Early postoperative serum Tg measurement

TOPIC: Prognostic value of early postoperative serum Tg measured without TSH stimulation in DTC

Title: Thyroglobulin measurement before rhTSH-aided 131 I ablation in detecting metastases from differentiated thyroid carcinoma.

Authors: Giovanella L, Ceriani L, Suriano S, Ghelfo A, & Maffioli M.

Reference: Clinical Endocrinology 69, 659–663, 2008



The practice of routinely treating DTC with radioiodine (RAI) for remnant ablation during thyroid hormone withdrawal is being replaced by a more selective use of rhTSH-assisted RAI treatment, targeted to patients with a higher risk for recurrent disease. Whereas serum Thyroglobulin (Tg) measured when endogenous TSH is high during thyroid hormone withdrawal has been shown to have prognostic value, the prognostic significance of Tg measured during L-T4 suppression “ onT4-Tg ” prior to rhTSH-assisted RAI treatment, or in low-risk patients in whom RAI is deferred, is less certain.


To evaluate whether serum Tg, measured 4-6 weeks after total thyroidectomy during L-T4 suppression and prior to rhTSH-assisted RAI treatment, has prognostic value.

Design, Setting and Patients

This was a retrospective study of Tg measurements made during L-T4 suppression therapy (TSH < 0.05 mIU/L) in 126 TgAb-negative DTC patients 4-6 weeks after total thyroidectomy prior to rhTSH-assisted 131 I-treatment (3.700 MBq). Patients were analyzed according to whether disease was detected by a post-RAI treatment scan (PT-WBS). Patients with positive PT-WBS received further surgery and/or RAI treatment, whereas patients who had negative PT-WBS were restaged at 12 months by neck ultrasound, basal “onT4-Tg” and rhTSH-stimulated serum Tg and diagnostic RAI WBS (Dx-WBS). The presence of disease was confirmed by cytology and Tg measured in the FNA needle washout fluid.

Main Outcome Measures

ROC curve analyses of “onT4-Tg” measurements made pre-RAI and at a 12-month restaging.


27/126 (21%) of patients had persistent disease detected by PT-WBS. There was no difference in RAI uptake between patients with or without a positive PT-WBS. However, there was a significant relationship between RAI uptake and “onT4-Tg” in the patients with negative PT-WBS. The patients with disease, detected by PT-WBS, had significantly higher serum Tg than patients with negative PT-WBS (mean 6.0 versus 1.5 µg/L*, respectively). No patient with pre-ablation Tg below 0.4 µg/L* had a positive PT-WBS. At the 12 month restaging, most (87/99) of the patients with a negative PT-WBS had both basal (onT4-Tg) and rhTSH-stimulated Tg below 0.4 µg/L* and no evidence of disease by both neck ultrasound and Dx-WBS. However, 9 of the 12 patients who had a rising Tg detected at 12 months had disease detected: 6 with lymph-node metastases detected by cytology + Tg wash-out measurements, 2 with mediastinal disease detected by Dx-WBS and 1 with a RAI-negative recurrence detected by 18 FDG-PET/CT. A retrospective analysis revealed that pre-RAI “onT4-Tg” levels were higher for patients who had recurrence detected at 12 months compared with those with a negative 12 month restaging (mean 1.9 ± 3.8 versus 0.7 ± 2.2 µg/L*, respectively; p< 0.001). ROC curve analysis found that a pre-RAI “onT4-Tg” of 1.2 µg/L* identified patients with persistent/recurrent disease with 82.9% sensitivity, 55.2 specificity, 97.8% negative predictive value (NPV) and 43.3% positive predictive value (PPV). Furthermore, multivariant analysis showed that node status and postoperative “onT4-Tg” were independent prognostic factors.


Serum Tg measured during L-T4 suppression therapy can help stratify patient’s risk of persistent/recurrent disease and can factor into the decision whether to give RAI treatment.

* The Tg values cited above have been corrected to reflect 1:1 CRM-457 standardization to allow comparison with other assays.


Serum Tg measurement is still technically challenging. Most assays suffer from suboptimal sensitivity, differences in specificity for tumor-derived Tg isoforms and interferences from heterophilic antibodies (HAMA) and thyroglobulin antibodies (TgAb). Standardization issues still persist despite the major international effort launched in the 1990s that resulted in the production of an international standard (CRM-457). Some assays, such as that used in this study (BRAHMS DYNOtest®), are still not directly standardized against CRM-457 (the manufacturer recommends a correction factor of 2). It is critical for publications to cite Tg values that are 1:1 CRM-457 standardized in order to allow a comparison of Tg cut-off values and assay functional sensitivity limits. A failure to standardize Tg values against CRM-457 adds confusion for physicians and means that Tg cut-offs established by ROC curves cannot be directly applied in clinical practice.

It has become common practice to use RAI remnant ablation in the early post-operative period when endogenous TSH is elevated by thyroid hormone withdrawal. Studies have reported that endogenous TSH-stimulated Tg levels have prognostic value. Recently, recombinant human TSH (rhTSH) has been approved as a means to elevate TSH and eliminate the need for thyroid hormone withdrawal prior to RAI (see Bryan Haugen et al. in Thyroid, 2008). The use of rhTSH-assisted RAI necessitates the measurement of Tg in a low TSH state and raises the question whether serum Tg measured without TSH stimulation (“onT4-Tg”) retains its prognostic value. This is a topical question given the current controversy regarding the need for RAI of low-risk patients and the growing practice of deferring RAI for low-risk patients. If “onT4-Tg” measurements have prognostic value, they could be factored into the decision concerning the need for RAI treatment.

Because DTC patients have a relatively low incidence of persistent/recurrent disease, protocols for post-operative management need to have a high negative predictive value (NPV) in order to limit unnecessary testing and focus investigations on the minority of individuals at risk for persistent/recurrent disease. Previous studies have reported that “onT4-Tg” measurements are less diagnostically sensitive than TSH-stimulated Tg measurements. More recently the use of more sensitive Tg assays has been shown to improve the clinical sensitivity of “onT4-Tg”. After correcting to CRM-457 standardization* the Tg method used for this study was not especially sensitive (0.4 µg/L*), however ROC curve analyses still found that an “onT4-Tg” cut-off value of 1.2 µg/L* had a clinical sensitivity of 83% and a NPV of 98%. These parameters were only slightly lower than the author’s previous study using endogenous TSH-stimulated Tg in which a sensitivity of 94% and NPV of 99% was reported (see Luca Giovanella et al. in Clin Chem Lab Med, 2005). Furthermore the NPV of 98% reported for the current study is comparable to the NPV typical of a rhTSH-stimulated Tg using a cut-off of 2.0 µg/L*.

In contrast to the high NPV reported for this study, the PPV for “onT4-Tg” was low (43.3%), again similar to rhTSH-stimulated Tg studies. It was significant that a rising “onT4-Tg” detected at the 12-month restaging was highly predictive of disease (9/12 patients). This confirms other studies that have also reported a low PPV associated with the use of a fixed Tg cut-off and improvement in PPV by using the Tg trend (if measured by the same assay under the same TSH conditions).

A post-thyroidectomy 4 to 6 week TSH-suppressed serum Tg below 1.2 µg/L* has a high NPV (98%). This suggests that the “onT4-Tg” can be factored into decisions concerning the need for RAI treatment. However, the low PPV (43%) emphasizes the need for periodic monitoring using onT4-Tg (measured using a sensitive assay) together with ultrasound.

Summary and commentary prepared by Carole Spencer (Related to Chapters 6[a] & 18 of TDM)

Download PDF

Thank you for using and supporting THYROID MANAGER

One click download of a complete current PDF version of this chapter is available by payment of $5.00 (including sales tax) to ENDOCRINE EDUCATION / MDTEXT.COM,INC.

Please note:

You will be directed to a Paypal site for entering payment information, and then returned immediately to this site for delivery of the PDF download. If you do not wish to secure the PDF version, you are of course free to download the material directly from the chapter on our Website without charge.

We welcome comments on this service, and this charge, to- ldegroot@earthlink.net. May we note that we must secure income from advertisements and chapter downloads in order to continue providing our (otherwise) totally free, comprehensive, authoritative, constantly up-dated, Endocrinology web-book to the thousands of physicians and trainees around the world who visit the website each day of the year. We also welcome contributions.