A single recombinant human thyrotropin-stimulated serum thyroglobulin measurement predicts differentiated thyroid carcinoma metastases three to five years later

Context: Testing for residual differentiated thyroid carcinoma relies heavily upon recombinant human ( rh) TSH- stimulated serum thyroglobulin ( Tg) levels, but the positive predictive value is often low. Objective: Our objective was to determine the accuracy of a single rhTSH- Tg measurement over time. Design and Setting: We conducted a prospective follow- up study at the University referral center. Patients: A total of 107 differentiated thyroid carcinoma patients were stratified according to their initial rhTSH- Tg as follows: group 1 with Tg less than 0.5 ( n = 68), group 2 with Tg of 0.6 - 2.0 ( n = 19), and group 3 with Tg greater than 2 ng/ ml ( n = 20). Intervention: Clinical evaluations were conducted over 0.9 - 5.2 yr as follows: Tg during thyroid hormone suppression ( n = 27), after rhTSH ( n = 59), and/ or after thyroid hormone withdrawal ( n = 15). Main Outcome: Tumor was identified in one patient in each of groups 1 ( 1.6%) and 2 ( 5.5%), and 16 in group 3 ( 80%), comprising 19 tumor locations: 11 locoregional, two mediastinal, five lung, and one brain. Tumor was found in 81% with an initial or follow- up rhTSH- Tg greater than 2 ng/ ml. TSH- stimulated Tg fell spontaneously to less than 0.5 ng/ ml in 50% of group 2 and 5% of group 3 over 1.7 - 5.0 yr. The positive predictive value of the initial rhTSH- Tg greater than 2 ng/ ml was 80%, and the negative predictive value was 98%. After retreatment, 100% of group 1, 74% of group 2, and 55% of group 3 had no evidence of tumor ( P = 0.0001). Conclusions: 1) A single rhTSH- Tg greater than 2 ng/ ml predicts persistent tumor, although no value entirely excludes future recurrence. 2) Repeated TSH- stimulated studies are appropriate for patients at risk of recurrence, especially those with an rhTSH- Tg greater than 1 ng/ ml. 3) A single rhTSH- Tg less than 0.5 ng/ ml without Tg antibody has an approximately 98% likelihood of identifying patients completely free of tumor, a large group in which TSH suppression to less than 0.1 mIU/ liter and frequent imaging and TSH- stimulated Tg testing are unnecessary.