The Effect of Combining Radiographs and DIAGNOdent With Visual Examination on Detection and Treatment Decisions of Noncavitated Occluso-dentinal Caries
The aim of this laboratory study was to evaluate the effectiveness of incorporating radiographic examination and DIAGNOdent with visual examination for the detection and treatment of noncavitated occluso-dentinal caries. Four examiners examined the occlusal surfaces of 160 extracted posterior teeth. Teeth were examined three times with a one-month interval in between. The first examination was visual (V), the second examination was visual with radiograph (VR), and the third examination was visual with radiograph and DIAGNOdent (VRD). Examiners were asked to detect the presence of caries (if any) and identify the extent of caries (if present; ie, enamel or dentin). The examiners were also asked to choose a treatment for each tooth. The examined teeth were later sectioned, and the presence of caries was charted as 0 = no caries, 1 = caries confined to enamel, 2 = caries in the outer dentin, and 3 = caries in the inner dentin. Sensitivity, specificity, area under the curve (Az values), rank correlation coefficient, interexaminer reproducibility, and McNemar χ2 tests were calculated. VR had statistically higher specificity and lower sensitivity than V and VRD. The means of Az values ranged from 0.58 to 0.63, with no statistical difference between the three examinations (p>0.05). The means of the rank correlation coefficients with histology for detection of dentinal caries were 0.48, 0.23, and 0.44 using V, VR, and VRD, respectively. Interexaminer reproducibility was best for V alone. The percentages of teeth indicated for restorative treatment were 31%, 33%, and 41% using V, VR, and VRD, respectively. The percentages of teeth correctly treated according to histology were 69.4%, 70.0%, and 67.5% using V, VR, and VRD, respectively. There was no difference in the accuracy of treatment decisions between the three examination points (p>0.05). The addition of radiographs and DIAGNOdent to visual examination did not have a significant effect on the improvement of the detection or treatment decisions of examiners of noncavitated occluso-dentinal carious lesions.SUMMARY
INTRODUCTION
Despite decreasing caries prevalence worldwide, occlusal caries remain the predominant form of dental caries at the present time.1 This has been mainly attributed to the anatomy of the fissure system and the difficulty of plaque removal in this area, especially during the stage of tooth eruption.2,3
Thirty-four percent of all dental patients have been found to have questionable occlusal caries that are difficult to diagnose.4 The early stages of occlusal caries are difficult to detect clinically, but later occlusal caries might also present as “questionable caries” where there is no frank cavitation but the surface is rough with opacities or staining.5 Histologically, these lesions might be confined to enamel or might have spread into dentin.6
To improve the frequency of detection of occlusal caries lesions, combining visual examination with other diagnostic aids has been recommended.7-10
Although radiographic examination has limited benefits when it comes to initial or early occlusal carious lesions, it is the most widely used technique for improving caries detection and the only technique available in most dental clinics.7,11 It has been shown that only one-third of occlusal dentinal lesions were detected clinically, while two-thirds were found clinically on bitewing radiographs as hidden caries.12 In addition, the presence of dentinal lesions on conventional bitewing radiographs has been found to be significantly associated with infected dentin, which is soft and wet clinically.13
A laser-stimulated fluorescence caries detection system (DIAGNOdent) has been introduced in an attempt to quantify carious lesions and better improve clinicians' diagnostic abilities.14 Clinically, DIAGNOdent has shown higher sensitivity and lower specificity at the D3 threshold (dentinal caries) compared with visual inspection and radiographs.15 As a result, DIAGNOdent has been recommended for use only as an adjunct to visual examination since it has been reported to yield more false-positive results than visual and radiographic methods.15,16
Caries detection has a tremendous impact on treatment decisions and is directly linked to the patient's dental management and treatment outcomes.17,18 Treatment and management of carious lesions should ideally be based on severity and activity.6,19 Unfortunately, recent studies have shown that initial carious lesions are still managed restoratively.19,20
Considering the fact that previous studies have focused on the accuracy of radiographs and DIAGNOdent on detecting occlusal carious lesions separately, the aims of this study were 1) to investigate whether combining radiographs and DIAGNOdent with visual examination would improve the detection of noncavitated occlusal caries, 2) to investigate whether combining radiographs and DIAGNOdent with visual examination would influence the treatment decisions for noncavitated occlusal carious lesions, and 3) to investigate the influence of combining radiographs and DIAGNOdent with visual examination on interexaminer reproducibility for the detection of noncavitated occluso-dentinal caries.
METHODS AND MATERIALS
Teeth
After preliminary visual examination, 160 posterior teeth (premolars and molars) were selected. The included teeth were free of restorations and had no proximal caries. Teeth with dental fluorosis, tetracycline staining, and hypoplasia were also excluded. The clinical appearance of the occlusal surface of each tooth ranged from sound to discolored, with white/brown discoloration and no cavitation. The teeth were cleaned with a prophylaxis brush using pumice slurry and rinsed thoroughly with a three-way syringe and then stored in water. Later, each tooth was embedded in an acrylic block and was randomly assigned a number from 1 to 160.
Examiners
Four examiners participated in the study. The examiners were university faculty with the following specialties: general dentistry, operative dentistry, pediatric dentistry, and prosthetic dentistry. The following clinical scenario was given to the examiners: “A healthy 18 year-old patient presents to your clinic for her/his scheduled dental appointment. Complete your clinical examination for the displayed tooth. How would you manage this tooth if the patient is at high risk for developing caries? The cost of treatment should not be a factor in your decision.” A pilot study of 25 teeth (not included in the study) was carried out. Teeth were examined three times with a one-month separation period. The first examination was visual only, the second was visual with radiographs, and the third was a combination of visual, radiographs, and DIAGNOdent as described below. In each of the examination sessions, the teeth were presented randomly to the examiners.
During each examination, participants were asked to decide on the diagnosis of the tooth using the following criteria: 0 = the tooth is sound, 1 = the tooth is carious but the caries is confined to enamel, 2 = the tooth is carious and the caries is extending into dentin. The examiners were then asked to choose one of the following treatment options for the occlusal surface of the tooth: 0 = no treatment, 1 = noninvasive treatment (fluoride application or fissure sealant), 2 = preventive resin restoration (conservative cavity preparation to remove the caries, followed by resin composite restoration of the cavity and fissure sealant of the remaining fissures), 3 = resin composite/amalgam restoration.
Visual Examination
Visual examination was carried out using a dental operating light, a blunt-end periodontal probe, and a 3-in-1 syringe. No explorer was used during the examination. The examination was done by all examiners in the same setting and at the same time. The examination time was almost two hours for all of the teeth. Air drying for each tooth was allowed for up to five seconds.
Visual and Radiographic Examination
After one month from the first examination, the examiners were presented with the teeth and their corresponding radiographs and were asked to propose a diagnosis and management for each tooth using the previously mentioned criteria. The radiographs were examined on a backlit screen with no magnification. Conventional radiographic imaging was obtained for each tooth separately using the paralleling technique, with Kodak Insight film of 30.5 mm × 40.5 mm (Kodak, Rochester, MD, USA) using a Planmeca ProX Intraoral X-ray Generator Unit (Planmeca Oy, Helsinki, Finland), which was operated at 63 kVp and 8 mA and an exposure time of 0.10 seconds. The film was then developed in a Dürr Dental XR 24 Pro automatic developing machine (Dürr Dental AG, Bietigheim-Bissingen, Germany), with a pass time of seven minutes.
Visual, Radiographic, and Laser Fluorescence Examination
The third examination was conducted a month after the completion of the second examination. In this case, the teeth were presented to the examiners with their corresponding radiographs and a laser fluorescence device (DIAGNOdent, Kavo, Biberach, Germany). The devices were initially calibrated before use against the ceramic reference as recommended by the manufacturer. Each tooth was then dried for five seconds, and the probe tip designated for occlusal surface was adjusted to each tooth separately by holding the tip against a sound smooth surface and pressing the ring cuff until calibration was complete. The probe tip was positioned over the occlusal fissure without applying pressure and pivoted in all directions. The tooth was scored based on the highest reading obtained. Each examiner was provided with the manufacturer's instructions on how to use the pen and the interpretation of the resulting reading. Using the information from the visual examination, the radiograph, and the DIAGNOdent, the examiners were asked to choose a diagnosis for the occlusal surface of the tooth and to choose a management option following the previously mentioned criteria.
Histological Validation
The teeth were sectioned through a predetermined point of interest using a water-cooled 0.3-mm diamond saw at low speed (Isomet low speed saw; Buehler, Lake Bluff, IL, USA). Teeth with more than one point of interest were sectioned more than once. Computerized images were obtained for each section with a digital camera (Leica DC200; Leica Microsystems Wetzlar, GmbH, Wetzlar, Germany) linked to a stereomicroscope (Leica MZ6, Leica Microsystems Wetzlar, GmbH), using a fixed distance and the same magnification (10×). The microscope was linked to a computer with a digital image analysis system (IM 500; Leica Microsystems Wetzlar, GmbH). Two examiners (the pediatric and restorative dentists) assessed the histological images using the following four-point scale (Downer): 0 = sound, 1 = caries in the outer half of the enamel, 2 = caries in the inner half of the enamel reaching up to the enamel-dentinal junction, 3 = caries in the outer half of the dentin, and 4 = caries in the inner half of the dentin.21 The highest score from the various sections was used to define the caries status of the surface. Calibration of both examiners was undertaken by reading 25 pilot samples. In cases in which the two investigators' ratings diverged, a joint assessment was performed until consensus was reached.
Statistical Analysis
The data were analyzed statistically with SPSS software, version 17.0 (SPSS Inc, Chicago, IL, USA). Descriptive exploratory analysis was done. The histological diagnosis for each tooth was dichotomized into two categories: sound and caries limited to enamel as one group and caries extending into dentin (D3) as a second group. Sensitivity, specificity, and the area under the receiver operating characteristics curve were also calculated and reported as Az values for the three examinations and for the four observers using the histological criteria at D3 (caries into dentin). The strength of the association between each examination and the histology scores was assessed using a rank correlation coefficient (Kendall's tau b). Comparisons among the three examination points were conducted using the McNemar χ2 test. The quality of the interexaminer repeatability was calculated using Cohen's weighted kappa test.22 With regard to treatment decisions, only caries extending into dentin (D3) were considered in need of operative treatment. The treatment options were also regrouped into preventive (no treatment, fluoride application, or fissure sealant) and restorative (preventive resin restoration or class I cavity preparation and restoration). The level of significance was set at α=0.05.
RESULTS
Histological examination of the 160 teeth revealed that 61 teeth had caries extending into dentin, while 99 teeth were either sound or had caries confined to enamel (Table 1). The McNemar χ2 test showed that there was a statistically significant difference between the three examination points and for the four examiners (except for the pediatric dentist between visual alone and visual combined with radiograph and DIAGNOdent). Table 2 presents the sensitivity, specificity, Az values, and the correlation coefficients for the visual examination alone, visual examination with radiographs, and visual examination with radiographs and DIAGNOdent for the four examiners for the detection of occlusal caries at the D3 diagnostic threshold. There was no statistical difference between the mean sensitivity of visual examination alone and visual examination combined with both radiograph and DIAGNOdent (p>0.05). The mean sensitivity of visual examination with radiograph was significantly lower than the mean sensitivity of both visual examination alone and visual examination with radiograph and DIAGNOdent (p<0.05). The mean specificity of visual examination with radiographs was significantly higher than that of visual examination alone or visual examination combined with radiographs and DIAGNOdent (p<0.05). There was no statistically significant difference between the Az values of the three examination points (p>0.05). However, the rank correlation with histology was lowest for visual examination combined with radiographs.
Table 3 presents the weighted Kappa values of interexaminer reproducibility for each examination point. For visual examination alone, Kappa values ranged from moderate to substantial (0.44-0.80), while for visual with radiographs and visual with radiographs and DIAGNOdent, the ranges were from fair to moderate (0.26-0.44 and 0.30-0.62, respectively).22
Table 4 shows the number (%) and type of treatments indicated by each examiner for management of occlusal surfaces, based on visual examination alone, visual examination with radiographs, or visual examination combined with radiographs and DIAGNOdent. The addition of radiographs to visual examination resulted in a decrease in the number of teeth indicated for restoration for the general dentist (p<0.001) and an increase in the same for the restorative dentist (p<0.001). There was no effect of adding the radiographs on the number of teeth indicated for restoration for the prosthodontist and the pediatric dentist (p>0.05). There was an increase in the number of teeth indicated for restoration when visual examination was combined with radiographs and DIAGNOdent for two of the examiners (general and restorative dentists; p<0.05), while there was no effect for the other two examiners (prosthodontist and pediatric dentist; p>0.05). The table also shows considerable variability among the four examiners in their treatment decisions regarding noncavitated occlusal caries.
Table 5 shows the relationship between the presence of caries, confirmed by the histological examination of the teeth, and the treatment decisions based on visual examination alone, visual examination with radiographs, and visual examination combined with both radiographs and DIAGNOdent. The number of correct treatment decisions was not different between the three examination points and for the four examiners (p>0.05). There was an increase in the number of overtreatment decisions when the visual examination was combined with both radiographs and DIAGNOdent for two of the examiners (general and restorative dentists; p<0.05).
DISCUSSION
In this study, we tested the hypothesis that the addition of radiographs and DIAGNOdent to visual examination will improve the detection and treatment decisions of noncavitated occlusal dentinal caries. The accuracy of detection and treatment of noncavitated occlusal dentinal caries was not improved by the addition of radiographs alone or the addition of radiographs with DIAGNOdent to the visual examination. Clinical visual examination can be used alone for the detection of noncavitated occlusal dentinal caries.
The study attempted to replicate a clinical setting: examiners were asked to use only visual examination in the first examination and to use both clinical examination and radiograph in the second examination. In the third examination, the examiners were requested to combine the visual examination with radiographs and DIAGNOdent readings; the three techniques were not used separately. Each of the four examiners was given the tooth, its radiograph, and a DIAGNOdent machine and was asked to use his or her clinical judgment to reach a diagnosis and treatment for the tooth. Previously, it has been recommended to consider the overall risk profile of the patient and site before using the cutoff points of the DIAGNOdent as the main source of information.23 Deciding on the type of restorative treatment for a specific lesion depends on a range of variables that include patient history, fluoride and dietary status, caries risk assessment, and the status of the tooth surface. In trying to eliminate the confounding effect of these factors on the decision of the investigators, a specific clinical scenario was created with a healthy patient who is at a high risk for caries and to whom the cost of the treatment was not applicable.
Four examiners with different specialties were the participants in this study. Previously, it has been found that some specialties tend to be more aggressive when treating occlusal caries than others.20 The results of this study also showed that the specialists of general and operative dentistry tended to overtreat occlusal caries more than the other two specialties (pediatrics and prosthodontics).
The sensitivity of visual examination alone was high, and the specificity was moderate. The area under the curve for visual examination alone was higher than the other two methods, and the rank correlation with histology was also higher than the other two methods. Previous studies have reported moderate sensitivity and substantial specificity for visual examination of dentinal occlusal caries.8,24-26 Differences between studies can be due to several factors, such as the degree of caries in the included teeth and the experience of the examiners. In this study, the four examiners are university teachers from different specialties, and all had received training on the International Caries Detection and Assessment System.
Sensitivity values of visual examination with radiographs in the study were lower than values for visual examination alone or visual examination combined with radiographs and DIAGNOdent. This is in agreement with what has been reported in the literature.24,27,28 Radiographs are two-dimensional images of a three-dimensional tooth, making the detection of early lesions difficult since they can be overlapped with sound tooth structure. In agreement with what has been reported previously in the literature, radiographs with visual examination had higher specificity than visual examination alone or visual examination combined with both radiographs and DIAGNOdent.25,27 The mean rank correlation coefficient of the visual examination with radiographs was 0.23, which was lower than the same for both the visual examination alone (0.48) or visual examination combined with radiographs and DIAGNOdent (0.44). A study by Toraman and others29 reported similar findings, while other studies reported higher values.24,27 Variability between studies might be attributed to differences in the extension of the carious lesions included. In this study, we had only 3.8% of our sample with caries in the inner dentin, while the remaining teeth were either sound, containing caries confined into enamel, or caries confined to the outer third of dentin. In addition, previous studies have reported that the DIAGNOdent can have a wide range of inter- and intraexaminer variability.10
In a review on the performance of DIAGNOdent in the detection of dentinal caries, it has been observed that there is wide variability in the literature in the reported sensitivity and specificity of this machine.14 In general, the DIAGNOdent usually exhibits higher sensitivity and lower specificity values than visual assessment. In our study, the same trend was observed, but the difference was not statistically significant between visual assessment alone and visual assessment combined with radiographs and DIAGNOdent. This may have been because examiners were asked not to depend on the DIAGNOdent readings alone but to use their clinical judgment as well when deciding on diagnoses and treatment. In addition, it has been reported that the performance of DIAGNOdent depends on the cutoffs used, and values obtained with the DIAGNOdent device are not the same as the values obtained with the DIAGNOdent pen.27 In this study, it was found that visual examination combined with radiographs and DIAGNOdent had high sensitivity and low specificity values, as has been reported.28 Furthermore, the correlation of the visual examination alone with the histology was found to be similar to that of the visual examination combined with both radiographs and DIAGNOdent, as has been reported in the literature.29 Other studies found the correlation between DIAGNOdent and histology to be better than for visual and radiographs.27,30 In previous reported studies, each diagnostic method was used separately, but in this study, the diagnostic methods were combined to mimic the clinical scenario, which may have had a different effect on results obtained.
There were no statistically significant differences in the Az values between the three examination points in detection of occlusal caries at the D3 level. Previous studies have reported DIAGNOdent to have higher Az values compared with visual and radiographs.27,28 Again, the reason for this difference might be attributed to the fact that the three diagnostic methods were combined.
The interexaminer reproducibility was moderate to substantial for visual examination alone, while it was fair to moderate for the visual examination combined with radiographs and DIAGNOdent, as has been previously reported.28 Other studies found similar or higher interexaminer reproducibility for DIAGNOdent compared with visual examination.26,29 Such differences between studies can be due to the level of caries included in the study, in which some studies were done at the D1 level while others were done at the D3 level. The level of experience of the examiners and the number and condition of teeth included in the study can also have an impact.23
This laboratory study supports many previous studies in the fact that there is still a wide variability among dentists in diagnosis and treatment decisions of initial occlusal caries.18,20,31,32 The addition of radiographs and DIAGNOdent to the visual examination did not improve the examiners' performance in detection and treatment decisions of noncavitated occluso-dentinal caries. Recently, similar results have been found in a clinical study on primary teeth, in which simultaneous combined strategies increased sensitivities but decreased specificities.33 Furthermore, no differences were observed in the accuracy of diagnosis or treatment.
The addition of both radiographs and DIAGNOdent to the visual examination resulted in more teeth treated restoratively (mainly as preventive resin restorations). Similar results have been reported in the literature.7 The addition of radiographs and DIAGNOdent did not increase the accuracy or the percentage of correctly treated teeth; conversely, it resulted in an increase in the percentage of overtreated teeth. The increase in the false-positive diagnosis, meaning some teeth will be treated without having the disease, has led others to conclude that DIAGNOdent should not be used as the primary source for diagnosis.11,14,24,32
This in vitro study has demonstrated that the use of radiographs and DIAGNOdent pen as adjuncts to visual examination adds little benefit in detection or treatment decisions of noncavitated occlusal carious lesions in comparison with the visual examination being carried out alone.28,34-39
CONCLUSIONS
-
Visual examination with radiograph had higher specificity and lower sensitivity than visual examination alone or visual examination combined with radiographs and DIAGNOdent.
-
There was no difference in the Az values between visual examination alone, visual examination with radiographs, or visual examination combined with radiographs and DIAGNOdent.
-
Visual examination with radiographs had the lowest rank correlation with histology.
-
There was an increase in the number of teeth indicated for treatment when the examiners used radiographs and DIAGNOdent with the visual examination.
-
The number of correct treatment decisions was not different between the three examination points.
Contributor Notes