Feasibility of measuring condylar inclination from facial scans using eccentric interocclusal records across...

Although digital prosthodontic workflows have improved anatomic data acquisition for articulator mounting and jaw simulation, virtual articulator systems still commonly rely on average intercondylar distance values and standardized mounting procedures, limiting the simulation of patient-specific mandibular movement.

The purpose of this in vitro study was to evaluate the feasibility of estimating sagittal condylar inclination (SCI) from facial scan-derived eccentric mandibular positional data in a temporomandibular joint (TMJ) simulation model and to investigate whether intercondylar distance influences measurement behavior.

This study used a TMJ simulation model with intercondylar distance set at 5 levels (90, 100, 110, 120, and 130 mm) and SCI set at 4 levels (0, 15, 30, and 45 degrees). For each condition, facial scans at maximum intercuspation position (MIP), protrusive, right lateral, and left lateral positions were acquired as a single acquisition cycle, with 15 repeated cycles per condition. Eccentric scans were aligned to the MIP scan, and mandibular anterior transformation matrices were derived and mapped to the hinge axis to calculate SCI and the Bennett angle. Agreement between measured and actual SCI was evaluated using a mixed-effects Bland-Altman approach. Supportive analyses included linear regression, signed and absolute error, root mean square error (RMSE), mixed-effects modeling of SCI absolute error, and intraclass correlation coefficient (ICC) for repeatability.

The mixed-effects Bland-Altman analysis showed a negative overall bias (-1.07 degrees; 95% confidence interval (CI), -1.78 to -0.36), with broad individual-measurement variability. Supportive regression analyses showed slopes not significantly different from 1 and intercept confidence intervals including 0. Pooled mean ±standard deviation absolute error was 4.56 ±4.34 degrees, and RMSE was 6.29 degrees. For SCI absolute error, actual SCI was significant, whereas the side, intercondylar distance, and intercondylar distance×actual SCI interaction were not. Overall ICC was 0.886 (95% CI, 0.819 to 0.919). Descriptively, the Bennett angle showed significant negative correlations with intercondylar distance and positive correlations with actual SCI in most, but not all, distance conditions.

Based on the findings of this in vitro TMJ simulation study, facial scan- and eccentric position-based SCI estimation showed generally stable measurement performance across the evaluated conditions, with no statistically detectable systematic effect of intercondylar distance on SCI error under the evaluated simulator-based conditions.