Cone-beam computed tomography with ultra-low dose protocols for pre-implant radiographic assessment: An in vitro study.

Purpose: To evaluate the ultra-low dose scanning protocols of a newly developed CBCT for pre-implant radiographic assessment.

Materials and methods: A total number of 32 CBCT scans were exposed at 90 kV using ProMax 3D's standard (GS) and three ultra-low dose scanning protocols: high, mid and low definition (UL-HD, UL-MD and UL-LD) on eight human dry skulls with simulated soft tissue. The mAs values and the voxel size were 96mAs/200 µm, 28mAs/150 µm, 22.4mAs/200 µm and 7.5mAs/400 µm for scanning protocols GS, UL-HD, UL-MD and UL-LD, respectively. The overall image quality and the precision of anatomical landmarks were assessed on a 4-rank ordinal scale by seven observers. Logistic regression analysis and post-hoc Scheffé test were applied to analyse possible differences in image quality and recognition of the anatomical structures for the three ultra-low dose protocols, compared with the standard protocol. Additionally, observers performed bone quality assessment and bone quantity measurement at 96 predefined 2D cross-sectional images. A Pearson Chi-square test and a paired t-test were used to analyse assessed bone quality and quantity using the four scanning protocols respectively.

Results: For the CBCT unit, ProMax 3D, UL-LD was not diagnostically acceptable for pre-implant assessment, whereas the UL-HD and UL-MD were diagnostically acceptable regarding overall image quality, visibility of most anatomical structures and bone quality assessment. However, to recognise the border of mandibular canal and the border of maxillary sinus, standard protocol may be indicated for some cases. No statistically significant differences in bone height measurement were found when applying standard protocol and the three ultra-low dose protocols.

Conclusions: Low-dose scanning protocols may be applied for pre-implant radiographic assessment, although image quality can be hampered if the radiation exposure is too low and the voxel size too large.