Lymph Node Clustering in Head and Neck Squamous Cell Cancer

Objective Detection of early regional metastasis of head and neck carcinoma is critical for tumor staging, prognosis, and treatment strategies. Clustered cervical lymphadenopathy portends a worse prognosis than isolated lymphadenopathy, but studies analyzing its effect are lacking. Key objectives include: 1) to establish criteria for lymph node clustering and 2) to assess the predictive value of lymph node clustering for metastasis. Methods This study retrospectively reviewed preoperative radiographic images of 29 patients with histopathologically proven metastatic head and neck squamous cell carcinoma between January 2006 and December 2007. Patients who had previous radiation or chemotherapy were excluded. CT, MRI, and PET CT images were assessed for lymph node size and number, neck level, and clustering, with respect to primary tumor location and size. A cluster was defined as 3 or more abutting nodes with no definable intervening fat planes. Results Statistics comparing histopathologic proven metastatic lymphadenopathy and radiographic clustering of nodes in each neck level were used to elucidate the positive predictive value via logistic regression analysis. Analysis revealed a positive predictive value for clustering of nodes greater than 1 cm to be 82.9%, with a negative predictive value of 100% in the level II region (p<0.05). Nodes in the remaining levels demonstrated less predictive values. Conclusions Clustered lymph nodes greater than 1cm in the level II region in head and neck squamous cell carcinoma have a high predictive value, suggesting that clustering is not only an important prognostic indicator but also an important radiographic feature that may assist surgeons in preoperative surgical planning.