Specimen selection and preparation
Thirty single-root premolars were randomly selected from a collection of extracted human teeth from a Chinese population sample based on mature apices without visible apical resorption and no prior endodontic treatment. These teeth were extracted because of periodontitis or orthodontic need. After understanding and written consent was obtained from patients, the extracted teeth were collected by the West China Hospital of Stomatology for teaching and research. The present study was approved by the Ethics Committee of the West China Hospital of Stomatology, and the premolars were selected from the teeth bank of the hospital. After extraction, tissue fragments and calcified debris were removed from the teeth by scaling and the teeth were stored in 0.1% thymol until used.
Preoperative radiographs of each selected tooth were first taken from the buccolingual and mesiodistal directions, then their canal system classifications were evaluated by an endodontist. After the radiographic evaluation and without probing the canals for patency so as to avoid modifying the canals' apical anatomy, premolars whose canal systems were classified into one of four categories (types I, II, III, IV) according to Weine's classification [
8] were scanned using a micro-CT system (μCT-80; Scanco Medical, Bassersdorf, Switzerland) with an isotropic voxel size of 36 μm. Images were acquired from 622 slices of each tooth. From these images, a 3-D model of the tooth and canal system was constructed with a framework system (MeVisLab 2.0, MeVis Research, Bremen, Germany) on a personal computer (Athlon II X2, 2.8 GHz CPU, 2 Gbyte RAM, Windows XP) according to a customized application framework using MeVisLab [
29]. The reconstructed 3-D model of the root canal system in each premolar was carefully examined. Unexpectedly, in addition to the four categories of Weine's classification [
8], we found one mandibular first premolar whose canal system was classified as type V based on the system of Yoshioka
et al. [
9], which was not recognized by the professional endodontist from the preoperative periapical radiographs. Finally, five representative single-root premolars (teeth A, B, C, D, and E), each of whose canal system was classified into one of five categories (types I, II, III, IV, and V), were included in this study. In addition to the reconstruction of the tooth and canal, we also calculated the root canal diameter and showed its morphology with a 3-D color-coded image of the prepreparation canal systems.
After preparing a standard access cavity with #2 and #4 high-speed round carbide burs, using ample water cooling, each root canal was passively negotiated with a #10 K-file to the apical foramen. The working length of the canal was determined by observing the tip of the file protruding through the apical foramen and subtracting 1 mm from the recorded length.
In tooth E, however, the entrance of the lingual canal formed a perpendicular angle with the buccal canal, so we could not directly explore this canal with a #10 K-file. To find the lingual canal orifice, we adjusted the procedure: based on the length and location of the canal identified through the 3-D tooth model, a #40 K-file was used to remove the overhanging dentin above the orifice of the lingual canal, filing toward the occlusal surface. Chelating agents were also used to soften the overhanging dentin during this procedure. After the overhanging dentin was cleared, we explored the lingual canal with #8 and #10 K-files, then the working length was determined as per the aforementioned method.
According to the manufacturer's instructions, all root canals were explored with a #15 K-file after exploring with a #10 K-file, then the canals were prepared using a set of new HPT instruments (Dentsply Maillefer). Instrumental sequences followed the manufacturer's instructions: first, canals were flared coronally with S1 (followed by SX if necessary), then their working lengths were measured and confirmed with a #15 K-file, after which they were prepared with S1, S2, F1, F2, and F3 at the working length, using a "modified balanced forces" motion.
During preparation, RC-Prep (Premier Products, Plymouth Meeting, PA) was used as the lubricant. Irrigation was performed with 2 ml of 1% sodium hypochlorite (NaOCl) solution after each instrument and canal patency was ascertained with a #10 K-file for each canal. All root canals were prepared by a single operator.
Micro-CT measurement and 3-D evaluation
After preparation with each HPT instrument at the working length, canals were dried with sterile paper points, then the teeth were scanned using the same micro-CT system. A series of micro-CT images were again obtained with the same isotropic voxel size of 36 μm.
To evaluate the efficacy of root canal preparation, volumes of interest were selected extending from the cemento-enamel junction to the apex of the roots. Using the customized application framework MeVisLab [
29], the canal models (pre- and postpreparation) were reconstructed and superimposed and the instrumentation characteristics were quantitatively measured in 3-D. Numeric values were obtained for canal surface area, volume, volume changes, percentage of untouched surface, dentin wall thickness, and the thickness of dentin removed. The location of dentin removed or a lack of change during preparation were also demonstrated using the 3-D color images. Preparation errors such as canal straightening, ledging, elbow formation, or zipping were also evaluated using the color-coded reconstruction.