INTRODUCTION A novel S-shaped rotary file was developed to conform to nonround canal cross sections. However, the instrument should also perform well in small curved canals. This study used micro-computed tomographic scans to test the effects of TRUShape (Dentsply Tulsa Dental Specialties, Tulsa, OK) and a conventional rotary on canal geometry. METHODS Twenty mandibular molars with independent mesial canals were submitted to preoperative micro-computed tomographic scans (20-μm resolution). Canals were randomly allocated to 2 groups using Vortex (Dentsply Tulsa Dental Specialties) and TRUShape for shaping according to the directions for use. Scans were obtained after size 20/.06 and 30/.06; the following outcome variables were calculated: dentin volume, smallest radicular wall thickness, treated canal surface, canal transportation, and accumulated hard tissue debris. The number of preparation errors was tabulated; data were statistically contrasted using repeated measures and factorial analyses of variance. RESULTS The initial canal sizes were similar in both groups. Both instruments promoted preparation with no overt procedural errors. Vortex removed significantly more dentin both at size #20 and #30 (P < .02). Significantly (P < .02) more radicular wall dentin remained toward the furcation in the TRUShape group (1.02 ± 0.25 mm vs 0.94 ± 0.22 mm). The untreated surface at size #20 was 28.5 ± 13.9 and 19.4 ± 8.9 (P < .05), and it was 15.1 ± 8.5 and 11.3 ± 4.9 (P > .05) at size #30 for TRUShape and Vortex, respectively. Canal transportation at size #30 varied between 85 ± 57 μm and 179 ± 65 μm; the overall transportation scores were significantly lower for TRUShape (P < .05). CONCLUSIONS TRUShape provided conservation of dentin by limiting dentin removal and in the absence of substantial canal transportation. In the present study, the use of TRUShape in the mesial canals of mandibular molars did not result in decreased amounts of unprepared canal surface.