Traditional risk factors and nodal yield—still relevant with high-quality risk-adapted adjuvant treatment for locally advanced head and neck cancer?

Patients with locally advanced head and neck cancer (LAHNC) often undergo multimodal therapy including radical resection of the primary tumor and neck dissection (ND) followed by risk-adapted adjuvant radio(chemo)therapy (R(C)T). Quality parameters influencing local control and survival of these patients have been postulated: resection status (R status), extranodal extension (ENE), interval to adjuvant treatment ≤6 weeks, R(C)T given when indicated, and nodal yield (NY) ≥18 lymph nodes per neck. For other solid tumors the trend is towards less extensive lymph node surgery to avoid toxicity such as lymphedema, damage to peripheral nerves, dysesthesia, or paresthesia. The present study aims to investigate whether the number of nodes removed during neck dissection for LAHNC is still predictive for outcome when patients receive risk-adapted adjuvant treatment according to current guidelines.

Between 2008 and 2015, 468 patients with LAHNC undergoing R(C)T with curative intent were prospectively registered in a database (UICC III/IV). Among them, 359 patients received adjuvant treatment and 295 underwent neck dissection. There were 119 (40%) patients with an oropharyngeal primary, 49 (17%) with cancer of the larynx/hypopharynx, 88 (30%) of the oral cavity, and 39 (13%) of the nasal/paranasal sinuses and cancer of unknown primary (CUP). Median follow-up was 45.6 months. Histopathology revealed an R1 status in 65 (22%) cases and ENE in 93 (31%) cases. 150 (51%) patients received RCT; the median time to adjuvant treatment from the day of tumor resection was 44 days (35–54) and overall treatment time (OTT; time from surgery to the last day of R(C)T) was 90 days (82–101). Factors influencing disease-free survival (DFS) were adjusted and analyzed using CART analysis (removed nodes, number of positive nodes, body mass index (BMI), ENE, T and N classification, R status, and primary site). Local control (LC), distant metastases-free survival (DMFS), and overall survival (OS) were analyzed using Kaplan–Meier statistics and multivariate analysis (MVA) for factors predictive for DFS and OS.

CART analysis (Classification and Regression Trees) showed that T classification (T3/4) is the most important predictor for DFS, followed by age (> 61 years) and BMI (< 17.4). Primary site (OPC vs. other) and number of removed nodes (< 17) were shown to be less important for DFS, while ECE, N classification, and R status seem to be of little relevance. MVA revealed number of positive nodes, non-OPC, and T3/4 to be negative predictive factors for DFS. For OS, the number of positive nodes and non-OPC primary were predictive. Five-year rates were 86.1% for LC, 87.9% DMFS, 76.5% DFS, and 67.2% for OS.

In this patient cohort, the number of removed nodes is not relevant for DFS and OS, while the number of positive nodes and T classification have a negative impact on these endpoints. The high-risk factors positive resection margin and ECE seem to lose their negative impact on DFS and OS. High-quality care in head and oncology is only possible within a close multidisciplinary team and network.