Impact of lymph node dissection extent on disease-free survival with postoperative nivolumab plus concurrent chemoradiotherapy in head and neck squamous cell carcinoma: A post-hoc analysis of the NIVOPOSTOP trial.

6001 Background: Postoperative nivolumab added to concurrent Chemo-Radiotherapy (CRT) after surgery has been shown to improve disease-free survival (DFS) in patients with resected head and neck squamous cell carcinoma (HNSCC) at high risk of relapse. However, extensive nodal dissection has long be suspected to hinder response to immunotherapy but prospective data remains scarce. Methods: We analyzed the surgical procedures of the patients included in the primary analysis of the NIVOPOSTOP study regarding neck lymph node dissection (LND). Patients were classified with either uni- or bilateral LND. The number of nodes in the pathological report was also studied. The extent of the LND on DFS was analyzed in univariate and multivariate analysis. Results: Surgical information was available for all 666 patients. Four patients did not undergo any LND. Of the 662 patients who underwent nodal surgery, 239 (36%) underwent unilateral LND and 423 (64%) bilateral LND. The proportion of patients with bilateral / unilateral LND was similar in both treatment. Patients who underwent bilateral LND had statistically higher stage tumors versus unilateral LND (76% vs 61% stage IV) and a significantly higher proportion of laryngeal and hypopharyngeal tumors (17% vs 3% and 16% vs 7% respectively). The median numbers of nodes removed were 39 overall, 25 on the right side of the neck, and 24 on the left. As compared to unilateral LND, bilateral LND was associated with worse DFS in univariate analysis (HR 1.56 (95%CI 1.18; 2.05)). After adjusting for performance status, tumor site and p16 status, clinical stage, pathological risk factors of relapse (nodal extracapsular extension, margin status, perineural invasion, ≥ 4 involved nodes), the association was no longer statistically significant: HR 1.26 (95%CI 0.92; 1.71), Wald test p-value 0.15. There was no interaction between the type of LND (unilateral or bilateral) and the type of treatment (without or with nivolumab) on DFS. The benefit of adding Nivolumab to CRT was similar for unilateral LND (HR 0.79 (95%CI 0.50; 1.26)) and bilateral LND (HR 0.77 (95%CI 0.57; 1.03)) in Cox model stratified for p16 status. Regarding the extent of LND, the benefit of adding Nivolumab to CRT in the 330 patients who underwent removal of more than 39 neck lymph nodes, on one or both sides, was similar to that of the whole population (HR 0.75 (95%CI 0.54; 1.05)). Conclusions: The DFS benefit of adding nivolumab to standard postoperative therapy (cisplatin-RT) was not changed by whether the LND was bilateral or unilateral and persisted in patients in whom more than 39 cervical lymph nodes were removed, on one or both sides. As such, no evidence supports reducing the extent of neck LND when immunotherapy is incorporated into the management of high risk resected HNSCC. Clinical trial information: NCT03576417 .