O. A. Peters, C. I. Peters, K. Schonenberger & F. Barbakow
Endodontic Division, Department of Preventive and Restorative Dental Sciences, University of California, San Francisco, USA.
Endodontic Department, University of The Pacific Dental School, San Francisco, USA.
Division of Endodontology, Department of Preventive Dentistry, Periodontology and Cariology, University of Zurich, Switzerland.

Introduction.
Modern engine-driven root canal preparation techniques claim to facilitate safe and efficient preparations. Such techniques have been evaluated in several studies using various experimental designs. Conventional analytical methods may employ reassembly techniques (Bramante et al. 1987) which evaluate cross-sections of root canals before and after preparation (Deplazes et al. 2001).
Recently, micro computed tomography (mCT) was introduced to evaluate not only cross-sections of roots, but also three-dimensional shapes of canals at resolutions as high as 36 mm (Peters et al. 2000, Rhodes et al. 2000, Bergmans et al. 2001, Gluskin et al. 2001). This innovation was achieved because new hardware and software was available to evaluate the metrical data created by mCT, thus allowing geometrical changes in prepared canals to be determined in more detail (Peters et al. 2000).
Evidence suggests that canal anatomy influences preparation outcomes: significantly more aberrations are recorded when preparing simulated canals with more acute curves in plastic blocks using various nickel-titanium (Ni-Ti) rotary instruments (Bryant et al. 1999). In addition, three-dimensional analysis using mCT indicated that canal transportation was more pronounced when shaping narrow curved canals than wider specimens (Peters et al. 2001a).
Nickel-Titanium rotary instruments such as ProTaper (Dentsply Maillefer, Ballaigues, Switzerland) have a modified cross-sectional design that resembles a K-File configuration instead of the U-shape common to many other rotary instruments. Rotary instruments with this geometry are claimed to cut dentine more effectively, and may therefore reduce torsional loads (Ruddle 2002). However, more aggressive cutting could produce increased canal transportation and, aside from clinical guidelines (Ruddle 2001), little information exists about these instruments. Consequently, the aim of this study was to assess the shaping potential of ProTaper instruments, and to evaluate the effect of normal canal anatomy on the final outcome of the shaped canal using different variables.

Preparation of specimens.
Eleventhree-rooted maxillary molars were selected from a pool of extracted teeth and stored in 0.1% thymol until used. Outer root surfaces were sealed from their apices to the cemento-enamel junctions, and the specimens were mounted on SEM stubs (014001-T, Balzers Union AG, Balzers, Liechtenstein). The canals were then scanned by mCT (see below), without probing the canals for patency to avoid modifying the canals apical anatomy. No attempt was made to locate or shape the second mesiobuccal (mb) canals because their anatomy was too variable for the purpose of this study (Fig.1). However, the mCT analysis indicated that a fourth canal was present in nine of the11specimens in this sample.
Canals were prepared using a set of ProTaper instruments, consisting of Shaping Files 1 and 2 (S1 and S2) and the Finishing Files 1 through 3 (F1-F3). Shaper X instruments were not available during the course of the study. Canals were prepared in a special torque-testing device (rotational speed 250 r.p.m.), which is described in detail elsewhere (Peters & Barbakow 2002). The operator was an endodontist with expertise in rotary techniques, and after a training period with ProTaper instruments and the torque-testing device (C.I.P.). During that phase, four S1 instruments fractured and subsequently, care was taken to enlarge the apical extent of the canals with a K-Flexofile size 015, prior to the use of ProTaper instruments.

Figure 1. Images of unshaped (A) and prepared (B) systems reconstructed from micro computed tomography (mCT) data. Note unusual canal configuration with mesiobuccal (mb), distobuccal (db) canals and second distobuccal (db2) system merging into the palatal (p).

Canal orifices were enlarged with Gates Glidden burs (insertion depth 3 mm, nos.3 and 2; Dentsply Maillefer), and pulp chambers were irrigated with 5 mL of tap water. Working lengths were then set by subtracting 0.5 mm from the lengths of size 010 K-Flexofiles (Dentsply Maillefer, Ballaigues, Switzerland) when their tips were just visible at the main apical foramina. Digital radiographs (Digora, Soredex, Helsinki, Finland) were also taken of each canal to verify file position and canal anatomy. Apical preparations began with size 015 K-Flexofiles, using Glyde (Dentsply Maillefer) as the lubricant. Subsequently, canals were enlarged with instruments S1 and S2 used in a gentle pumping and brushing action. The mb and distobuccal (db) canals were prepared to a F2 (D1 diameter 0.25 mm), whilst palatal (p) canals were shaped to a F3 (D1 diameter 0.3 mm). Tap water served as the irrigant after each instrument, delivered by means of a gauge 27 needle, allowing for adequate back flow. After preparing one specimen (three canals), each set of ProTaper instruments was discarded and replaced by a new set.
Micro CT measurements and evaluations Scanning and evaluation procedures have been described elsewhere in detail (Peters et al. 2000, 2001a). Briefly, specimens were scanned at an isotropic resolution of 36 mm using a mCT system (mCT-20, Scanco Medical, Bassersdorf, Switzerland), and then, binary images of the root canals were constructed (Fig. 1) after filtering and thresholding. The canals were again scanned, as above, after shaping, so that each canal served as it sown control.
Whilst a special mounting device ensured almost exact repositioning of the pre- and postpreparation images, precision was perfected by superimposing both sets of segmented root canals manually over each other. Finally, the best superimposition was automatically detected, with a precision better than one voxel, by varying the relative translation in x-, y-, and z-directions.
Subsequently, matched root canals were evaluated as follows: changes in volume and surface area of the root canals were determined from the triangulated data. The same models were also used to determine the Structure Model Index (SMI) of the canals. This index characterizes the structure of an object as having an ideal ribbon-like shape, corresponding to an SMI score of 0 or cylindrical shape, corresponding to an SMI of 3. Furthermore, ‘thickness’ of the canals were determined using recently described distance transformation techniques (Peters et al. 2000) and related to canal length in order to construct ‘thickness’ profiles (Fig. 2). Canal tapers, as line slopes, were determined from thickness profiles by fitting simple regressions.

Figure 2. Canal 'thickness' profiles detailing clinical diameters along canal length (mean scores from 11 canals each).

Based on an overall median canal volume of 2.94 mm3, 32 canals in 11 of 15 teeth were divided into ‘wide’ (mean volume: 5.54 _ 2.04 mm3) and ‘constricted’ (mean volume: 2.03 _0.66 mm3) groups. One p canal was not included in this analysis due to ambiguous surface area scores. Then, ‘centres of gravity’ of the canals, calculated for each slice, were connected along the z-axis by a fitted line. Canal transportations (CM shifts in millimetres, Fig. 3), were calculated by comparing the centres of gravity before and after treatment for the apical-, mid- and coronal-thirds of the canals. From a polynomic equation, describing a fitted line for each canal, curvatures were calculated as second derivatives.
Finally, matched images of the surface area voxels of the canals, before and after preparation, were examined to evaluate the amount of uninstrumented surface (Fig. 4). This parameter was calculated by subtracting the number of static surface voxels from the total number of surface voxels. Scores expressed as mean _ SD were compared using one- and two-way anovas with Scheffe. tests for posthoc comparisons. When appropriate, repeated-measures anovas were constructed. A level of P < 0.05 was considered significant.

Figure 3. Canal transportation (CM) shift in mesiobuccal (mb), distobuccal (db) and palatal (p) canals, split by canal third (mean +SD, n =11 in each group). *P < 0.05, two-way anova.



Figure 4. Matched and superimposed root canal systems shown in Fig.1. Prepared canal areas indicated by green colour.

Scanning of unprepared and instrumented canals yielded highly detailed three-dimensional canal images (Fig.1). No obvious procedural errors such as apical zips perforations or ledges were detected after canal preparation with ProTaper. In addition, no clear changes in overall shape were detected in eight of 32 canals. This phenomenon was most prevalent in the apical parts of wider canals. However, distinct changes were recorded in the coronal parts of all canals, and were most likely due to preflaring the canal orifices.
Overall, median initial canal volume and surface area were 2.94 mm3 and 24.15 mm2, respectively. Table 1 details preoperative mean scores, indicating that mb canals were significantly (P < 0.05)more ribbon-shaped than the other two canal types, as shown by respective SMI scores. Repeated-measures anova revealed that preparation significantly increased canal volumes, surface areas and SMI scores (Table 2). At the same time, SMI increase was highest in mb canals and anova indicated significant differences between canal types in this respect (P < 0.05; T able 2). Figure 2 illustrates diameters of canals, evaluated as ‘thickness’, by plotting means against canal lengths and yielding canal dimension estimates, before and after preparation. Overall ‘thickness’ increased significantly whilst canals were prepared, and resulting profiles indicated smooth tapers of .08 for the apical 5 mm in mb and db canals as wellas .10 for p canals. Canal ‘thickness’ at the 5-mm level increased from 0.38 to 0.65, 0.42 to 0.66 and 0.57 to 0.79 mm, for mb and db and p canals, respectively.
Using superimposed canal models, mean centre of mass shift (CM shift) scores were calculated for the coronal, middle and apical thirds, and scores ranged from 0.021 to 0.407 mm. Figure 3 shows mean scores for canals that were further grouped into ‘constricted’ and ‘wide’ groups with the former comprising seven of 11 mb canals, seven of11db canals and two of10 p canals. The ‘wide’ group included the remaining 16 canals. Two-way anova indicated a significant difference between the ‘wide’ and ‘constricted’ groups in the coronal two thirds, whilst CM shifts in the apical thirds were similar (0.120 _ 0.07 vs. 0.135 _0.09 mm, Fig. 3).
Most canals used in this study were moderately curved, curvature being metrically described as the second derivative of a fitted line through successive canal centres. However, canal preparation led to various degrees of straightening (Table 3). This effect was most pronounced in those canals, which had higher initial degrees of curvature (Fig.1, note curved p canal). Statistically, no significant differences were recorded when comparing the ‘wide’ and ‘constricted’ groups or between mb, db and p canals.
Superimposed images, with colour-coded static areas, designated relatively large untreated areas (Fig. 4).These areas tended to be mid-root at the convex side and apically at the concave side of the curvature. Finally, amounts of static surface voxels, or untreated areas were calculated (Table 4). The db canals had the lowest numbers of untreated voxels when compared to mb and p canals. This difference was not significant whilst constricted canals had a significantly lower number of static voxels in comparison to large canals (32 _22% vs.52 _26%, P < 0.05).

Table 1. Morphometric data determined for untreated maxillary root canals.



Table 2. Increase in morphometric scores after preparation of maxillary root canals.



Table 3. Relative degree of canal straightening.



Table 4. Numbers and percentages of static voxels recorded by superimposing matched images, before and after preparation.

Discussion.
ProTaper instruments were recently introduced and embody two new concepts. Firstly, in cross-section, the instruments do not have a U-file design and secondly, the instrument’s shaft has variable tapers along its cutting surface (Ruddle 2001). This concept minimizes the number of instruments per set and is claimed to decrease torsional loads by reducing the frictional surface, thereby increasing cutting efficiency. No previous reports have evaluated the shaping ability of ProTaper Ni-Ti instruments using mCT. The mCT is emerging in several endodontic research facilities as a nondestructive and accurate method to analyse canal geometry and the relative effects of shaping techniques (Rhodes et al. 2000, Bergmans et al. 2001, Gluskin et al. 2001, Peters et al. 2001b). Accuracy and reproducibility of the system used in this study has been verified previously (Peters et al. 2000). This project is part of a larger study testing torque and force produced when canals are shaped using extracted human maxillary molars in a special testing device (Peters & Barbakow 2002).
Originally, ProTaper sets included five instruments, namely, Shaping files 1 and 2 and Finishing files 1-3, and such sets were used in the present study. However, an additional instrument (Shaper X) was subsequently introduced, whose task is to relocate canal orifices and shape the coronal part of a canal. In the present study, this coronal preflaring was completed with Gates Glidden drills. Whilst the modified cutting flute of ProTaper instruments might reduce friction and consequently torque (Blum et al. 1999), this design may also increase the incidence of procedural errors and overall canal transportation. The present study addressed this question and furthermore evaluated the effect of canal anatomy on preparation outcome.
No obvious procedural errors were detected in this study, confirming findings reported in two earlier studies, in which four other Ni-Ti preparation systems were evaluated using mCT (Peters et al. 2001 a, b). However, when transportation was expressed as CM shifts, varying degrees of canal straightening were recorded with ProTaper, as was the case with other previously described preparation techniques (Peters et al. 2001 a, b).
Overall canal anatomy, as described by volumes, surface areas and SMI, was statistically similar in the present study, compared to canals evaluated earlier using the same analytical methods (Peters et al.2001 a, b). Pro- Taper preparation removed dentine volumes varying from1.40 to1.76 mm3 compared to a preoperative canal volume of 2.42-5.27 mm3.Whilst some of these values for individual canals are lower than those previously described (Nielsen et al. 1995), the differences are probably due to varying regions of interest (ROI). However, calculations for a perfect 10-mm long cone with a 0.25- mm tip diameter and a.10 taper result in a theoretical file volume of 4.42 mm3. This result should correspond to the volume of a perfectly prepared canal of similar dimensions. In principle, it seems intriguing to refer to these volumes when deciding on irrigation parameters (Yamada et al.1983).
In this study, canal diameters were described as ‘thickness’, which was calculated by fitting spheres into reconstructed canals as described previously (Peters et al. 2000, 2001a). Specifically, maximum local sphere diameter relates to a specific file tip size, which a clinician would select to gauge the apical region (Ruddle 2002). ProTaper instruments adequately opened canals 5 mm from their apices, with sizes varying from 0.65 to 0.79 mm. Spreaders and pluggers with size 0.5-mm tips could readily be used during obturation of root canals with such apical preparations. Deep instrument penetration is considered critical for both lateral (Alison et al.1979) and vertical (Ruddle 2002) compaction. Canal ‘thickness’ is also an important parameter when considering how far into a canal irrigation needles can be safely inserted to allow for back flow.
Although procedural errors were not obvious, some canal transportation was evident in the present study. In fact, CM shifts were slightly larger than those reported previously (Peters et al. 2001a) for rotary instruments with a U-file cross-sectional geometry. This is possibly significant when the smaller apical sizes of ProTaper instruments (0.25-0.30 mm) are considered. Importantly, there was no significant difference in apical transportation when the findings for ‘wide’ canals were compared to those of their ‘constricted’ counterparts. However, results suggested that coronal preflaring with relocation of canal orifices was sufficient to avoid apical preparation errors. Whilst some canal straightening occurred, there was again no difference between the various canal types (p, db, mb) when graded as ‘wide’ or ‘constricted’.
The impact of preoperative canal anatomy was most prominent when assessing the amount of uninstrumented canal areas after preparation. Canals graded as ‘wide’ had significantly larger untouched areas (Fig. 4; Table 4), amounting to 43-49% of their total compared to their ‘constricted’ counter parts. Similar findings have been reported earlier for other techniques using mCT reconstructions (Peters et al. 2001a) as well as from canal cross-sections (Tucker & Wenckus 1997). It has been proposed that canals be prepared to sufficiently large apical sizes, firstly to optimize irrigation and disinfection (Ruddle 2002), and to facilitate elimination of microbes mechanically (Dalton et al. 1998). However, the clinical significance of the parameter ‘prepared surface’ is not yet clarified considering that viable microbes penetrate deeper into dentinal tubules and may persist during root canal treatments (Peters et al. 2002).

References.

Alison DA, Weber CR, Walton RE (1979) The influence of the method of canal preparation on the quality of apical and coronal obturation. Journal of Endodontics 5, 298-304.
Bergmans L, Van Cleynenbreugel J, Wevers M, Lambrechts P (2001) A methodology for quantitative evaluation of root canal instrumentation using microcomputed tomography. International Endodontic Journal 34,390-8.
Blum JY, Machtou P, Micallef JP (1999) Location of contact areas on rotary ProFile instruments in relationship to the forces developed during mechanical preparation on extracted teeth. International Endodontic Journal 32,108-14.
Bramante CM, Berbert A, Borges RP (1987) A methodology for evaluation of root canal instrumentation. Journal of Endodontics 13, 243-5.
Bryant ST, Dummer PMH, Pitoni C, Bourba M, Moghal S (1999) Shaping ability of .04 and .06 taper ProFile rotary nickel-titanium instruments in simulated root canals. International Endodontic Journal 32,155-64.
Dalton BC, Qrstavik D, Philips C, Pettiette M, Trope M (1998) Bacterial reduction with nickel-titanium rotary instrumentation. Journal of Endodontics 24,763-7.
Deplazes P, Peters O, Barbakow F (2001) Comparing apical preparations of root canals shaped by nickel-titaniumrotary and nickel-titanium hand instruments. Journal of Endodontics 27,196-202.
Gluskin AH, Brown DC, Buchanan LS (2001) A reconstructed computerized tomography comparison of Ni-Ti rotary GT files versus traditional instruments in canals shaped by novice operators. International Endodontic Journal 34,476-84.
Nielsen RB, Alyassin AM, Peters DD, Carnes DL, Lancaster JL (1995) Microcomputed tomography: an advanced system for detailed endodontic research. Journal of Endodontics 21, 561-8.
Peters OA, Barbakow F (2002) Dynamic torque and apical forces of ProFile .04 rotary instruments during preparation of curved canals. International Endodontic Journal 35,379-89.
Peters OA, Laib A, Gohring TN, Barbakow F (2001b) Changes in root canal geometry after preparation assessed by high-resolution computed tomography. Journal of Endodontics 27,1-6.
Peters OA, Laib A, Ruegsegger P, Barbakow F (2000) Three dimensional analysis of root canal geometry using high-resolution computed tomography. Journal of Dental Research 79, 1405-9.
Peters OA, Schonenberger K, Laib A (2001 a) Effects of four Ni-Ti preparation techniques on root canal geometry assessed by micro computed tomography. International Endodontic Journal 34, 221-30.
Peters LB, van Winkelhoff A-J, Bujis JF, Wesselink PR (2002) Effects of instrumentation, irrigation and dressing with calciumhydroxide on infection in pulpless teeth with periapical bone lesions. International Endodontic Journal 35,13-21.
Rhodes JS, Pitt Ford TR, Lynch PJ, Liepins PJ, Curtis RV (2000) A comparison of two nickel-titanium instrumentation techniques in teeth using microcomputed tomography. International Endodontic Journal 33, 279-85.
Ruddle C (2001) The ProTaper endodontic system: geometries, features and guidelines for use. Dentistry Today 20,60-7.
Ruddle C (2002) Cleaning and shaping the root canal system. In: Cohen S, Burns RC, eds. Pathways of the Pulp, 8th edn. St. Louis, MO: Mosby, 231-92.
Tucker DM, Wenckus CS (1997) Canal wall planning by engine driven nickel-titanium instruments, compared with stainless-steel hand instrumentation. Journal of Endodontics 23, 170-3.
Yamada RS, Armas A, Goldmann M, Lin PS (1983) A scanning electron microscopic study of a high-volume final flush with several irrigation solutions. Part 3. Journal of Endodontics 9, 137-42.