The rate of occult nodal disease might be relatively low, and it does not frequently progress to clinical recurrent disease. Clinical nodal status might be valid for deciding the extent and radicality of node dissection. Prophylactic (central) lymphadenectomy should be carried out without radicality-associated morbidity. Macroscopic nodal disease warrants more rigorous, compartment-oriented lymphadenectomy. There is no rationale for detection of occult disease and micrometastasis by frozen section or immunohistochemistry.