QUANTITATIVE EVALUATION OF SURGICAL FIDELITY IN VSP-GUIDED MANDIBULAR RECONSTRUCTION: A REGION-SPECIFIC ERROR ANALYSIS OF COMPLEX DEFECTS
Abstract
Resume. Objective: To quantify the three-dimensional spatial accuracy of virtual surgical planning (VSP)-guided fibula free flap mandibular reconstruction and to characterize the directional distribution of residual geometric errors across anatomical zones through axis-specific decomposition, with particular focus on secondary and complex defects. Materials and Methods: A retrospective analysis was conducted on 12 consecutive patients who underwent segmental mandibular resection and vascularized fibula free flap reconstruction between February 2024 and December 2025. A standardized CAD/CAM workflow was employed, incorporating patient-specific SLA cutting guides and pre-operatively adapted 2.4 mm titanium reconstruction plates contoured against stereolithographic resin mandibular models. Postoperative CBCT datasets were registered against the preoperative virtual plan in Materialise 3-matic software, with superimposition anchored on unoperated native mandibular regions. Deviations were decomposed along the mediolateral (X), superoinferior (Y), and anteroposterior (Z) axes. Patients were stratified by the Brown classification into less complex (Class I–II, n=6) and complex (Class III–IV, n=6) subgroups, and between-group differences were evaluated by independent samples t-test. Results: The mean global 3D deviation across the cohort was 3.40 ± 1.02 mm. No statistically significant difference was identified between the less complex (3.55 ± 1.12 mm) and complex (3.29 ± 1.10 mm) subgroups (P > 0.05), confirming that defect severity does not adversely affect overall reconstructive accuracy when a standardized digital workflow is employed. Axis-specific analysis, however, revealed pronounced and anatomically localized failure patterns not captured by global metrics. Posterior proximal segments (condyle and ramus) accounted for 83% of maximum deviations, with errors manifesting as two mechanically distinct transverse displacement patterns: medial collapse in five patients, driven by unopposed medial pterygoid contractile forces, and lateral flaring in four patients following condylectomy, with a maximum lateral displacement of 9.13 mm in the neocondyle case. Inferior displacement of the anterior arch was identified in seven of twelve cases along the Y-axis, with a maximum vertical drop of 6.98 mm attributable to gravitational loading from the fascio-cutaneous paddle combined with suprahyoid muscular traction. Posterior symphyseal drift of 10.38 mm was recorded in one case following hardware destabilization and unopposed digastric contraction. Signed mean deviations along the X- and Y-axes approached zero due to cancellation of opposing directional errors, whereas absolute mean deviations reached 4.50 ± 2.91 mm (posterior ΔX) and 3.78 ± 1.51 mm (anterior ΔY), demonstrating that aggregate global scores systematically mask clinically significant zone-specific displacements. Conclusion: The combined VSP, SLA cutting guide, and pre-bent plate protocol achieves reproducible global accuracy across a spectrum of defect complexities, including salvage reconstructions, at a cost-effective alternative to patient-specific milled implants. However, axis-specific analysis demonstrates that this protocol does not neutralize postoperative muscular forces on isolated bone segments. Proximal segment instability — manifesting as medial collapse or lateral flaring — and anterior arch inferior displacement constitute distinct, targetable failure modes requiring adjunctive intraoperative fixation strategies, including temporary maxillomandibular fixation and condylar seating maneuvers, beyond standard cutting guide application. Key words: Mandibular reconstruction; Virtual Surgical Planning (VSP); Fibula free flap; CAD/CAM; Stereolithography (SLA); Directional error analysis; Biomechanics; Condylar reconstruction; Brown classification; Medial collapse; Lateral flaring; Reconstruction plate; Osteotomy.