Because prospective, randomized studies on papillary carcinoma of the thyroid are lacking, the results of retrospective studies from the basis of our present knowledge. To assess differences in treatment modalities, the patient groups must be comparable. A prognostic score has been devised by the use of multivariate analysis which can accurately distinguish patients at various levels of risk for disease-related mortality. To obtain the score, the calculated contributions of four significant variables are added. These variables are Age, tumor Grade, Extent, and Size (AGES). Using this scoring system, locally recurrent papillary thyroid carcinoma was studied in 816 patients primarily treated at the Mayo Clinic between 1946 and 1970. Locally recurrent disease developed in 54 (6.6%) patients, with the highest risk of occurrence in the first five years following initial thyroidectomy. Both in low- and high-risk patients, patients who had undergone unilateral thyroid resection had significantly more local recurrences than those with bilateral resections. However, there was no such difference when comparing subtotal to total thyroidectomy. Even though local disease was the cause of death in over one-half of the patients who died of thyroid cancer, this represented less than two percent of the entire series. From analysis of these data, bilateral resection - but not necessarily total thyroidectomy - with specific care to preserve both recurrent laryngeal nerves and at least one parathyroid gland in situ, seems to provide optimal treatment. When recurrence does occur, it appears worthwhile to attempt further surgical resection.