Is It Safe to Omit Neck Dissection of Level Ib for the Treatment of Clinically Node-Negative Human Papillomavirus-Negative Oropharyngeal Squamous Cell Carcinoma?

BACKGROUND

Selective neck dissection (SND) of levels II-IV is traditionally recommended for the treatment of clinically node-negative (cN0) oropharyngeal squamous cell carcinoma (OPSCC). The necessity of including level Ib in cN0 neck dissection remains debated.

PURPOSE: The purpose of this study was to estimate and compare neck recurrence (NR) rates and survival outcomes between subjects with cN0 human papillomavirus-negative OPSCC who underwent SND including level Ib and excluding level Ib.

STUDY DESIGN, SETTING, SAMPLE: This was a single-center retrospective cohort study conducted at a population-based hospital named Ninth People's Hospital affiliated to Shanghai Jiao Tong University School of Medicine. The study included subjects with cN0 human papillomavirus-negative OPSCC who underwent primary resection and SND. Exclusion criteria were prior induction chemotherapy. Subjects in group II-IV were matched to those in group Ib-IV at a 1:2 ratio using iterative expanding radius matching on pathological T/N classification and age.

EXPOSURE: The exposure variable was SND and it was divided into 2 groups: I-IV group and Ib-IV group.

MAIN OUTCOME VARIABLE: The primary outcome variable was time to NR.

SECONDARY OUTCOMES: overall survival, disease-free survival, and disease-specific survival.

COVARIATES: Covariates were age, sex, subsites, T/N classification, depth of invasion, histologic grade, lymphovascular invasion, perineural invasion, radiotherapy, chemoradiotherapy, and immunotherapy/target therapy.

ANALYSES: Kaplan-Meier for survival; Cox proportional hazard models for NR risk factors. P < .05 was statistically significant.

RESULTS

After matching, the final sample comprised 198 subjects. For the entire cohort, the 5-year NR incidence was 7.6% (5/66) in group II-IV and 2.3% (3/132) in group Ib-IV (P = .074). In the multivariable model, SND was not associated with NR (hazard ratio: 3.26; 95% CI: 0.74 to 14.46, P = .1). For the tumor from the base of the tongue, NR was higher in group II-IV (12.2%, 5/41) than in group Ib-IV (2.1%, 2/95, P = .015). In the multivariable model, SND was not associated with an increased risk for NR (hazard ratio: 9.36; 95% CI: 1.36 to 64.34; P = .088).

CONCLUSIONS AND RELEVANCE: NR risk was higher in subjects with SND excluding level Ib than those including it, but the difference is not statistically significant.