The aim of this study was to investigate the accuracy of minimally invasive endodontic cavity preparation with the guidance of cone-beam computed tomography (CBCT) and three-dimensional printed templates through analyzing distance deviation and angle deviation.

Forty extracted human single-rooted maxillary first premolars with double root canals were selected. Preoperative CBCT scanning of all specimens was used for designing access cavities. All teeth were numbered and assigned into two groups according to the random number table. There were twenty teeth including forty root canals in each group. The minimally invasive endodontic cavity was designed as "Truss" access cavity. Preparation for access cavities in group A was guided by three-dimensional printed templates which were fabricated according to the CBCT images. Access cavities in group B were prepared according to the marks on occlusal surface after the analysis of CBCT images. Distance deviations and angle deviations in buccopalatal and mesiodistal directions as well as the areas of access cavities at occlusal and pulp horn levels were measured and calculated in the CBCT images. Statistical analysis for the data was performed by SPSS 20.0.

The data were analyzed with Mann-Whitney U test, since they were not satisfied with the normal distribution and the homogeneity of variance. No statistic difference was observed between the two groups in the areas of access cavities at occlusal level [group A: (1.34 ± 0.18) mm², group B: (1.30 ± 0.15) mm²; U=0.876, P=0.393] or at pulp horn level [group A: (1.74 ± 0.20) mm², group B: (1.67 ± 0.24) mm²; U=1.290, P=0.194]. The mesiodistal distance deviations at occlusal level in group B [ (0.37 ± 0.26) mm] were less than that in group A [ (0.52 ± 0.30) mm], which had significant difference (U=2.237, P=0.024) . There was no significant difference between the two groups in the buccopalatal distance deviations of occlusal level [group A: (0.45 ± 0.40) mm, group B: (0.41 ± 0.28) mm; U=0.385, P=0.697], the buccopalatal distance deviations of pulp horn level [group A: (0.32 ± 0.25) mm, group B: (0.41 ± 0.30) mm; U=1.290, P=0.199], the mesiodistal distance deviations of pulp horn level [group A: (0.38 ± 0.24) mm, group B: (0.35 ± 0.26) mm; U=0.905, P=0.500], as well as the buccopalatal angle deviations [group A: (4.76 ± 3.04) °, group B: (4.72 ± 3.61) °; U=0.404, P=0.679] or the mesiodistal angle deviations [group A: (3.04 ± 1.97) °, group B: (3.05 ± 2.45) °; U=0.467, P=0.637].

CBCT scanning was able to accurately guide the preparation of the minimally invasive endodontic cavities for the teeth with normal root canal anatomy. The guidance of CBCT scanning with templates did not significantly enhance the accuracy of access cavities, compared to that of CBCT scanning alone. Moreover, no significant difference was detected in the actual sizes of "Truss" access cavities prepared with the guidance of the two approaches.