545 patients with differentiated thyroid carcinoma were followed up for periods ranging up to 25 years after first treatment (mean 8.1 years, 65% for over 5 years). 72% of patients with papillary carcinoma (n = 270), but only 52% with follicular carcinoma (n = 275) remained tumor-free during the further course. Residual malignancies persisted for more than the first year in 6% and 17% of patients respectively; there were tumor recurrences after an apparently tumor-free interval in 22% and 31% respectively, the latest after 12 and 27 years respectively. 6% and 19% of patients respectively died as a direct result of the tumor (and a group of equal size from other causes), half due to residual and half due to recurrent carcinoma. With regard to residual tumors, few significant risk factors were found preoperatively, comprising distant metastases (factor = 34 and 20 for papillary and follicular tumors respectively), age over 50 years (F = 6.4 and 5), infiltrating growth of primary tumor (F = 4 and 4.3), and regional lymph node involvement (F = 1.2 and 2). However, these factors were of little use in predicting the risk of the more frequently observed tumor recurrence, with maximum factors of 2 (for T4 and N+ stage) for papillary thyroid cancers and 1.5 for follicular cancers. At risk for recurrence were patients in whom total thyroidectomy was not performed (F = 2.3 and 2) and those who did not receive postoperative radioiodine treatment (F = 3), irrespective of age and tumor stage. Therefore, any individualizing regimen beginning with the first treatment has a bearing not only on residual tumor's 50% contribution to mortality. The equally large contribution of recurrences to tumor death can be influenced only by thyroidectomy or, more realistically, by strumectomy combined with early ablation of thyroid remnants with radioiodine. Postoperative radiotherapy of the neck region did not prevent tumor recurrence, and although hormonal suppression was never given the results compared well with the best of published long-term follow-up studies. There were no acute or late complications that could be ascribed to radioiodine treatment. However, a strict strategy of the reducing the administered doses was adopted: the ablation dose was half that used previously (1.5 GBq, i.e. 45 mCi on average), tumor treatment was halted even where residual uptake was observed scintigraphically (in 44% of patients treated) and radioiodine was no longer used for follow-up investigations.(ABSTRACT TRUNCATED AT 400 WORDS)