To evaluate 3D changes in the condyle (CO) and glenoid fossa (GF) in growing patients with Class II malocclusion treated with the Twin Block (TB) appliance, using cone-beam computed tomography (CBCT) to compare pre-treatment (T0), end-of-treatment (T1: 12 months), and post-removal (T2: 12 months) time points, in order to assess post-treatment skeletal stability.

It was a retrospective longitudinal pilot study. Patients were recruited between August 2015 and 2022 and followed up until 2023. Twelve patients (11.92 ± 1.08 years; cervical vertebral stages CS3–CS4) were evaluated with CBCT at T0, T1, and T2. Complete 3D models at T0, T1, and T2 were superimposed in Dolphin Imaging software. Partial 3D models of the CO and GF were segmented in ITK-SNAP software. A coordinate system (x-, y-, and z-axes) was created for each complete 3D model at T0 using Geomagic Qualify software, followed by the alignment of the partial models of the CO and GF to the complete models. For each CO and GF, a 3D centroid was generated, and its displacement was evaluated across the T0–T1, T0–T2, and T1–T2 intervals. The non-parametric Wilcoxon and Friedman tests were used with a 95% confidence level.

Significant posterior displacement of the CO was observed during the T0–T1 interval (right: 1.33 mm; left: 1.54 mm), T1–T2 (right: 1.01 mm) and T0–T2 (right: 2.34 mm; left: 2.16 mm). Superior displacement of the CO was also significant during the three intervals. The GF showed significant posterior displacement at T0–T1 and T0–T2. The P -values were adjusted using the Bonferroni correction; 95% confidence intervals are reported in the results tables.

These preliminary findings suggest that TB therapy induces adaptive remodelling of both the CO and GF, with the CO changes exceeding those of the GF. The CO was displaced posteriorly and superiorly, whereas the GF was displaced posteriorly and inferiorly. However, the absence of a control group precludes a definitive distinction between treatment effects and normal craniofacial growth. Future controlled studies with larger samples are necessary to confirm these observations.