Risk Stratification for Extranodal Extension in Head and Neck Cancers-Implication for Treatment Intensification.

Abstract

Background: Extranodal extension (ENE) is an important adverse prognostic indicator for head and neck cancers. However, ENE needs further risk stratification in terms of optimum cutoff for the extent of ENE and interaction with clinicopathological risk factors.

Methods: Retrospective single-center study on patients with treatment-naïve head and neck cancer with final histology indicating ENE.

Results: Pathological ENE was observed in 122 patients (12.4%). With the 2 mm cutoff separating Mi-ENE and Ma-ENE, no difference was observed in overall survival (OS) and disease-free survival (DFS) for unmatched (OS: 40.9% vs. 33.6%, p = 0.7; DFS: 34.05% vs. 26.12%, p = 0.5) as well as propensity score-matched cohort (OS: HR 1.08, p = 0.82; DFS: HR 0.95, p = 0.89). Receiver-operator curve (ROC) analysis showed the highest area under the curve with a 4 mm cutoff for the extent of ENE (AUC: 0.52). On assessing the impact of adjuvant chemoradiation (CRT) on Ma-ENE and Mi-ENE, only Ma-ENE showed OS (HR: 0.42, 95% CI: 0.18-0.9) and DFS (HR: 0.33, 95% CI: 0.15-0.70) benefit with CRT, which was statistically significant. The revised 4 mm cutoff was predictive of therapeutic benefit with adjuvant CRT (HR: 0.27, 95% CI: 0.1-0.73). Clinicopathological factors with statistically significant interaction with ENE in worsening OS and DFS were tumor location in tongue/floor of mouth (OS), T3/T4 category (OS), depth of invasion greater than 10 mm (OS), ≥ 5 nodes with metastasis (OS), and male sex (DFS).

Conclusion: The standard 2 mm cutoff for the extent of ENE failed to reveal sufficient hazard discrimination for OS/DFS. Instead, the 4 mm cutoff determined on ROC analysis was found to have the best predictive ability for DFS.