Evaluation of Diagnodent Accuracy in Detecting Approximal Caries in Primary Molars

Introduction

Dentists usually rely on visual, tactile and radiographic methods to detect dental caries¹. Caries in inter-dental surfaces are frequently detected by bite-wing radiography^2-4, but this method has got some limitations. Radiographic images cannot reliably show the real depth of the lesions and are not appropriate for detection of the initial stages of enamel caries. Exposure to X-ray is another concern with radiography (2).

Visual method has shown a high specificity, though low sensitivity and reproducibility for detection of proximal caries compared to other methods (1,2,5).

International Caries Detection and Assessment System (ICDAS) is introduced for standardization and improvement of dental caries detection by visual method⁶. This system is expected to increase the sensitivity and reliability of this method. Some studies on detection of occlusal caries have found this international system highly valuable during recording of results by visual method^7,8; although using ICDAS index for diagnosis of approximal caries by visual method is not valuable (2,5).

Recently, the need for more conservative treatments and procedures to suppress and inverse the decay-producing process has led to considerable efforts to develop technologies for early detection of dental caries. One of the new technologies to achieve this goal is using pen-type laser fluorescence device (LF-pen). This device provides a quantitative method for detection of occlusal and approximal caries (5,9,10).

Novaes et al. assessed bitewing radiography and LF-pen and showed a similar efficacy for detection of approximal caries, but Lussi et al. found laser fluorescence to be superior to radiography in detection of proximal caries in permanent teeth.

Most studies on detection of caries by LF-pen have assessed occlusal caries (11, 14).There are few in vivo studies on detection of approximal caries especially in primary molars (2, 16), on the other hand in vitro reconstruction of contact areas in posterior teeth is not a precise method, so we decided to compare different methods of caries detection, i.e. visual, bitewing radiography and laser fluorescence for detection of approximal caries in primary molars in an in vivo study.

Materials and methods

The research protocol was approved by the Ethical Committee of Yazd medical university (protocol≠ p/17/1/21967). This in vivo diagnostic cross sectional study was performed on 36 children referred to the pediatric dentistry department of Shahid Sadoughi University of Medical Sciences, Yazd, Iran. Patients were referred for dental examinations or therapeutic procedures. After primary examinations, children 5-10 year-old who fulfilled the inclusion criteria (according to the forms filled by the parents) entered the study. An informed consent was obtained from the parents. Examinations were performed on approximal surfaces of the primary molars.

Those with the following conditions were excluded from the study: approximal restorations, hypoplastic pits, extensive approximal caries completely destroying the marginal ridge, extensive caries on smooth or occlusal surfaces, and lack of adjacent tooth.          

Eventually, 229 approximal surfaces were assessed. Caries lesions on approximal surfaces were assessed by visual, radiographic and laser fluorescence methods. All evaluations were performed by two dentists. Each examiner separately assessed the surfaces and recorded the results. Both examiners were blinded to the results of each other. Before beginning the study, aforementioned diagnostic methods were performed on two children as pilot samples and these two children did not include in the study.

Before examinations, approximal surfaces were cleaned by dental floss, prophylactic paste, and rubber cup. Initially in order to assess the visual method, approximal surfaces were evaluated after cleansing under sufficient illumination. Each surface was first assessed when it was wet and then it was dried by air pressure and assessed again.

ICDAS index was used for recording the visual results [Ismail et al., 2007].

For radiographic evaluation, bitewing radiographs were taken from dental surfaces including maxillary and mandibular primary molars (device was set on 70Kv, 80mA and 0.3 exposure time), using 22×35mm Kodak films and bitewing film preservative (XCP).

After radiography, caries in each surface were recorded using the criteria previously described [Ekstrand et al., 1997].

Surfaces were assessed by laser fluorescence LF- pen (DIAGNOdent, Kavo, Biberach, Germany) as well. A tip for approximal surfaces was used for evaluation. Laser was first calibrated on the porcelain reference of the device and then on the healthy teeth. Afterwards, the tooth contact area was dried by air for 5 seconds and DIAGNOdent tip was placed first near proximal area of buccal and the lingual surfaces. The score was read each time and the largest score was recorded by each examiner.

After recording the results of three diagnostic methods, in order to use a standard reference method, it was necessary to separate the contact area of primary molars so as the teeth be separated at least 0.5 – 1 mm from each other. In this condition, it is possible to accurately define the amount of caries in approximal surfaces. So we used orthodontic separators between primary molars for 7 days. After this period, the contact area was cleaned by dental floss and was examined again using a mirror and a probe. The results of the evaluations were categorized in 3 groups:

0-       Intact and healthy surface: without any change in enamel lucency and absence of surface discontinuity after air drying.

1-       Presence of white spot lesions: White or brown discoloration in dried or wet surfaces without surface discontinuity.

2-       Cavitation: loss of surface integrity clearly observed or observed after using a probe.

When there was inconsistency between two examiners, surfaces were assessed once more to reach to a common opinion.

Proximal surfaces were considered as statistical units which were assessed for comparison between different diagnostic methods.

Sensitivity, specificity and accuracy were calculated for each method and each examiner separately using defined cut-off points.

Receiver operating characteristic (ROC) curves were drawn for data extracted from LF-pen diagnostic method and the area under the ROC curve (A-z) was calculated. Then agreement coefficient for radiography and visual method was measured and inter-examiner reliability was evaluated by using intra-class correlation coefficient.

Results

After assessment of data from the standard method, 87 surfaces (38%) were sound, but 84 (36.7%) and 58 (25.3%) showed white-spot lesions and cavitation, respectively.

Table 1 compares the results of visual method with the standard method.

In table 2, results of radiographic method are compared with standard method.

All surfaces showing radiolucency in the middle third of dentin (score 3) were found to have cavity.

The best cut-off point for LF-pen device in this study included: Sound surfaces: 0-5, white-spot: 6-15, cavitation: ≥16.   

The Az value for laser method was 0.97 and 0.98 for the first and second examiners, respectively, which shows a high efficiency for this method. High ICC value (0.99) in LF-pen method and high agreement coefficient (0.96 for radiography and 0.90 for visual method) showed a high inter-examiner reliability for these three methods.  As table 3 shows, in white-spot category, the sensitivity of LF-pen device by both examiners was significantly higher than other methods.

The sensitivity of visual method was lower than other two methods, although its specificity in white-spot category was higher than other methods. Totally, all three methods in this category have got a high specificity (tables 3). In the cavitation category, the specificity of radiography was significantly higher than LF-pen.

The highest accuracy for detection of caries in approximal surfaces was observed in LF-pen, radiography and visual methods, respectively.

Table 1. Frequency of visual scores by ICDASII index and results from standard method, by two examiners.

Table 2. Frequency of radiographic scores and results from standard method, by twoxaminers

Table3. Sensitivity, specificity and accuracy of caries detectio by each examiner (1 and 2) for all methods (visual inspection, DIAGNOdent and radiography)

*the numbers inside the brackets are calculated with a 95% confidence coefficient