observed for Actinobacillus Actinomycetemcomitans and Fusobacterium Nucleatum that are correlated with

the periodontal disease. Porphyromonas Gingivalis which is also correlated with the periodontal disease shows only a

very low band at about 625nm. The bacterium Streptococcus Mutans that is correlated with dental caries shows

emission bands at about 580nm and 640nm of somewhat lower intensity than Actinobacillus Actinomycetemcomitans

and Fusobacterium Nucleatum. Interesting is that for this bacterium the ratio of the red to green luminescence is

significantly higher than for enamel. Thus the ratio of the red to green luminescence as suggested earlier can be a much

better indicator for dental caries than the red intensity itself. For the Lactobacillae no significant luminescence has been

found.

4. CONCLUSIONS

It has been shown in the previous section exemplary that plaque, carious lesions, dental calculus and subgingival

concrements emit red light under the excitation with light of 405nm. The band structure of the emission spectra,

however, is complex and differs often from one area to another. On the other hand enamel, dentin and healthy root

surfaces without concrements show a fluorescence that decreases significantly with increasing wavelength from the

green at about 510nm to the red. Further it has been found that at least some of the relevant bacteria that are correlated

with dental diseases show enhanced red fluorescence. Thus the described fluorescent camera is probably a very good

tool to distinguish healthy dental surfaces from infected sites where the bacterial concentration is out of balance or

healthy limits. The fluorescence camera has the advantage that large areas of dental surfaces can be inspected with high

spatial resolution within seconds, which is impossible i.e. with the Diagnodent (Kavo) that uses only a fibreoptic probe

to inspect fluorescence intensities in the IR point by point by moving the probe. Due to this high spatial resolution it is

possible to identify even small sites with this bacterial imbalance within short time and to store and document them on a

PC. For the detection of these sites it is advantageous to compare the ratio of the red to green fluorescence at these sites

with the ratio of uninfected sites. It is especially of an advantage, that for the examples shown above the sensitivity is

that high that already the plaque formation can be diagnosed, from which later on the dental disease could follow. Thus

it seems also appropriate to use the fluorescence camera to identify the hygienic status of patients. Additionally the

images with the luminescent plaque can be used to motivate the patients to enhance their activities in the removal of

plaque. Further, after removal of the plaque from the surfaces of the teeth caries lesions in an early stage can be

diagnosed with the camera and care can be taken to remineralize the defects. However, studies that prove the correlation

of dental diseases like caries with the output of the fluorescence camera over a large number of patients are necessary.

For the Diagnodent system, that excites porphyrins in the long wavelength Q-band at 655nm and measures the emission

intensity above 680nm such studies already have been made for dentin caries and show in general a high specificity and

sensitivity in comparison with histology, microradiography or clinical opening. It should be noted here that the

excitation process of the porphyrin is more efficient in the Soret band that is used in the fluorescence camera than in the

Q-bands at longer wavelength that is used by the Diagnodent system. Further the output power of the excitation light is

in the case of the fluorescence camera 60mW whereas the optical output power of the Diagnodent system is below 1

mW.

With a miniaturized camera head the fluorescence camera also has the potential to identify harmful subgingival plaque

in the tooth pocket. This is supported by the fact that it has been found that at least Actinobacillus

Actinomycetemcomitans and Fusobacterium Nucleatum which are correlated with the periodontal disease show

enhanced red fluorescence compared with the healthy root surface.

From the point of the application it is also interesting to point out that the excitation light with the wavelength of 405nm

has a penetration depth of the order of at least millimeters into healthy enamel and dentin. Thus the excitation light can

excite fluorophores that are not directly on the surface of the tooth but more inside the volume of the tooth. In addition

enamel and dentin show low absorption to emitted red light. Therefore it is possible to excite and collect with the

fluorescent camera signals from points that have a distance of the order of millimeters from the surface. Due to light

scattering in the enamel and dentin the emission signals of such deeper points are spatially unsharp and diffuse.

Nevertheless there is a chance to find defects or lesions inside the tooth that are not directly located on the surface of the

tooth like it is i.e. the case for secondary caries between fillings and the enamel or dentin of for defects on directly

inaccessible approximal surfaces. At one patient for example it was possible to identify a caries lesion below the enamel

of an incisor with the fluorescence camera. The same lesion was impossible to identify by the usual visual inspection

because the caries lesion started from the approximal side between the incisors. However, upon mechanical probing the

enamel over the cavity was broken and the cavity became visible.

The red light emission originates from fluorophores that are produced by the metabolism of some intraoral bacteria.

Since the emission spectra show similarities with the emission spectra of porphyrins it is reasonable to conclude that the

relevant fluorophores are porphyrins. However, there is in principal a large variety of porphyrins with different side

groups can be produced. Therefore it would be interesting to identify the actually produced type of porphyrin for each

bacterial species that gives rise the emission of red light. Some of the bacteria that give rise to the red fluorescence have

been identified. But there is a much larger variety of bacteria present in the oral cavity. Therefore additional effort is

needed in future to identify more of the bacterial species that show red fluorescence upon excitation in the Soret band of

the prophyrins. Of special interest are here bacteria like Streptococcus Sobrinus, Actinomyces Odontologica,

Actinomyces Naeslundi that are associated with caries and Bacteroides Forsythus and Prevotella Intermedia that are

associated with periodontitis.

5. ACKNOWLEDGEMENTS

The author likes to acknowledge the technical support and the assistance that was provided for the setu

Maybe my publication will help?

Hartstone-Rose A, Kuhn B, Werdelin L, Nalla S, Berger L. 2013. A new species of fox from the Australopithecus sediba type locality, Malapa, South Africa. Transactions of the Royal Society of South Africa. 68(1):1-9.

Szuma E (2000) Variation and correlation patterns in the dentition of the red fox from Poland.  Ann Zool Fennici 37:113-127

you might be interested in a new intraoral scanner

Optics and Lasers in Engineering

Volume 54, March 2014, Pages 187-196