Caries detection and diagnosis: Novel technologies

Transcription

1 journal of dentistry 34 (2006) available at journal homepage: Review Caries detection and diagnosis: Novel technologies Iain A. Pretty * Dental Health Unit, 3A Skelton House, Lloyd Street North, Manchester Science Park, Manchester M15 6SH, UK article info Article history: Received 30 January 2006 Received in revised form 30 May 2006 Accepted 1 June 2006 Keywords: Caries Detection QLF ECM Accuracy Reliability Sensitivity Specificity abstract Recent years have seen an increase in research activity surrounding diagnostic methods, particularly in the assessment of early caries lesions. The drive for this has come from two disparate directions. The first is from the dentifrice industry who are keen to develop techniques that would permit caries clinical trials (CCTs) to be reduced in duration and subject numbers to permit the investigation of novel new anti-caries actives. The second is from clinicians who, armed with the therapies to remineralise early lesions are now seeking methods to reliably detect such demineralised areas and implement true preventative dentistry. This review examines novel technologies and the research supporting their use. Techniques based on visual, optical, radiographic and some emerging technologies are discussed. Each have their benefits although systems based on the auto-fluorescence (such as QLF) of teeth and electrical resistance (such as ECM) seem to offer the most hope for achieving reliable, accurate detection of the earliest stages of enamel demineralisation. # 2006 Elsevier Ltd. All rights reserved. 1. Introduction Our understanding of the caries process has continued to advance, with the vast majority of evidence supporting a dynamic process which is affected by numerous modifiers tending to push the mineral equilibrium in one direction or another, i.e. towards remineralisation or demineralisation. 1 All of these interactions are taking place in the complex biofilm overlaying the tooth surface which comprises of the pellicle as well as the oral microflora of the plaque. 2 The modifiers of this system are well known and are summarised in Table 1 with Fig. 1 presenting an overview of the dynamics of the caries process. 2 With this greater understanding of the disease, comes an opportunity to promote preventative therapies that encourage the remineralisation of non-cavitated lesions resulting in inactive lesions and the preservation of tooth structure, function and aesthetics. Central to this vision is the ability to detect caries lesions at an early stage and correctly quantify the degree of mineral loss, ensuring that the correct intervention is instigated. 3,4 The failure to detect early caries, leaving those detectable only at the deep enamel, or cavitated stage has resulted in poor results and outcomes for remineralisation therapies. A range of new detection systems have been developed and are either currently available to practitioners or will shortly be made so. It is a crucial distinction that the systems described within this review are correctly classified as caries detection systems, rather than diagnostic systems. Diagnosis is a decision process that rests with the clinician and is informed by, initially, detection of a lesion and should be followed by an assessment * Tel.: ; fax: address: iain.pretty@man.ac.uk /$ see front matter # 2006 Elsevier Ltd. All rights reserved. doi: /j.jdent

2 728 journal of dentistry 34 (2006) Table 1 Risk and modifying factors for caries Primary risk factors Saliva (1) Ability of minor salivary glands to produce saliva (2) Consistency of unstimulated (resting) saliva (3) ph of unstimulated saliva (4) Stimulated salivary flow rate (5) Buffering capacity of stimulated saliva Diet (6) Number of sugar exposures per day (7) Number of acid exposures per day Fluoride (8) Past and current exposure Oral biofilm (9) Differential staining (10) Composition (11) Activity Modifying factors (12) Past and current dental status (13) Past and current medical status (14) Compliance with oral hygiene and dietary advice (15) Lifestyle (16) Socioeconomic status of the patient s caries risk which may include the number of new caries lesions, past caries experience, diet, presence or absence of favourable or unfavourable modifying factors (salivary flow, mutans streptococci counts, oral hygiene) and qualitative aspects of the disease such as colour and anatomical location. 5 These detection systems are therefore aimed at augmenting the diagnostic process by facilitating either earlier detection of the disease or enabling it to be quantified in an objective manner. Visual inspection, the most ubiquitous caries detection system, is subjective. Assessment of features such as colour and texture are qualitative in nature. These assessments provide some information on the severity of the disease but fall short of true quantification. 6 They are also limited in their detection threshold and their ability to detect early, non cavitated lesions restricted to enamel is poor. It is this ability to quantify and/or detect lesions earlier that the novel diagnostic systems offer to the clinician. Pitts provides a useful visual description of the benefits of early caries detection. 7 Using the metaphor of an iceberg, it can be seen that traditional methods of caries detection result in a vast quantity of undetected lesions. There is a clinical argument about the significance of these lesions, with some authors believing that only a small percentage will progress to more severe disease, however, it is a undisputed fact that all cavitated lesions with extension in pulp began their natural history as an early lesion. Fig. 2 demonstrates the Pitts iceberg. From this it can be seen that as the sensitivity of the detection device increases, so does the number of lesions detected. It can also be seen that the new detection tools are required to identify those lesions that would be amenable to remineralising therapies. 8 When assessing the effectiveness of such methods, the preferred reporting metrics are those of traditional diagnostic science; namely specificity, sensitivity, area under the ROC curve and the correlation with the truth (the true state of the disease, established using a gold standard). The reliability or reproducibility of the test can be established using either intra-class correlation or kappa coefficients depending on the nature of the metric output, i.e. either continuous or ordinal. 9,10 Novel diagnostic systems are based upon the measurement of a physical signal these are surrogate measures of the caries process. Examples of the physical signals that can be used in this way include X-rays, visible light, laser light, electronic current, ultrasound, and possibly surface roughness. 11 For a caries detection device to function, it must be capable of initiating and receiving the signal as well as being able to interpret the strength of the signal in a meaningful way. Table 2 demonstrates the physical principles and the detection systems that have taken advantage of them. 11 It is worthwhile to take an overview of the performance of the traditional caries detection systems and these are shown, in terms of sensitivity and specificity in Figs. 3 and 4. Fig. 3 demonstrates the methods performance irrespective of the severity of the lesion, with Fig. 4 presenting the same data for lesions confined to enamel. These data are based on the excellent systematic review by Bader et al. 12 who restricted his assessment of studies to those that employed histological validation. This therefore indicates that while the true diagnostic outcome is not in doubt, these studies were conducted in vitro and hence the actual values in clinical Fig. 1 Demineralisation and remineralisation cycle for enamel caries (adapted from Mount and Hume 56 ).

3 journal of dentistry 34 (2006) Fig. 2 The iceberg of caries and the influence of detection system (modified from Pitts, ). Table 2 Methods of caries detection based on their underlying physical principles Physical principle Application in caries detection X-rays Visible light Laser light Electrical current Ultrasound Modified from literature. 51 Digital subtraction radiography Digital image enhancement Fibre optic transillumination (FOTI) Quantitative light-induced fluorescence (QLF) Digital image fibre optic transillumination (DiFOTI) Laser fluorescence measurement (DiagnoDent) Electrical conductance measurement (ECM) Electrical impedance measurement Ultrasonic caries detector practice are likely to be poorer. A scant assessment of the figures indicates that while specificity is adequate, the sensitivity scores of the traditional methods are poor, with many being significantly less than chance; i.e. a guess would provide the same or a better result in many cases. These figures serve to illustrate the need for detection devices that are objective, quantitative, sensitive and enable early lesions to be monitored over time. This longitudinal monitoring is especially important when one considers the treatment of early caries lesions. The following review describes those systems that have potential to meet the aims of clinicians and researchers for enhanced sensitivity and objective, metric, continuous measures of mineralisation status. 2. Detection systems based on electrical current measurement Every material possesses its own electrical signature; i.e. when a current is passed through the substance the properties of the material dictate the degree to which that current is conducted. Conditions in which the material is stored, or physical changes to the structure of the material will have an effect on this conductance. 11 Biological materials are no exception and the concentration of fluids and electrolytes contained within such Fig. 3 Effectiveness of traditional caries detection systems based on lesions of any severity, after Bader et al. 12 Fig. 4 Effectiveness of traditional caries detection systems based on lesions restricted to enamel only, after Bader et al. 12

4 730 journal of dentistry 34 (2006) materials largely govern their conductivity. 13 For example, dentine is more conductive than enamel. In dental systems, there is generally a probe, from which the current is passed, a substrate, typically the tooth, and a contra-electrode, usually a metal bar held in the patient s hand. Measurements can be taken either from enamel or exposed dentine surfaces. 14 In its simplest form, caries can be described as a process resulting in an increase in porosity of the tissue, be it enamel or dentine. This increased porosity results in a higher fluid content that sound tissue and this difference can be detected by electrical measurement by decreased electrical resistance or impedance. 3. Electronic caries monitor (ECM) The ECM device employs a single, fixed-frequency alternating current which attempts to measure the bulk resistance of tooth tissue 15 (see Fig. 5). This can be undertaken at either a site or surface level. When measuring the electrical properties of a particular site on a tooth, the ECM probe is directly applied to the site, typically a fissure, and the site measured. During the 5 s measurement cycle, compressed air is expressed from the tip of the probe and this results in a collection of data over the measurement period, described as a drying profile, that can provide useful information for characterising the lesion. An example of this is shown in Fig. 6. While it is generally accepted that the increase in porosity associated with caries is responsible for the mechanism of action for ECM, 15 there are some points to consider: (1) Do electrical measurements of carious lesions measure the volume of the pores, and if so, is it the total pore volume or Fig. 6 A demonstration of an ECM profile obtained from a primary root caries lesion in vitro demonstrating the sites assessed. just a portion, perhaps the superficial portion, that is measured? (2) Do electrical measurements measure pore depth? If this is the case, what happens during remineralisation where the superficial layer may remineralise, leaving a pore beneath? (3) Is the morphological complexity of the pores a factor in the measurement of conductivity? There are also a number of physical factors that will affect ECM results. These include such things as the temperature of the tooth, 16 the thickness of the tissue, 17 the hydration of the Fig. 5 The ECM device (Version 4) and its clinical application. (a) The ECM machine, (b) the ECM handpiece, (c) site specific measurement technique, (d) surface specific measurement technique.

5 journal of dentistry 34 (2006) Table 3 ECM ROC areas under the curve ROC area Diagnostic threshold Tooth type Surface or site specific measurement Study 0.82 D1 Premolars Site specific 0.80 D1 Molars Site specific 0.84 D3 Premolars Site specific 0.82 D3 Molars Site specific 0.80 D1 Premolars Surface specific 0.67 D1 Premolars Surface specific 0.94 D3 Premolars Surface specific 0.79 D3 Molars Surface specific material (i.e. one should not dry the teeth prior to use) and the surface area. 15 An excellent review of the performance of ECM was undertaken in 2000 by Huysmans 18 who collated information from a variety of validation studies. She was unable to perform a meta-analysis of these data; stating that aspects of the studies such as version of equipment, storage medium, cutoffs and tooth type prevented direct comparisons. A summary of her findings are presented in Table 3, these demonstrate a good to excellent range of area under the curves (AUC s) with the exception of surface specific premolars when assessing at the D1 level (lesions restricted to enamel). The sensitivity and specificity values were assessed from a number of studies; for site specific measurements these were; sensitivity 74.8( 11.9) and specificity 87.6( 10) and for surface specific measurements; 63( 2.8) and 79.5( 9.2). The lower efficacy in surface specific measurements has led to this area of research being neglected, with the vast majority of publications concentrating on site specific measurements. The reproducibility of the device has been assessed in a number of publications and has been rated as good to excellent for both measurement techniques. The intra-class correlation coefficients for site specific were 0.76 and 0.93 for surface specific. 19 It is important to note that these high figures relate to the use of the device in a controlled, laboratory setting. Further studies in vitro are required before the device can be used for monitoring lesions longitudinally. For example, some authors have stated that the limits of agreement can be as much as 580 kv for surface specific measurements. If the range for an occlusal surface is considered as kv then this could be a substantial source of error. 20 A clinical trial has been undertaken using the ECM device on root caries, and the successful outcome of this study suggests that dentine may be a more suitable tissue for ECM. The study assessed the effect of 5000 ppm fluoride dentifrice against 1100 ppm on 201 subjects with at least 1 root caries lesion. These were site specific measurements taken using the airflow function of the ECM unit. After 3 and 6 months, there was statistical difference between the two groups, with the higher fluoride group showing a better remineralising capability than the lower fluoride paste users 21 (see Fig. 7). This is good evidence to suggest that ECM is capable of longitudinal monitoring and that clinicians may be able to employ the device to monitor attempts at remineralising, and thus potentially arresting, root caries lesions in their patients. 21 A further application of electronic monitoring of caries is that of Electrical Impedance Spectroscopy or EIS. Unlike ECM which uses a fixed frequency (23 Hz), EIS scans a range of electrical frequencies and provides information on capacitance and impendence among others. 22 This process provides the potential for more detailed analysis of the structure of the tooth to be developed, including the presence and extent of caries. A prototype has been developed and is being commercially exploited and a market release is expected in Radiographic techniques 4.1. Digital radiographs Digital radiography has offered the potential to increase the diagnostic yield of dental radiographs and this has manifested itself in subtraction radiography. A digital radiograph (or a traditional radiograph that has been digitised) is comprised of a number of pixels. Each pixel carries a value between 0 and 255, with 0 being black and 255 being white. The values in between represent shades of grey, and it can be quickly appreciated that a digital radiograph, with a potential of 256 grey levels has significantly lower resolution than a conventional radiograph that contain millions of grey levels. This would suggest that digital radiographs would have a lower diagnostic yield than that of traditional radiographs. Research has confirmed this; with sensitivities and specificities of digital radiographs being significantly lower than those of regular radiographs when assessing small proximal lesions. 23 However, digital radiographs offer the potential of image Fig. 7 ECM values from a root caries study using high and low concentrations of fluoride dentifrices. The increasing ECM values relate to a reduction in porosity and increase in electrical resistance.

6 732 journal of dentistry 34 (2006) Fig. 8 Comparison of regular and enhanced digital radiographs. (a) Digital radiograph, (b) enhanced radiograph where the interproximal lesions between first molar and second premolar can be seen more clearly. enhancement by applying a range of algorithms, some of which enhance the white end of the grey scale (such as Rayleigh and hyperbolic logarithmic probability) and others the black end (hyperbolic cube root function). When these enhanced radiographs are assessed their diagnostic performance is at least as good as conventional radiographs, 24 with reported values of 0.95 (sensitivity) and 0.83 (specificity) for approximal lesions. See Fig. 8 for an example of this enhancement. When these findings are considered, one must remember that digital radiographs offer a decrease in radiographic dose and thus offer additional benefits than diagnostic yield. Digital images can also be archived and replicated with ease Subtraction radiology As described above, using digital radiographs offers a number of opportunities for image enhancement, processing and manipulation. One of the most promising technologies in this regard is that of radiographic subtraction which has been extensively evaluated for both the detection of caries and also the assessment of bone loss in periodontal studies. 25 The basic premise of subtraction radiology is that two radiographs of the same object can be compared using their pixel values. If the images have been taken using either a geometry stabilising system (i.e. a bitewing holder) or software has been employed to register the images together, then any differences in the pixel values must be due to change in the object. 26 The value of the pixels from the first object are subtracted from the second image. If there is no change, the resultant pixel will be scored 0; any value that is not 0 must be attributable to either the onset or progression of demineralisation, or regression. Subtraction images therefore emphasise this change and the sensitivity is increased. It is clear from this description that the radiographs must be perfectly, or as close to perfect as possible, aligned. Any discrepancies in alignment would result in pixels being incorrectly represented as change. 27 Several studies have demonstrated the power of this system, with impressive results for primary and secondary caries. However, uptake of this system has been low, presumably due to the need for well aligned images. Recent advances in software have enabled two images with moderate alignment to be correctly aligned and then subtracted. 27 This may facilitate the introduction of this technology into mainstream practice where such alignment algorithms could be built into practice software currently used for displaying digital radiographs. An example of a subtraction radiograph is shown in Fig. 9. Fig. 9 Example of a subtraction of two digital bitewing radiographs. (a) Radiograph showing proximal lesion on mesial surface of first molar, (b) follow up radiograph taken 12 months later, (c) the areas of difference between the two films are shown as black, i.e. in this case the proximal lesion has become more radiolucent and hence has progressed.

7 journal of dentistry 34 (2006) Fig. 11 FOTI equipment. Fig. 10 Example of early lesions before (a) and after (b) drying. 5. Enhanced visual techniques 5.1. Fibre optic transillumination (FOTI and DiFOTI) The basis of visual inspection of caries is based upon the phenomenon of light scattering. Sound enamel is comprised of modified hydroxyapatite crystals that are densely packed, producing an almost transparent structure. The colour of teeth, for example, is strongly influenced by the underlying dentin shade. When enamel is disrupted, for example in the presence of demineralisation, the penetrating photons of light are scattered (i.e. they change direction, although do not loose energy) which results in an optical disruption. In normal, visible light, this appears as a whiter area the so called white spot. 28 This appearance is enhanced if the lesion is dried; the water is removed from the porous lesion. Water has a similar refractive index (RI) to enamel, but when it is removed, and replaced by air, which has a much lower RI than enamel, the lesion is shown more clearly. This demonstrates the importance of ensuring the clinical caries examinations are undertaken on clean, dry teeth 29 (see Fig. 10). Fibre optic transillumination takes advantage of these optical properties of enamel and enhances them by using a high intensity white light that is presented through a small aperture in the form of a dental handpiece. Light is shone through the tooth and the scattering effect can be seen as shadows in enamel and dentine, with the device s strength the ability to help discriminate between early enamel and early dentine lesions (see Fig. 11). A further benefit of FOTI is that it can be used for the detection of caries on all surfaces; and is particularly useful at proximal lesions. The research around FOTI is somewhat polarised, with a recent review finding a mean sensitivity of only 14 and a specificity of 95 when considering occlusal dentine lesions, and 4 and 100% for Fig. 12 Example of FOTI on a tooth. (a) Normal clinical vision, (b) with FOTI.

8 734 quantified using proprietary software and has been shown to correlate well with actual mineral loss; r = The source of the auto-fluorescence is thought to be the enamel dentinal junction the excitation light passes through the transparent enamel and excites fluorophores contained within the EDJ. Studies have shown that when underlying dentine is removed from the enamel, fluorescence is lost, although only a small amount of dentine is required to produce the fluorescence seen. 37 Decreasing the thickness of enamel results in a higher intensity of fluorescence. The presence of an area of demineralised enamel reduced the fluorescence for two main reasons. The first is that the scattering effect of the lesion results in less excitation light reaching the EDJ in this area, and the second is that any fluorescence from the EDJ is back scattered as it attempts to pass through the lesion. The QLF equipment is comprised of a light box containing a xenon bulb and a handpiece, similar in appearance to an intraoral camera, see Fig. 13. Light is passed to the handpiece via a liquid light guide and the handpiece contains the bandpass filter. 38 Live images are displayed via a computer and accompanying software enables patient s details to be entered and individual images of the teeth of interest to be captured and stored. QLF can image all tooth surfaces except interjournal of dentistry 34 (2006) proximal lesions. 30 This is in contrast to other studies where sensitivity was recorded at 85% and specificity at 99%. 31 Many of the differences can be explained by the nature of the ordinal scale used to record the subjective visual assessment and the gold standard used to validate the method. However, one would expect FOTI to be at least as effective as a visual examination. Recent developments in ordinal scales for visual assessments, such as the ICDAS scoring system, 32 may enable a more robust framework for visual exams into which FOTI can be added (Fig. 12). One would expect that FOTI would enable discrimination of occlusal lesions to be improved (particularly dentine lesions), as well as detection of proximal lesions (in the absence of radiographs) to be higher. 29 As a technique FOTI is an obvious choice for translation into general practice; the equipment is economical, the learning curve is short and the procedure is not time consuming. Indeed, some work has been undertaken trialling the use of FOTI in practice with encouraging results. 33 However with the simplicity of the FOTI system come limitations; the system is subjective rather than objective, there is no continuous data outputted and it is not possible to record what is seen in the form of an image. Longitudinal monitoring is, therefore, a complex matter and some degree of training is required in order to be competent at this level of FOTI usage. In order to address some of these concerns, an imaging version of FOTI has been developed; digital imaging FOIT (DiFOTI). This system comprises of a high intensity light and grey scale camera which can be fitted with one of two heads; one for smooth and one for occlusal surfaces. Images are displayed on a computer monitor and can be archived for retrieval at a repeat visit. However, there is no attempt within the software to quantify the images, and analysis is still undertaken visually by the examiner who makes a subjective call based on the appearance of scattering Fluorescent techniques 6.1. Visible light fluorescence QLF Quantitative light-induced fluorescence (QLF) is a visible light system that offers the opportunity to detect early caries and then longitudinally monitor their progression or regression. Using two forms of fluorescent detection (green and red) it may also be able to determine if a lesion is active or not, and predict the likely progression of any given lesion. Fluorescence is a phenomenon by which an object is excited by a particular wavelength of light and the fluorescent (reflected) light is of a larger wavelength. When the excitation light is in the visible spectrum, the fluorescence will be of a different colour. In the case of the QLF the visible light has a wavelength (l) of 370 nm, which is in the blue region of the spectrum. The resultant auto-fluorescence of human enamel is then detected by filtering out the excitation light using a bandpass filter at l > 540 nm by a small intra-oral camera. This produces an image that is comprised of only green and red channels (the blue having been filtered out) and the predominate colour of the enamel is green. 35,36 Demineralisation of enamel results in a reduction of this auto-fluorescence. This loss can be Fig. 13 QLF Equipment. (a) The QLF unit light box, demonstrating the handpiece and liquid light guide; (b) a close-up of the intra-oral camera featuring a disposable mirror tip that also acts as an ambient light shield.

9 journal of dentistry 34 (2006) Fig. 14 Example of QLF images. (a) White light image of early buccal caries effecting the maxillary teeth, (b) QLF image taken at the same time as (a), note the improved detection of lesions as a result of the increased contrast between sound and demineralised enamel, (c) 6 months after the institution of an oral hygiene programme, the lesions have resolved. proximally. See Fig. 14 for an example of QLF images that have been merged to create a montage on the anterior teeth demonstrating resolution of buccal caries over a 1 month period following supervised brushing. Once an image of a tooth has been captured, the next stage is to analyse any lesions and produce a quantitative assessment of the demineralisation status of the tooth. This is undertaken using proprietary software and involves using a patch to define areas of sound enamel around the lesion of interest. Following this the software uses the pixel values of the sound enamel to reconstruct the surface of the tooth and then subtracts those pixels which are considered to be lesion. This is controlled by a threshold of fluorescence loss, and is generally set to 5%. This means that all pixels with a loss of fluorescence greater than 5% of the average sound value will be considered to be part of the lesion. Once the pixels have been assigned sound or lesion the software then calculates the average fluorescence loss in the lesion, known as %DF, and then the total area of the lesion in mm 2, a the multiplication of these two variables results in a third metric output, DQ. See Fig. 15 for an example of the analysis and the resultant lesion. When examining lesions longitudinally, the QLF device employs a video repositioning system that enables the precise geometry of the original image to be replicated on subsequent visits. QLF has been employed to detect a range of lesion types. For occlusal caries sensitivity has been reported at 0.68 and specificity at 0.70, and this compares well with other systems. Correlations of up to 0.82 have also been reported for QLF metrics and lesion depth. Smooth surfaces, secondary caries and demineralisation adjacent to orthodontic brackets have all been examined. The reliability of both stages of the QLF process; i.e. the image capture and the analysis; have been examined and has been shown to be substantial. Intraclass correlation coefficients have been reported as 0.96 for image capture, with analysis at 0.93 for intra-examiner and 0.92 for inter-examiner comparisons. Again, these compare well to other systems. The QLF system offers additional benefits beyond those of very early lesion detection and quantification. The images acquired can be stored and transmitted, perhaps for referral purposes, and the images themselves can be used as patient motivators in preventative practice. For clinical research use, the ability to remotely analyse lesions enables increased legitimacy in trials; permitting, for example, a repeat of the analyses to be conducted by a third-party. QLF is one of the most promising technologies in the caries detection stable at present, although further research is required to demonstrate its ability to correctly monitor lesion changes over time. There is also a great deal of interest in red fluorescence, and whether or not this can be a predictor of lesion activity and again, research is currently being undertaken in this area Laser fluorescence DIAGNODent The DIAGNODent (DD) instrument (KaVo, Germany) is another device employing fluorescence to detect the presence of caries. Using a small laser the system produces an excitation wavelength of 655 nm which produces a red light. This is carried to one of two intra-oral tips; one designed for pits and fissures, and the other for smooth surfaces. The tip both emits the excitation light and collects the resultant fluorescence. Unlike the QLF system, the DD does not produce an image of the tooth; instead it displays a numerical value on two LED displays. The first displays the current reading while the second displays the peak reading for that examination. A small twist of the top of the tip enables the machine to be reset and ready for another site examination and a calibration device is supplied with the system. There has been some debate over what exactly the DD is measuring; it is not employing the intrinsic changes within the enamel structure in the same way as QLF; this has been demonstrated by the inability of DD to detect artificial lesions in in vitro settings. Instead the system is thought to measure the degree of bacterial activity; and this is supported by the fact that the excitation wavelength is suitable for inducing fluorescence from bacterial porphyrins; a byproduct of metabolism (Figs. 15 and 16). Initial evaluations of the device suggest that it may be a promising tool for clinical use; correlation with histological depth of lesions was substantial at 0.85 and the sensitivity and specificity for dentinal lesions were 0.75 and 0.96, respectively. 39 Reliability of the device measured by Kappa was for intra-examiner and for inter-examiner. 40 Further in vitro studies have found that the area under the ROC was significantly higher for DD (0.96) than that for conventional radiographs (0.66). 41 However, the device is not

10 736 journal of dentistry 34 (2006) Fig. 15 An example of lesion analysis using QLF. (a) Lesion on the occlusal surface of a premolar is identified and the analysis patch placed on sound enamel, (b) the reconstruction demonstrates correct patch placement as the surface now looks homogenous, (c) the subtracted lesion is demonstrated in false colour indicating the severity of the demineralisation, (d) the quantitative output from this analysis at a variety of fluorescent threshold levels. without its confounders, and, like many novel caries detection devices, requires teeth to be clean and dry. The presence of stain, calculus, plaque and, when used in the laboratory, the storage medium, have all be shown to have an adverse effect on the DD readings. 39 Most confounders tend to cause an increase in the DD reading, leading to false-positives. The literature surrounding the DD device was recently assessed in a systematic review. 42 The authors found that, for dentinal caries, the DD device performed well, although there was a great deal of heterogeneity in the studies and they were all undertaken in vitro. The authors stated that these results could not be extrapolated into the clinical setting and then detected a worrying trend for the device to produce more false-positives than traditional diagnostic systems. Their conclusion was therefore that there was insufficient evidence to support the use of the device as a principle means of caries diagnosis in clinical practice. 42 It should be noted that the DD device has not been employed in a clinical trial, so there are no data indicating that the system can detect a dose response Other optical techniques There are a number of other techniques for detecting caries using optical methods. These systems are in their infancy and many are based solely in laboratories. However, such technologies may prove useful in the future. Examples include optical coherence tomography (OCT), and near infra-red imaging. OCT has been shown to be able to image early enamel caries lesions in extracted teeth, 43 and also on root lesions. 44 Like many other novel techniques, it is likely that stain will adversely effect OCT. 45 Work has just begun on using near infra-red, but initial results look promising. 46 There is significant work involved in developing these systems into clinically and commercially acceptable applications and so it could be some time until these new methodologies can be properly assessed in clinical trials Ultrasound techniques Fig. 16 The DIAGNODent device. The use of ultrasound in caries detection was first suggested over 30 years ago, although developments in this field have been slow. The principle behind the technique is that sound

11 journal of dentistry 34 (2006) waves can pass through gases, liquids and solids and the boundaries between them. 45 Images of tissues can be acquired by collecting the reflected sound waves. In order for sound waves to reach the tooth they must pass first through a coupling mechanism, and a number of these have been suggested, but those with clinical applications include water and glycerine. 45 A number of studies have been undertaken using ultrasound, with differing levels of success. One study reported that an ultrasound device could discriminate between cavitated and non-cavitated inter-proximal lesions 47 in vitro. A further study found that ultrasonic measurements at 70 approximal sites in vitro resulted in a sensitivity of 1.0 and a specificity of 0.92 when compared to a histological gold standard. 48 Further histological validation has been undertaken by using transverse microradiography and ultrasound. 49 A final in vivo study was undertaken using a device described as the Ultrasonic Caries Detector (UCD) which examined 253 approximal sites and claimed a diagnostic improvement over bitewing radiography. 50 Despite these encouraging findings, no further research has been undertaken using the device and the research has only been published as abstracts. 7. Conclusion A range of caries detection systems have been covered in this review. A summary of their performance is presented in Fig. 17. The pattern of dental caries is changing, with an increasing incidence in occlusal surfaces. This shift has rendered traditional detection systems, particularly bitewing radiographs less useful in the diagnostic protocols of clinicians. High concentration fluoride varnishes have been demonstrated to arrest the progression of early lesions, but often traditional methods of detection are too insensitive to permit the most efficacious use of these products. Caries clinical trials involving thousands of subjects over several years employ are no longer commercially viable. For all of these reasons, there is a real need for a range of caries detection and quantification systems to augment the clinician s diagnostic pathway. Fig. 17 A summary of the diagnostic performance (validity and reliability) of a range of novel caries detection systems based on D3 lesions, in vitro, on occlusal surfaces, after Pretty 57 The evidence supporting each of the systems is currently limited; often by virtue of the in vitro nature of the studies, or due to a failure of standardisation of approach to study design making meta-analyses impossible. However, if we can state with some confidence that the systems do permit earlier detection of enamel lesions, and systems such as QLF and DiFOTI enable images of these lesions to be stored and viewed at a later date. It is worthwhile considering what the purpose may be for supplementing or even replacing well established dental diagnostic systems; they must offer improved diagnostic efficiency, better patient care pathways or perhaps comply with legislative changes. There is a paradigm shifty occurring in dentistry; we are slowly moving away from a surgical model into one more medically based. The devices described within this paper, by enabling early detection of caries enable the remineralising therapies to be correctly prescribed and for their success to be measured. The advent of dental auxiliaries who can undertake an increasing number of procedures emphasises the important role that the dentist plays as the leader of the dental team. This leadership role is critically tied to the fact that dental clinicians retain the sole right of diagnosis and thus the devices and approaches described within the current paper serve only to augment the diagnostic skills of the clinician. Making the right decision about the presence or absence of a lesion, its degree of severity and its likely activity combined with the sociobehavioural aspects of the patient, their risk and modifying factors will continue to rest with the dentist. references 1. Holt RD. Advances in dental public health. Primary Dental Care 2001;8(3): Featherstone JD. The caries balance: the basis for caries management by risk assessment. Oral Health & Preventive Dentistry 2004;2(Suppl. 1): al-khateeb S, Oliveby A, de Josselin de Jong E, Angmar- Mansson B. Laser fluorescence quantification of remineralisation in situ of incipient enamel lesions: influence of fluoride supplements. Caries Research 1997;31(2): Amaechi BT, Higham SM. In vitro remineralisation of eroded enamel lesions by saliva. Journal of Dentistry 2001;29(5): Kidd EA. The operative management of caries. Dental Update 1998;25(3): Maupome G, Pretty IA. A closer look at diagnosis in clinical dental practice. Part 4. Effectiveness of nonradiographic diagnostic procedures and devices in dental practice. Journal of the Canadian Dental Association 2004;70(7): Pitts NB. Clinical diagnosis of dental caries: a European perspective. Journal of Dental Education 2001;65(10): Pitts NB. Diagnostic tools and measurements impact on appropriate care. Community Dentistry and Oral Epidemiology 1997;25(1): Pretty IA, Maupome G. A closer look at diagnosis in clinical dental practice. Part 2. Using predictive values and receiver operating characteristics in assessing diagnostic accuracy. Journal of the Canadian Dental Association 2004;70(5): Pretty IA, Maupome G. A closer look at diagnosis in clinical dental practice. Part 1. Reliability, validity, specificity and sensitivity of diagnostic procedures. Journal of the Canadian Dental Association 2004;70(4):251 5.

12 738 journal of dentistry 34 (2006) Verdonschot EH, Angmar-Mansson B. Advanced methods of caries diagnosis and quantification. In: Fejerskov O, Kidd E, editors. Dental Caries. The disease and its clinical management. Oxford: Blackwell Munksgaard; Bader JD, Shugars DA, Bonito AJ. Systematic reviews of selected dental caries diagnostic and management methods. Journal of Dental Education 2001;65(10): Ekstrand KR, Ricketts DN, Kidd EA, Qvist V, Schou S. Detection, diagnosing, monitoring and logical treatment of occlusal caries in relation to lesion activity and severity: an in vivo examination with histological validation. Caries Research 1998;32(4): Verdonschot EH, Rondel P, Huysmans MC. Validity of electrical conductance measurements in evaluating the marginal integrity of sealant restorations. Caries Research 1995;29(2): Longbottom C, Huysmans MC. Electrical measurements for use in caries clinical trials. Journal of Dental Research 2004:83. Spec no. C:C Huysmans MC, Longbottom C, Christie AM, Bruce PG, Shellis RP. Temperature dependence of the electrical resistance of sound and carious teeth. Journal of Dental Research 2000;79(7): Wang J, Sakuma S, Yoshihara A, Kobayashi S, Miyazaki H. An evaluation and comparison of visual inspection. Electrical caries monitor and caries detector dye methods in detecting early occlusal caries in vitro study. Journal of Dental Health 2000;50: Huysmans MC. Electrical measurements for early caries detection. In: Stookey GK, editor. Early Caries Detection II. Proceedings of the fourth annual Indiana conference Huysmans MC, Longbottom C, Hintze H, Verdonschot EH. Surface-specific electrical occlusal caries diagnosis: reproducibility, correlation with histological lesion depth, and tooth type dependence. Caries Research 1998;32(5): Ekstrand KR, Ricketts DN, Kidd EA. Reproducibility and accuracy of three methods for assessment of demineralization depth of the occlusal surface: an in vitro examination. Caries Research 1997;31(3): Baysan A, Lynch E, Ellwood R, Davies R, Petersson L, Borsboom P. Reversal of primary root caries using dentifrices containing 5,000 and 1,100 ppm fluoride. Caries Research 2001;35(1): Huysmans MC, Longbottom C, Pitts NB, Los P, Bruce PG. Impedance spectroscopy of teeth with and without approximal caries lesions an in vitro study. Journal of Dental Research 1996;75(11): Verdonschot EH, Kuijpers JM, Polder BJ, De Leng-Worm MH, Bronkhorst EM. Effects of digital grey-scale modification on the diagnosis of small approximal carious lesions. Journal of Dentistry 1992;20(1): Verdonschot EH, Angmar-Mansson B, ten Bosch JJ, Deery CH, Huysmans MC, Pitts NB, et al. Developments in caries diagnosis and their relationship to treatment decisions and quality of care. ORCA Saturday Afternoon Symposium Caries Research 1999;33(1): White SC, Yoon DC, Tetradis S. Digital radiography in dentistry: what it should do for you. Journal of California Dental Association 1999;27(12): Wenzel A, Pitts N, Verdonschot EH, Kalsbeek H. Developments in radiographic caries diagnosis. Journal of Dentistry 1993;21(3): Ellwood RP, Davies RM, Worthington HV. Evaluation of a dental subtraction radiography system. Journal of Periodontal Research 1997;32(2): Choksi SK, Brady JM, Dang DH, Rao MS. Detecting approximal dental caries with transillumination: a clinical evaluation. Journal of American Dental Association 1994;125(8): Cortes DF, Ellwood RP, Ekstrand KR. An in vitro comparison of a combined FOTI/visual examination of occlusal caries with other caries diagnostic methods and the effect of stain on their diagnostic performance. Caries Research 2003;37(1): Bader JD, Shugars DA, Bonito AJ. A systematic review of the performance of methods for identifying carious lesions. Journal of Public Health Dentistry 2002;62(4): Mitropoulos CM. A comparison of fibre-optic transillumination with bitewing radiographs. British Dental Journal 1985;159(1): Pitts N. ICDAS an international system for caries detection and assessment being developed to facilitate caries epidemiology, research and appropriate clinical management. Community Dental Health 2004;21(3): Davies GM, Worthington HV, Clarkson JE, Thomas P, Davies RM. The use of fibre-optic transillumination in general dental practice. British Dental Journal 2001;191(3): Schneiderman A, Elbaum M, Shultz T, Keem S, Greenebaum M, Driller J. Assessment of dental caries with digital imaging fiber-optic transillumination (DIFOTI): in vitro study. Caries Research 1997;31(2): de Josselin de Jong E, Sundstrom F, Westerling H, Tranaeus S, ten Bosch JJ, Angmar-Mansson B. A new method for in vivo quantification of changes in initial enamel caries with laser fluorescence. Caries Research 1995;29(1): Ando M, Hall AF, Eckert GJ, Schemehorn BR, Analoui M, Stookey GK. Relative ability of laser fluorescence techniques to quantitate early mineral loss in vitro. Caries Research 1997;31(2): van der Veen MH, de Josselin de Jong E. Application of quantitative light-induced fluorescence for assessing early caries lesions. Monography in Oral Science 2000;17: Angmar-Mansson B, ten Bosch JJ. Quantitative light-induced fluorescence (QLF): a method for assessment of incipient caries lesions. Dentomaxillofac Radiology 2001;30(6): Shi XQ, Tranaeus S, Angmar-Mansson B. Validation of DIAGNOdent for quantification of smooth-surface caries: an in vitro study. Acta Odontologica Scandinavica 2001;59(2): Lussi A, Imwinkelried S, Pitts N, Longbottom C, Reich E. Performance and reproducibility of a laser fluorescence system for detection of occlusal caries in vitro. Caries Research 1999;33(4): Shi XQ, Welander U, Angmar-Mansson B. Occlusal caries detection with KaVo DIAGNOdent and radiography: an in vitro comparison. Caries Research 2000;34(2): Bader JD, Shugars DA. A systematic review of the performance of a laser fluorescence device for detecting caries. Journal of American Dental Association 2004;135(10): Ngaotheppitak P, Darling CL, Fried D. Measurement of the severity of natural smooth surface (interproximal) caries lesions with polarization sensitive optical coherence tomography. Lasers in Surgery and Medicine 2005;37(1): Amaechi BT, Podoleanu AG, Komarov G, Higham SM, Jackson DA. Quantification of root caries using optical coherence tomography and microradiography: a correlational study. Oral Health & Preventive Dentistry 2004;2(4): Hall A, Girkin JM. A review of potential new diagnostic modalities for caries lesions. Journal of Dental Research 2004:83. Spec no. C:C Fried D, Featherstone JD, Darling CL, Jones RS, Ngaotheppitak P, Buhler CM. Early caries imaging and monitoring with near-infrared light. Dental Clinics of North America 2005;49(4): vi. 47. Bab I, Fuerstein O, Gazit D. Ultrasonic detector of proximal caries. Caries Research 1997;31:322. (Abstract).

13 journal of dentistry 34 (2006) Ziv V, Gazit D, Beris D, Fuerstein O, Aharnonov L, Bab I. Correlative ultrasonic histologic and Roentgenographic assessment of approximal caries. Caries Research 1998;32(2):294. (Abstract). 49. Ng SY, Ferguson MW, Payne PA, Slater P. Ultrasonic studies of unblemished and artificially demineralized enamel in extracted human teeth: a new method for detecting early caries. Journal of Dentistry 1988;16(5): Bab I, Ziv V, Gazit D, Fuerstein O, Findler M, Barak S. Diagnosis of approximal caries in adult patients using ultrasonic waves. Journal of Dental Research 1998;77:255. (Abstract). 51. Angmar-Mansson BE, al-khateeb S, Tranaeus S. Caries diagnosis. Journal of Dental Education 1998;62(10): Rock WP, Kidd EA. The electronic detection of demineralisation in occlusal fissures. British Dental Journal 1988;164(8): Ricketts DN, Kidd EA, Wilson RF. Electronic diagnosis of occlusal caries in vitro: adaptation of the technique for epidemiological purposes. Community Dentistry and Oral Epidemiology 1997;25(3): Verdonschot EH, Wenzel A, Truin GJ, Konig KG. Performance of electrical resistance measurements adjunct to visual inspection in the early diagnosis of occlusal caries. Journal of Dentistry 1993;21(6): Pereira AC, Huysmans MC, Verdonschot EH. Occlusal caries diagnosis using the DiagnoDent: in vitro reproducibility and comparison with electrical and visual diagnosis. Journal of Dental Research Spec A: 233 (Abstract). 56. Mount GJ, Hume WR. Preservation and restoration of tooth structure. 2nd ed. Queensland: Knowledge Books and Software; Pretty IA. A review of the effectiveness of QLF to detect early caries lesions. Indianapolis, Indiana: Indiana University Press; 2005.