Diagnostic accuracy of interproximal caries using NIR imaging with OS750 in an ex vivo setting

Extracted intact human teeth (n=20) were assembled in the typodont from the second molar to the canine for each of 4 quadrants and a total of 32 interproximal surfaces were used for the assessment (Figure ​(Figure1).1). A small lesion with approximately 1.0 mm diameter and 0.5~1.0 mm depth was created on 16 interproximal surfaces using a #2 carbide bur. These holes were packed with the hydroxyapatite powder (mock lesion, Figure ​Figure5-A1,5-A1, -A2). Another 16 interproximal surfaces were kept as intact (Figure ​(Figure5-A3).5-A3). Two microliters of 1/100 dilution of the stock solution of OS750 was applied to all the interproximal surface lesion. The NIR illumination light (730nm, Laser diode, Fig. ​Fig.5B)5B) was excited over the interproximal area at 90-degree angle (Figure ​(Figure5C).5C). Emission fluorescence was captured by the CCD camera (780nm, Figure ​Figure5C5C and ​and5D)5D) and visible on a head-mount display (Figure ​(Figure5E,5E, 5-F1, 5- F2, and 5-F3). Two dentists repeated the assessment 3 times with 1 week in between and repeatability was calculated as Cohen's κ. Eight dentists with clinical experience of 3~20 years examined each interproximal surface and graded them as either “Intact” or “Lesion” using NIR images and the sensitivity, specificity and inter-examiner error were calculated. Sample size (p = 0.05, power = 0.90) was calculated ¹¹.

Caries imaging system using the head-mount display. A: radiograph (A1: intact, A2 and A3: mock lesion), B and C: The Near-Infrared (NIR) illumination light (750nm, Laser diode) was excited over the interproximal area at 45~90 degree angle. D: Emission fluorescence was captured by the CCD camera (780nm) from the buccal side. E: The signal was sent to the head-mount display. F1 and F2 shows NIR images for the interproximal lesion, F3 indicates image of intact surface.