The present study sample was comprised of 21 human mandibular molar teeth with no restorations and with pit-fissure occlusal caries that were extracted due to periodontal loss or pericoronitis. Informed consent was obtained for the use of all extracted teeth. Teeth were cleaned of calculus and debris, disinfected in 2% sodium hypochlorite solution for 20 min, and stored in distilled water. Each tooth was embedded in a block of silicone impression material. All teeth were imaged using analog X-ray film, CCD digital X-ray, two different CBCT units and a micro-CT unit. Pilot studies were conducted to determine image acquisition exposure parameters, with visibility of the pulpal root canal, dentine, and enamel used as an indication of optimal image quality for all images. Intra-oral radiographs were taken using Size 2 E/F Kodak Insight Film (Eastman Kodak Co, Rochester, NY) and an RVG 5.0 CCD (Trophy, Marne la Valle, France) with a size 1 sensor and 10–15 lp mm spatial resolution. All intra-oral images were exposed under reproducible conditions using an Evaluation X 3000-2C X-Ray unit (New Life Radiology Srl, Grugliasco, Torino, Italy) at 70 kVp, 8 mA, and a focus-receptor distance of 30 cm behind a plexiglass soft-tissue equivalent. Exposure times were 0.4 s for film and 0.1 s for CCD. Films were automatically processed on the same day as exposure using an Extra-x Velopex (Medivance Instruments Ltd., London, England) and fresh chemicals (Hacettepe, Ankara, Turkey) in accordance with the manufacturer's instructions. Intraoral images obtained with RVG Trophy CCD sensor are shown in Fig. .
Intraoral images obtained with RVG Trophy CCD sensor
CBCT images were taken using Iluma (3M Imtec, Oklahoma, USA) and 3D Accuitomo 170 (3D Accuitomo; J Morita MFG. Corp, Kyoto, Japan) systems. Iluma images were obtained at 120 kV, 3.8 mA, and an exposure time of 40 s with a 24.4
19.5 cm amorphous silicon flat-panel image detector and reconstructed using a voxel size of 0.2 mm3
. Accuitomo images were obtained at 63 kV, 2 mA, an exposure time of 30.8 s, and a voxel size of 0.125 mm using 6 cm
6 cm field of view (FOV). Axial scans and multiplanar reconstructions were obtained for both CBCT systems, and volumetric data was reconstructed using the systems' own software programs to provide serial coronal and sagittal sections along each tooth plane. CBCT images obtained with Accuitomo 170 unit are shown in Fig. .
CBCT images obtained with Accuitomo 170 unit
Micro-CT images of tooth crowns were obtained using a SkyScan 1174 (SkyScan, Kontich, Belgium) at 50 kV, 800 μA and a pixel size of 18-μm. Axial images were exported as bitmap (bmp) files to the Mimics software program (Materialise, Leuven, Belgium) in order to obtain serial coronal and sagittal images. While every scanning procedure, micro-CT system created a .log file which has whole scanning parameters in internationally calibrated metric system. Additionally, when uploading files to Mimics software, we also aligned and adjusted the sections according to parameters displayed in software interface. Micro-CT images obtained with SkyScan 1174 unit are shown in Fig. .
Micro-CT images obtained with SkyScan 1174 unit
Following imaging, teeth were embedded in acrylic and serially sectioned mesio-distally in parallel to the long axis using a water-cooled diamond saw at low speed. For each tooth, the section with the deepest carious lesion was scanned using a high-resolution scanner (Epson Perfection V750-M Pro Scanner) at 600 dpi resolution and ×15 magnification and viewed using the AxioVision Rel. 4.7 image software (Carl Zeiss Imaging Solutions, Göttingen, Germany). Histological sections obtained from the teeth are shown in Fig. .
Histological sections obtained from the teeth
Images were evaluated separately by three oral radiologists experienced in image interpretation. Each image set (film, CCD, Iluma CBCT, Accuitomo CBCT, micro-CT, and histology) was viewed in a random order in a different session at 1-week intervals. Intra-observer agreement was assessed by having each observer view all image sets twice, with a 7-week interval between viewing to eliminate memory bias. Conventional film radiographs were evaluated using a light box and magnifier (×2), and a digital caliper (Absolute Digimatic, Mitutoyo Corp., Kawasaki, Japan) was used to measure lesion depth, defined as the distance from the enamel surface to the innermost part of the lesion. Digital 2-D and 3-D images were viewed on a 22" LG Flatron monitor (LG, Seoul, South Korea) at a 1,440
900 pixel screen resolution and 32-bit color depth. CCD and CBCT measurements of the largest dimensions of occlusal caries lesions were performed using the built-in measurement tools provided by the imaging software. Micro-CT measurements were obtained from the available sections using the Mimics software measurement tools, and histological measurements were performed using the AxioVision Rel. 4.7 imaging software.
Intra- and inter-rater reliability was assessed with the Bland/Altman method by calculating Intraclass Correlation Coefficients from residual variances of one-way ANOVA (with reading as the main variable) and two-way ANOVA (with the variables reading, observer and the interaction between the two). One-way ANOVA for each observer was also used to calculate a within-variability for each of them. Statistical analysis was performed using SAS 9.0 (PROC GLM), and Bland/Altman plots were drawn with the limits of agreement using MedCalc.