Radiologic-Pathologic Correlation of Extranodal Extension in Patients With Squamous Cell Carcinoma of the Oral Cavity: Implications for Future Editions of the TNM Classification

To evaluate the accuracy and prognostication of the presence of radiologic extranodal extension (rENE) versus pathologic extranodal extension (pENE) in patients with oral cavity squamous cell carcinoma (OSCC). A retrospective review was conducted for all newly diagnosed OSCC patients who underwent neck dissection in our institution from 2010 to 2015 with available records of preoperative computed tomography or magnetic resonance imaging. Two head and neck neuroradiologists reviewed the presence of rENE (defined as ill-defined lymph node borders) on imaging independently, being blinded regarding the pathology report. The impact of the imaging-surgery interval, imaging modalities, and intrarater and interrater concordance of rENE was assessed. The diagnostic accuracy of rENE versus pENE was evaluated. Overall survival (OS) was compared between those with and without rENE. Multivariate analysis evaluated the prognostic value of rENE. Among the 508 patients, rENE and pENE were identified in 57 and 121 cases, respectively. The diagnostic accuracy of rENE versus pENE was identical (73%) for cases with the imaging-surgery interval ≤4 weeks (n = 276) and 4 to 8 weeks (n = 207) but lower (48%) for those >8 weeks (n = 25). Computed tomography displayed higher accuracy on rENE assessment versus magnetic resonance imaging (80% vs 63%, P = .011). Interrater and intrarater concordance (n = 93) was good (κ = 0.79) and excellent (κ = 0.94), respectively. Excluding the 25 cases with a >8 weeks imaging-surgery interval, the sensitivity, specificity, positive predictive value, and negative predictive value of rENE versus pENE in the remaining 483 cases were 52%, 96%, 93%, and 66%, respectively. Patients with rENE (n = 55) had inferior OS versus those without rENE (n = 202), and both had lower OS than node-negative (n = 226) patients (3-year OS: 31% vs 68% vs 81%, P < .001). Multivariate analysis, adjusted for age, T category, N category, and performance status, confirmed the prognostic value of rENE for OS (hazard ratio 3.3, 95% confidence interval 2.4-5.3, P < .001). This large cohort study shows a high specificity but low sensitivity of rENE for pENE. Similar to pENE, the presence of rENE is associated with reduced survival in OSCC.