Application of an auto-edge counting method for quantification of metal artifacts in CBCT images: a multivariate analysis of object position, field of view size, tube voltage, and metal artifact reduction algorithm.

OBJECTIVE

The objective was to assess the effects of object position, field of view (FOV) size, peak kilovoltage (kVp), and a metal artifact reduction (MAR) algorithm on metal artifacts in cone beam computed tomography (CBCT) as measured with an auto-edge counting method.

STUDY DESIGN

A titanium implant and a stainless steel intracanal post in a root were inserted in bovine rib blocks. CBCT scans were acquired with changes in object position (incisor, canine, and premolar-molar areas), FOV, kVp, and MAR (on or off) mode. Images were quantitatively analyzed in MATLAB by using the Canny edge detection algorithm. Four-way analysis of variance and Tukey tests were applied for data analysis.

RESULTS

The implant produced no significant differences in number of artifacts among the object positions through changing the kVp and MAR mode for all FOV sizes (P > .05). The intracanal post scanned with the medium-sized FOV, high kVp, and MAR off mode generated significant differences among object positions (P = .033). Among the variables assessed, FOV size and MAR mode had a significant influence on the number of artifacts (P ≤ .039).

CONCLUSION

Reduction of FOV size and application of the MAR tool significantly decreased the number of streak artifacts. The Canny edge detection algorithm could be an efficient method of metal artifact quantification.