Impact of Targeted Computed Tomography Evaluation on Detection of Mandibular Nonunion

Introduction: Bony nonunion occurs in 1-5% of patients undergoing VBG reconstruction of the mandible. CT identification of mandibular nonunion is crucial after VBG head and neck reconstruction, yet the reliability of this modality remains unclear. This study examines the accuracy of CT radiographic findings of nonunion associated with VBG mandibular reconstruction and whether accuracy improves when radiologists are specifically assessing for nonunion.

Methods: This retrospective observational study included patients who underwent revision surgery for mandibular hardware removal following VBG reconstruction between January 2000 and January 2024 at Ronald Reagan UCLA Medical Center. 60 charts met inclusion criteria, with 43 containing original CT films. Primary outcomes were radiologic findings and surgical presence of mandibular nonunion. A second radiologist, blinded to original radiographic and surgical findings, reread imaging studies while specifically assessing for nonunion. Interobserver agreement was determined using Cohen’s kappa. Using surgical nonunion as the reference standard, diagnostic performance of original and reread imaging was evaluated using sensitivity and specificity. Paired comparisons of sensitivity and overall diagnostic accuracy were performed using McNemar’s test.

Results: Of 60 charts, 43 underwent blinded reread. Agreement between original and reread radiographic findings was moderate (κ =0.50; 95% CI 0.26-0.74). Blinded reread significantly improved sensitivity for detection of nonunion from 36.8% (95% CI 16.3-61.6%) to 78.9% (95% CI 54.4-93.9%), with similar specificity (original: 54.2% [95% CI 32.8-74.4%]; reread: 58.3% [95% CI 36.6-77.9%]). Sensitivity and overall diagnostic accuracy were significantly higher with blinded rereads (McNemar's p = 0.021 and p = 0.012).

Conclusions: When radiologists are specifically asked to assess for mandibular nonunion, detection on imaging improves and may better inform surgical planning. However, findings should be interpreted with clinical assessment when determining the need for revision surgery.