Reducing the need for rhTSH-testing in the follow-up of patients with differentiated thyroid carcinoma

Topic: Sensitive Tg assays and recombinant human TSH (rhTSH)

Title: Monitoring thyroglobulin in a sensitive immunoassay has comparable sensitivity to recombinant human TSH-stimulated thyroglobulin in the follow-up of thyroid cancer patients.

Authors: Smallridge RC , Meek SE , Morgan MA , Gates GS , Fox TP , Grebe S , & Fatourechi V

Reference: Journal of Clinical Endocrinology & Metabolism, 92: 82-87, 2007



Most thyroglobulin (Tg) assays have functional sensitivities in the 0.5 to 1.0 ng/mL range. Patients with residual disease and undetectable serum Tg reported by such assays sometimes exhibit a detectable Tg after rhTSH-stimulation. Guidelines recommend that patients with TSH-stimulated serum Tg above 2 ng/mL should have imaging studies to exclude the presence of disease.


To determine whether there was still a need for rhTSH-stimulation when a sensitive Tg assay (with a functional sensitivity of 0.1 ng/mL) was used.


Eighty (out of 194) consecutive TgAb-negative patients undergoing treatment for differentiated thyroid carcinoma at two different Mayo Clinic facilities, selected as having a baseline Tg below 0.1 ng/mL (during l-T4 treatment) and a rhTSH-stimulated Tg test result.


Charts were retrospectively reviewed over a 4.5 year period of using a sensitive Tg assay. Patients were selected who had baseline and rhTSH-stimulated Tg test results, follow-up evaluations and imaging studies to determine the presence of disease.


Only 2 of the 80 (2.5%) patients with a baseline Tg <0.1 ng/mL had a positive rhTSH-stimulated Tg response (>2 ng/mL) and neither of these patient had evidence of disease by imaging studies and follow-up. Only one patient with a baseline Tg below 0.1 ng/mL had a recurrence, and this was detected by ultrasound and not rhTSH-stimulation testing (rhTSH-stimulated Tg: 0.3 ng/mL).


When a sensitive Tg assay is used, the costs of rhTSH-testing can be dramatically reduced because patients with baseline Tg below 0.1 ng/mL rarely have disease and usually have a negative rhTSH-stimulated Tg response (< 2 ng/mL).


Conventional Tg assays have suboptimal sensitivity for detecting disease in thyroidectomized patients. This is evident from the proximity of the lower assay reference limit for euthyroid individuals with intact thyroid glands (2-4 ng/mL) to the typical assay functional sensitivity (FS) limit of 0.5-1.0 ng/mL. Since recombinant human TSH (rhTSH) became available in 1999, it has become common practice to use rhTSH stimulation to compensate for Tg assay insensitivity, much as TRH-stimulated TSH testing was used to overcome TSH assay insensitivity before the current sensitive 3rd generation immunometric assays became available. Analogous with TRH-stimulated TSH testing there is a strong relationship between the basal and stimulated Tg levels, so that when basal Tg (measured on LT4 Rx.) can be measured reliably, stimulation testing becomes unnecessary. Tg assays with ten-times greater sensitivity (FS <0.1 ng/mL) than conventional assays became available about five years ago. One of these assays was used for the present retrospective study of rhTSH-stimulated Tg tests. When rhTSH was developed to replace the need for thyroid hormone withdrawal prior to radioiodine whole body scanning (WBS) the rhTSH-stimulated Tg response was considered to be a secondary endpoint. Currently, diagnostic WBS is being rapidly replaced by ultrasound imaging which has been shown to be more diagnostically sensitive for detecting lymph node disease. RhTSH is primarily used for rhTSH-stimulated Tg testing in which rhTSH-stimulated Tg values above 2 ng/mL are considered a risk factor for disease. This study focused on the cohort of 80 patients with an undetectable (<0.1 ng/mL) basal Tg online pokies machines measured by sensitive assay. Most (59%) of these patients also had an undetectable rhTSH-stimulated Tg. Only two patients (2%) had a positive rhTSH-stimulation test (rhTSH Tg > 2 ng/mL) and neither of these patients had detectable disease. In one patient, lymph node disease was detected by ultrasound but not rhTSH stimulation (rhTSH-Tg = 0.3 ng/mL) in accord with other studies showing that ultrasound is often more diagnostically sensitive than rhTSH-stimulated Tg testing. The study clearly demonstrates that when the Tg assay can reliably establish that the basal Tg is below 0.1 ng/mL, rhTSH-stimulation becomes unnecessary. The lower the basal Tg level, the greater the likelihood of having a negative rhTSH-stimulation test, in accord with other studies reporting that there is a strong relationship between basal and rhTSH-stimulated Tg (6 to 10 fold stimulation). The economic impact of rhTSH-stimulation testing is considerable! The authors of this study estimated that the cost of an rhTSH stimulation test was close to $2000. Given the low risk of disease in patients with basal Tg values below the conventional assay limit of ~ 1.0 ng/mL, and the fact that low-risk patients have primarily local or lymph node disease that can be detected by ultrasound, it is clear that the cost-benefit ratio of using rhTSH stimulated Tg testing needs to be reassessed.

Diagnostic sensitivity and specificity depends on the Tg cut off used to define disease and the imaging modalities employed for detecting disease. Current thyroid cancer guidelines include -an undetectable Tg- as a parameter for assigning a disease-free status. There has been resistance to using more sensitive Tg assays because many patients thought to have an -undetectable- Tg when using assays with a functional sensitivity of ~1.0 -g/L (ng/mL) have Tg detected in the 0.1 to 1.0 -g/L (ng/mL) range when more sensitive assays are used, as was the case with this study. Patients who have not received adjuvant radioiodine (RAI) treatment are the especially likely to have low levels of Tg detected by these more sensitive assays. In fact, one justification for using RAI ablation has been to -eradicate- residual remnant tissue and render Tg measurement more specific for tumor. This rationale is flawed given that normal remnant tissue often persists even after high dose RAI administration. Furthermore, the use of routine adjuvant RAI treatment is declining, following recent retrospective studies showing a lack of efficacy of RAI treatment of low-risk patients who constitute the majority of cases of DTC . It follows that as more sensitive Tg assays are adopted it will become necessary to accept that a low but detectable Tg level [<1.0 -g/L (ng/mL)] is probably clinically inconsequential, and instead use the trend in serum Tg (measured during L-T4 suppression of TSH) instead of the mere presence of Tg, as a post-operative tumor marker. A number of studies report that serum Tg falls spontaneously following thyroidectomy (during L-T4 treatment) even when patients do not receive RAI. This presumably represents atrophy of TSH-dependent tissue (normal remnant or tumor) in response to a reduction of the trophic influence of TSH. It follows that a rising trend in Tg in the face of a chronically low TSH provides a better marker for tumor growth and recurrence than whether the Tg is -detectable-, because Tg -detectability- merely relates to the functional sensitivity of the assay used.

( Summary and commentary prepared by Carole Spencer )

Present summary & commentary are related to Chapter N- 18 (see section -follow-up-) of the TDM