Near infrared transillumination compared with radiography to detect and monitor proximal caries: A clinical retrospective study

Abstract

Objective

To compare near infrared transillumination device, DIAGNOcam (DC) and bitewing radiography (BW) for the detection of proximal caries.

Materials and methods

This retrospective analysis of DC and BW images of 18 students in dental medicine who had consented to the anonymous use of their dental record. The data included BW and DC images performed for a check-up in 2013, and corresponding follow-up images performed in 2015. Two observers rated 376 proximal surfaces on a 4-level dentin lesion scale and reached a unanimous rating for each surface. Calculated measures of agreement for each assessment method over time provided the reproducibility of the information obtained by each method.

Results

Agreement between 2013 and 2015 within each method was excellent (intraclass correlation coefficient, BW: 0.86, DC: 0.90). Agreement between DC and BW was similar for dentin lesion detection, but was low for enamel caries detection; DC detected more enamel caries than BW. Agreement between DC and BW was modest (0.33 in 2013 and 0.36 in 2015), chiefly because DC identified more enamel caries.

Conclusion

This study shows that DC is as reliable as BW to detect proximal dentin lesions. DC detects proximal enamel lesions at an earlier stage than BW. DC enables clinicians to differentiate lesions limited to the enamel from lesions that have reached the enamel dentin junction. Regular monitoring with DC should help provide individualized preventive measures and early non-invasive caries management.

Clinical significance

The early detection of enamel lesions with near infrared transillumination can help clinicians undertake early non invasive treatments to prevent or slow down the progression of initial proximal lesions.

Introduction

In recent decades, improvement in dental hygiene and regular patient follow-up have led to important changes in caries presentation, with more concealed lesions and “hidden” caries, prompting the need to detect proximal carious lesions at earlier stages. Clinical examination, intraoral radiographs and temporary teeth separation can detect proximal dental caries [ ]. X-ray examination, the most common method for caries detection [ ], has advantages and limitations. Wenzel et al. [ ] have suggested that 30–40% of enamel has to be demineralized before an enamel lesion is visible by Rx examination and Yang and Dutra [ ] have stated that as much as 40–60% of tooth decalcification is required for the lesion to be detected on a radiograph. Bitewing (BW) X-rays seem more efficient at detecting caries that have reached dentin than at early diagnosis of initial enamel lesions [ ]. Moreover, BW radiographs cannot determine whether a detected lesion is active and/or cavitated [ ].

Increased emphasis on using less ionizing radiation [ ] and minimally invasive treatment of early caries has prompted research into the potential of light-based caries detection methods.[ ]

Transillumination is one of the oldest alternative caries detection methods besides radiographs.[ , ] Enamel’s optical properties are modified by very slight increases in enamel porosity, resulting in increased scattering when light passes through enamel.[ , , ] Visible light, used for fiber optic transillumination (FOTI), has evolved to digital fiber optic transillumination (DIFOTI) to allow digital recording and monitoring of early enamel lesions. Both techniques have been extensively described and compared with radiography and clinical examination. [ , ]

In the last decade, near infrared (NIR) light has been developed for occlusal and proximal caries.[ , ] Different NIR light wavelengths have been tested. [ , ] The DIAGNOcam (DC) device (KaVo, Biberach Germany) using 780 nm NIR transillumination technology recently entered the marketplace. This intraoral camera has two flexible extensions: one NIR light that transilluminates the tooth through the periodontal tissues, and a camera that captures the images from the occlusal surface of the examined tooth. This device, by reducing the need for BW radiography offers a safe alternative to ionizing radiography to detect caries in young patients and pregnant women.[ , ] Furthermore, a recent clinical study found that, when compared with BW radiographs, DC can detect more proximal enamel lesions. [ ] DC also provides information on the exact location of the lesion in the bucco-lingual dimension, showing the extent of the lesion in enamel until it reaches the enamel dentin junction, but DC cannot precisely determine the depth of the lesion in dentin. DC is thus more suited for enamel caries detection and monitoring, allowing early minimally invasive intervention if a lesion progresses towards the enamel-dentin junction.

The aim of this study was to assess the reproducibility of proximal caries detection using NIR transillumination device and BW radiography in a population at low risk for caries, and to assess the agreement between the two methods. We assumed that the progression rate of early lesions would be slow over a 2 year timeframe, thus providing reliable information on using NIR transillumination for caries detection and monitoring over time.

Methodology

Materials

BW radiographs were taken by different practitioners in 2013 and 2015, exposure parameters were set at 70 kV, 7 mA, 0.16–0.20 s. Various digital imaging plates size 2 and CS 7600 scanning system (Carestream Health, Rochester, NY, USA) were used. For NIR transillumination images, DIAGNOcam 2170U (Kavo, Biberach, Germany) was used with the KaVo integrated desktop (KID) software V 2.4.2 with the replaceable tip for adults.

Data collection

After obtaining ethical approval, retrospective analyses of the digital dental records of final year undergraduate dental students (n = 18) were carried-out. The students had previously consented to the anonymous use of their dental record including BW and DC images. Images were recorded between 2013 and 2015 and collected in 2016. Eighteen digital records were screened for BW radiography (approximately 24 months’ interval between first and second X-ray examinations), and 6 patients were excluded due to incomplete or missing information. Data from 12 patients (ages 22 to 32) with BW and DC images performed for a check-up in 2013, and corresponding BW and DC follow-up images performed in 2015, were analysed. For each patient, a PowerPoint file with 4 slides was made (separating the 4 quadrants). One slide, missing the DC images from 2015, was excluded. The images were a priori anonymized, the 47 slides were mixed, randomly reorganized, then numbered, conferring complete irreversible anonymity. Each of the final 47 slides with the BW and DC for 2013 and 2015 were then re-distributed on 5 slides for separate analysis of the following data: BW 2013, BW 2015, DC 2013, DC 2015, and full data. The final data set included 188 teeth (376 proximal surfaces). Each surface was assessed by BW and DC in 2013 and again in 2015.

Data interpretation

Two trained dentists, with extensive experience in radiology and NIR transillumination images, were calibrated to rate the DC and BW images. BW and DC images were analysed separately, and each interpreted proximal surface was given a letter score. The two observers agreed on one score for each image, which was recorded in an Excel file for the final statistical analysis. The two observers evaluated each image for each diagnostic method and year. Table 1 shows the scoring system and examples of BW and DC images for each score.

Statistical analysis

The descriptive data analysis was obtained. The correlation between both diagnostic methods was tested using the Kappa statistic, for dichotomous ratings, and intra-class correlation coefficient (ICC) (two-way, mixed as the methods were considered fixed factors but the tooth surfaces were random factors, and absolute level of agreement) for the 4-level scale. This method accounts for the severity of disagreement, represented by the number of categories for each discrepancy. All the data were analysed with SPSS software version 17 (Illinois, Chicago, USA).

Methodology

Materials

BW radiographs were taken by different practitioners in 2013 and 2015, exposure parameters were set at 70 kV, 7 mA, 0.16–0.20 s. Various digital imaging plates size 2 and CS 7600 scanning system (Carestream Health, Rochester, NY, USA) were used. For NIR transillumination images, DIAGNOcam 2170U (Kavo, Biberach, Germany) was used with the KaVo integrated desktop (KID) software V 2.4.2 with the replaceable tip for adults.

Data collection

After obtaining ethical approval, retrospective analyses of the digital dental records of final year undergraduate dental students (n = 18) were carried-out. The students had previously consented to the anonymous use of their dental record including BW and DC images. Images were recorded between 2013 and 2015 and collected in 2016. Eighteen digital records were screened for BW radiography (approximately 24 months’ interval between first and second X-ray examinations), and 6 patients were excluded due to incomplete or missing information. Data from 12 patients (ages 22 to 32) with BW and DC images performed for a check-up in 2013, and corresponding BW and DC follow-up images performed in 2015, were analysed. For each patient, a PowerPoint file with 4 slides was made (separating the 4 quadrants). One slide, missing the DC images from 2015, was excluded. The images were a priori anonymized, the 47 slides were mixed, randomly reorganized, then numbered, conferring complete irreversible anonymity. Each of the final 47 slides with the BW and DC for 2013 and 2015 were then re-distributed on 5 slides for separate analysis of the following data: BW 2013, BW 2015, DC 2013, DC 2015, and full data. The final data set included 188 teeth (376 proximal surfaces). Each surface was assessed by BW and DC in 2013 and again in 2015.

Data interpretation

Two trained dentists, with extensive experience in radiology and NIR transillumination images, were calibrated to rate the DC and BW images. BW and DC images were analysed separately, and each interpreted proximal surface was given a letter score. The two observers agreed on one score for each image, which was recorded in an Excel file for the final statistical analysis. The two observers evaluated each image for each diagnostic method and year. Table 1 shows the scoring system and examples of BW and DC images for each score.

Statistical analysis

The descriptive data analysis was obtained. The correlation between both diagnostic methods was tested using the Kappa statistic, for dichotomous ratings, and intra-class correlation coefficient (ICC) (two-way, mixed as the methods were considered fixed factors but the tooth surfaces were random factors, and absolute level of agreement) for the 4-level scale. This method accounts for the severity of disagreement, represented by the number of categories for each discrepancy. All the data were analysed with SPSS software version 17 (Illinois, Chicago, USA).

Results

Data collected from retrospectively-analysed dental records of 12 patients yielded 376 proximal surfaces acquired with 2 distinct caries detection methods; BW and DC with an average 2-year interval.

Each surface had 4 different readings (DC 2013, DC 2015, BW 2013, BW 2015). At each reading the surface was scored with one of the codes described in Table 1 . Table 2 shows, for all proximal surfaces, the distribution frequency for each year by each method. For each method and each year, Missing, Filled and Non-interpretable surfaces were excluded for the statistical analysis. Consequently, 192 surfaces were evaluable by both methods in 2013, 193 in 2015, and 207 surfaces were evaluable by BW in both years, 208 were evaluable by DC in both years, and 139 were evaluable for all 4 assessments. After the exclusion of Missing, Filled and Non-interpretable surfaces, a second frequency table was obtained for the 139 surfaces with all 4 ratings available ( Table 3 ).

Table 2
Frequency (%) distributions for all proximal surfaces (376 in total) including missing, non interpretable and filled surfaces.
Frequency (%) BW 2013 BW 2015 DC 2013 DC 2015
N= 376 376 376 376
Normal 171 (45.5) 187 (49.7) 61 (16.2) 67 (17.8)
Enamel lesion 69 (18.4) 63 (16.8) 121 (32.2) 98 (26.1)
Junction lesion 14 (3.7) 13 (3.5) 63 (16.8) 67 (17.8)
Dentine lesion 7 (1.9) 3 (0.8) 1 (0.3) 2 (0.5)
Missing 31 (8.2) 29 (7.7) 57 (15.2) 82 (21.8)
Not interpretable 75 (19.9) 66 (17.6) 66 (17.6) 50 (13.3)
Filled 9 (2.4) 15 (4.0) 7 (1.9) 10 (2.7)
Only gold members can continue reading. Log In or Register to continue

Stay updated, free dental videos. Join our Telegram channel

Jun 17, 2018 | Posted by in General Dentistry | Comments Off on Near infrared transillumination compared with radiography to detect and monitor proximal caries: A clinical retrospective study

VIDEdental - Online dental courses

Get VIDEdental app for watching clinical videos