Midline involvement and perineural invasion predict contralateral neck metastasis that affects overall and disease-free survival in locally advanced oral tongue squamous cell carcinoma.

Abstract

Although patients with oral squamous cell carcinoma who develop contralateral neck metastasis (CLNM) have worse survival outcomes than those without CLNM, accurate prediction of occult CLNM in clinically negative contralateral neck (contralateral cN0) remains difficult. This study aimed to identify clinicopathological factors that could reliably predict CLNM in patients with locally advanced (clinical T3 and T4a) tongue squamous cell carcinoma (TSCC). The medical data of 32 patients with cT3-4a TSCC who underwent curative surgery between 2010 and 2017 were retrospectively analyzed. The correlation of clinicopathological variables with CLNM was examined using logistic regression analysis. The diagnostic performance of significant variables was evaluated using the area under the receiver operating characteristic curves (AUC). Overall survival (OS) and disease-free survival (DFS) were assessed using a Cox proportional hazards model. CLNM was eventually confirmed in 11 patients (34.4%). Multivariate logistic regression showed that midline involvement [odds ratio (OR) = 23.10, P = 0.017] and perineural invasion (PNI, OR = 14.96, P = 0.014) were independent predictors of CLNM. Notably, the prediction model comprising a combination of midline involvement and PNI exhibited superior diagnostic performance with an even higher OR of 80.00 (P < 0.001), accuracy of 90.3%, and AUC of 0.876. The multivariate Cox hazards model revealed independent significance of CLNM as an unfavorable prognostic factor for both OS [hazard ratio (HR) = 5.154, P = 0.031] and DFS (HR = 3.359, P = 0.038), as well as that of PNI for OS (HR = 5.623, P = 0.033). Our findings suggest that coexisting midline involvement and PNI of the primary tumor is highly predictive of CLNM development, which independently affects both OS and DFS in patients with locally advanced TSCC. Such reliable prediction enables efficient control of CLNM by optimizing management of the contralateral cN0 neck, which will likely contribute to improved prognosis of those patients without unnecessarily compromising their quality of life.