Near-infra-red transillumination (NIR) for proximal caries detection-A pilot study

Method: Radiographic images of thirty three macroscopically noncavitated caries lesions from 16 patients  (>14 years) were obtained on standardized digital bitewing radiographs after receiving written informed consents. Consequently, NIR images (illuminating wavelength 780 nm) of these lesions were acquired using DIAGNOcam.  Two independent observers examined the NIR images and bitewing radiographs to score the carious lesions. The scoring system was as follows: 0: no lesion visible, 1: lesion less than halfway through enamel, 2: lesion more than halfway through enamel, but not involving dentin, 3: lesion involving dentin, but not less than halfway to the pulp, 4: lesion involving dentin, more halfway to the pulp. Each examiner assessed all images for three times. Wilcoxon signed-rank test was used to compare the scorings of DIAGNOcam images and bitewing radiographs.  The correlation between the DIAGNOcam and bitewing scores was assessed with Spearman’s rho correlation coefficient, whereas interexaminer agreement was determined with Cohen’s Kappa. All statistical analyses were performed using SPSS 10.0 for Windows (SPSS, Chicago, IL, USA), p was set as 0.05.

Result: There were no statistically significant differences between the DIAGNOcam and bitewing scores (p=0.545>0.05). Spearman’s rho correlation coefficient was r=0.822, revealing a strong correlation between scorings of NIR and bitewing images (p=0.00<0.05). Kappa values for DIAGNOcam and bitewings were 0.92 (p<0.05) and 0.83 (p<0.05), respectively, indicating a “good” interexaminer agreement with both methods.

Conclusion: The results of this preliminary study showed that NIR transillumination had a  similar efficacy for the detection of proximal noncavitated caries lesions in comparison to that of bitewing radiography. DIAGNOcam appears to provide clinically efficient imaging without biohazards of radiation.

(DIAGNOcam was kindly provided by KaVo, Biberach)