TOPIC: Methodological issues with serum thyroglobulin measurements
Title: Comparison of seven serum thyroglobulin assays in the follow-up of papillary and follicular thyroid cancer patients
Authors: Schlumberger M, Hitzel A, Toubert ME, Corone C, et al.
Reference: Journal of Clinical Endocrinology & Metabolism 92:2487-2495, 2007
Serum thyroglobulin (Tg) is the most sensitive and specific marker of differentiated thyroid cancer after initial treatment and TSH stimulation increases its sensitivity for the diagnosis of recurrent disease. Variability in the sensitivity and reproducibility of different commercial assays makes serum Tg measurement a delicate issue when interpreting the results. The goal of this study is to compare the diagnostic values of seven methods for serum Tg measurement for detecting recurrent disease both during L-T 4 treatment and after TSH stimulation.
Thyroid cancer patients with no evidence of persistent disease after initial treatment were studied at 3 months on L-T 4 treatment (Tg1) and then at 9-12 months after withdrawal or recombinant human TSH stimulation (Tg2). Sera with anti-Tg antibodies or with an abnormal recovery test result were excluded from Tg analysis with the corresponding assay. The results of serum Tg determination were compared to the clinical status of the patient at the end of follow-up. All measurements were performed using seven commercial Tg assays with different functional sensitivity: 0.9 ng/ml (3 assays), 0.3-0.2-0.11-0.02 ng/ml (1 kit, respectively).
Thirty recurrences were detected among 944 patients. A control 131 I total body scan had a low sensitivity, a low specificity, and a low clinical impact. Assuming a common cut-off for all Tg assays at 0.9 ng/ml, sensitivity ranged from 19-40% and 68-76% and specificity ranged from 92-97% and 81-91% for Tg 1 and Tg2, respectively. Using assays with a functional sensitivity at 0.2-0.3 ng/ml, sensitivity was 54-63% and specificity was 89% for Tg1. Using the two methods with a lowest functional sensitivity at 0.02 and 0.11 ng/ml resulted in a higher sensitivity for Tg1 (81% and 78%), but at the expense of a loss of specificity (42% and 63%); finally, for these two methods, using an optimized functional sensitivity according to receiver operating characteristic curves at 0.22 and 0.27 ng/ml resulted in a sensitivity at 65% and specificity at 85-87% for Tg1.
Using an assay with a lower functional sensitivity may give an earlier indication of the presence of Tg in the serum on L-T 4 treatment and may be used to study the trend in serum Tg without performing any TSH stimulation. Serum Tg determination obtained after TSH stimulation still permits a more reliable assessment of cure and patient-s reassurance.
Serum Tg is recognized as the most sensitive and specific marker of residual disease in differentiated thyroid cancer patients treated with total thyroidectomy and 131-I ablation. Since suppressive l-T4 therapy may reduce serum Tg concentrations even to undetectable levels (false negative results), stimulation by endogenous or exogenous TSH may be required to increase the diagnostic value of serum Tg measurement. Such stimulation is particularly required in patients with undetectable basal serum Tg at the time of the first control (usually 8-12 months after initial treatment).
The study by Schlumberger et al. compared the diagnostic accuracy of seven different commercial Tg kits both during l-T4 treatment and after TSH stimulation (endogenous or exogenous). Using different kits resulted in different sensitivity in detecting residual disease, both on l-T4 and after TSH stimulation, but with any kit, the results confirmed low sensitivity of serum Tg during l-T4 therapy and better sensitivity after TSH stimulation. The results also confirmed that many patients with detectable Tg concentrations are not associated with any detectable disease (false positive Tg? or disease that is to small to be localized?).
The novel and important finding of the study is that using new kits with a lower sensitivity (ultrasensitive assays) allowed an earlier discovery of detectable Tg in the serum and eventually the localization of persistent disease by imaging modalities, but, on the other hand, increased the number of patients with detectable Tg and no evidence of disease (lowering the specificity of the test). Thus, for the moment it is not possible to advocate the use of ultrasensitive Tg assays in alternative to TSH stimulation, which still permits a more reliable assessment of cure.
Related to Chapter 18.