Imaging modalities for detecting jaw bone invasion in oral squamous cell carcinoma: a systematic review and diagnostic test accuracy meta-analysis

To evaluate the diagnostic accuracy of preoperative imaging modalities for detecting jaw-bone invasion in oral squamous cell carcinoma (OSCC), with particular emphasis on studies using histopathology as the reference standard. PubMed, Scopus, and Web of Science were searched. Diagnostic accuracy was assessed using bivariate random-effects models and hierarchical summary receiver operating characteristic curves. Imaging modalities were compared using meta-regression. Additional analyses were restricted to OSCC-only cohorts and studies with histopathological verification. Thirty-eight studies contributed 78 modality-specific diagnostic datasets. Computed tomography (CT; 30 datasets) yielded pooled sensitivity and specificity of 78.7% and 84.2%, respectively, compared with 84.3% and 81.6% for magnetic resonance imaging (MRI; 21 datasets). Single-photon emission computed tomography (SPECT; nine datasets) showed high sensitivity (96.7%) but lower specificity (57.5%). Corresponding estimates were 88.6% and 77.1% for cone-beam CT (five datasets), 80.2% and 81.5% for PET/CT (six datasets), and 65.7% and 90.0% for panoramic radiography (seven datasets). In analyses restricted to OSCC-only cohorts with histopathological verification, sensitivity and specificity were 78.9% and 81.7% for CT (17 datasets) and 87.2% and 75.4% for MRI (15 datasets). The formal CT comparison did not establish a significant difference between OSCC-only and mixed-site cohorts. CT and MRI showed consistent diagnostic performance for jaw-bone invasion. SPECT was highly sensitive but produced more false-positive findings. Evidence for CBCT and PET/CT remains limited. Imaging selection should consider the clinical question and the consequences of false-positive and false-negative results.