Font Size: 

Abstract Metastasis is the process through which cancer extends to different sites in the body and it is the main cause of death in individuals suffering from cancer. Detection of metastasis at early stages is difficult but central for appropriate treatment. The currently adopted methods in the prognosis of head and neck squamous cell carcinomas (HSCC), such as staging and grading, do not allow for a sufficiently and more reliable prediction of disease progression, recurrence, and metastasis formation. In this paper, we propose robust statistical methods to determine a gene signature of metastasis from a follow-up study of 117 patients with primary head and neck squamous cell carcinomas. Results suggest a list of 107 genes which may improve the accuracy of prediction of metastasis with respect to staging and grading.