[Lower jaw virtual reposition accuracy research]

The aim is to study of the accuracy of the virtual reposition of the mandible (LF) by comparing the virtual and real position of the LF heads, as well as the analysis of factors contributing to the occurrence of errors.

An experimental model of a human skull with a movable LF was made of radiopaque material. The resulting computed tomogram (CT) of the upper and lower jaws and temporomandibular joint (TMJ) was combined with dental scans of dental rows and CT scans and a report on the position of the heads of the LF was obtained. Then the LF was shifted to another position and this position was fixed. Optical bite recorders were obtained using an intraoral scanner. An experimental model with a fixed new position of the LF was placed in a tomograph and repeated CT was obtained. In the program, according to optical registers, a virtual reposition of the LF was carried out. The correspondence of the virtual and real position of the mandibular heads was compared.

The absolute difference of glenoid fossa space between the actual and virtual reconstruction was 0.542±0.3354 mm in the anterior joint section (M±δ, M - average mean, δ - standard deviation); for the upper joint section - 0.309±0.2246 mm; and 0.260±0.1862 mm for the posterior section. The integral indicator of surface coincidence (q) when the virtual and actual position are directly compared was 0.296±0.1254 (s.u.).

The total margin of error depends on the following factors: repeated scanning error (teeth arches and bite scans), repeated CBCT error, a combination of scan and CBCT error, a combination of teeth arch scan and bite scan error, error of TMJ separation. The latter two factors were the most significant.