Predictive value and dynamic risk stratification of high sensitive basal or stimulated thyroglobulin assay in a long-term thyroid carcinoma cohort

Abstract

Abstract Purpose To evaluate the predictive value of the rhTSH thyroglobulin stimulation test (rhTSH-Tg) compared to basal high-sensitive thyroglobulin (hs-Tg) under TSH suppressive therapy at 12 months after the completion of initial treatment to predict the long-term response and Dynamic Risk Stratification (DRS) at the last follow-up visit in a long-term DTC cohort. Methods Prospective study in 114 DTC patients (77.2% women, mean age 46.4 ± 14.1 years old, median/IQR evolution 6.7[3.1–8.0] years) from 2013 to 2020 undergoing total thyroidectomy and radioiodine ablation in whom hs-Tg and rhTSH-Tg was performed 12 months after completing initial treatment. Pearson correlation, receiving operating characteristics (ROC) and DRS at initial and last follow-up visit were analyzed. Results hs-Tg and rhTSH-Tg show a strong positive linear correlation (r = 0.864, p < 0.001). The diagnostic performance of initial hs-Tg and rhTSH-Tg levels were evaluated via ROC-AUC as a predictor of excellent response (ER) in the last follow-up visit. Hs-Tg showed a better AUC (0.969, 95%CI = 0.941–0.997) than rhTSH-Tg (0.944, 95%IC = 0.905–0.984; p < 0.001). The hs-Tg and rhTSH-Tg cutoff point of highest sensitivity (S) and specificity (E) was 0.110 and 0.815 ng/dl, respectively. Hs-Tg showed a higher diagnostic accuracy than rhTSH-Tg (S = 100% vs 96.8%, E = 84.3% vs 84.3%, NPV = 100% vs 98.6%, PPV = 70.5% vs 69.7%; p < 0.05). The DRS based on initial hs-Tg showed better ability to predict ER (93.3% vs 86.7%) and biochemical incomplete response (53.3%vs13.3%) in the last follow-up visit compared to rhTSH-Tg. Conclusions Both initial hs-Th and rhTSH-Tg were good predictors of long-term ER. In patients with hs-Tg, the rhTSH-test did not provide relevant prognosis information. An ER after initial treatment was associated with a very high NPV at subsequent follow-up.