Perineural Invasion and Lymph Node Ratio Quartile Are Associated With Extranodal Extension in Oral Cavity Squamous Cell Carcinoma.

Abstract

Background: In contrast to extranodal extension (ENE), tumor characteristics can often be evaluated preoperatively in patients with oral cavity squamous cell carcinoma (OCSCC). If correlations exist between primary tumor characteristics and the presence of ENE, evaluating these factors could help provide more accurate patient counseling and adjuvant treatment planning before undergoing surgical therapy.

Purpose: The study's purpose was to identify associations between OCSCC pathologic characteristics and ENE.

Study design, setting, sample: We performed a retrospective cohort study of adult patients who underwent upfront curative-intent surgery for OCSCC at the University of Iowa Hospitals and Clinics from 2004 to 2018. The treating service was Otolaryngology-Head and Neck Surgery. Exclusion criteria included patients who did not undergo neck dissection, primary tumor occurring outside the oral cavity, pathologic N0, cancer previously treated with radiation or surgically at another institution, and gross disease remaining after surgery.

Predictor variable: The predictor variable was a set of tumor characteristics, including oral cavity subsite, T- and N- classification, perineural invasion (PNI), lymphovascular invasion, bone invasion, and positive lymph node ratio (LNR).

Main outcome variable: The main outcome variable was ENE status, defined as positive (at least 1 lymph node reported to have ENE) or negative, identified from pathology reports.

Covariates: Covariates included subject sex, age, and smoking history.

Analyses: Bivariate comparisons and multivariate logistic regression analyses were performed to identify correlations between predictor variables/covariates and presence of ENE. Statistical significance was set at P = .05.

Results: The sample comprised 233 subjects with a mean age of 60.5 (SD 12.5) years, and 154 (66.1%) were male. Of 233 subjects with nodal metastasis, 122 (52.4%) had ENE in at least 1 node, and the median (interquartile range) positive LNR was 0.083 (0.094). On bivariate analysis, PNI (relative risk = 1.49; 95% CI, 1.14 to 1.96; P = .002), bone invasion (relative risk = 1.42; 95% CI, 1.13 to 1.80; P = .005), LNR quartile (P < .001), and pathologic T-class (eighth edition; P = .001) were significantly correlated with ENE. On multivariate logistic regression analysis, PNI (odds ratio = 2.29; 95% CI, 1.21 to 4.31; P = .01) and LNR quartile (P < .001) remained significantly correlated with ENE.

Conclusions and relevance: This study offers insight into important clinicopathologic details of lymph node metastases in OCSCC with an emphasis on tumor characteristics and odds of ENE in 1 or more lymph nodes. This information may be beneficial for patient counseling and treatment planning, especially in patients with PNI and high positive LNR but no evidence of obvious ENE before surgery.