Serum thyroglobulin (Tg) measurement is used in the follow-up of patients with differentiated thyroid cancer following total thyroidectomy and radioactive iodine ablation. In athyrotic patients, thyroglobulin is an excellent tumor marker because it is produced exclusively by the follicular cells of the thyroid. However, thyroglobulin measurement is not recommended for the screening or initial diagnosis of thyroid cancer due to the significant overlap between the levels in benign thyroid diseases and those in thyroid cancer patients.
The clinical utility of thyroglobulin testing can be negatively affected by the presence of anti-thyroglobulin auto-antibodies (TgAb), which are detected in up to 30% of patients with differentiated thyroid cancer. TgAb interference is characterized by falsely low or undetectable thyroglobulin levels using immunometric Tg assays. Due to this problem various practice guidelines, including those from the NACB and ATA, stressed that TgAb should be measured on all samples tested for Tg. Failure to detect TgAb interference in the presence of an undetectable Tg could impact patient management as disease recurrence might go undiagnosed. It is also recommended that screening for TgAb is performed by immunoassay method and not by a recovery test.
A temporary withdrawal from the market of the TgAb assay used by a large number of laboratories force the adoption of alternative TgAb assays. Unfortunately, this change in methodology caused disruption of the clinical practice. In our institution, we selected an alternate assay that had the highest qualitative agreement with the previous assay (95% positive agreement and 96% negative agreement). Even with the excellent concordance between assays, there were cases where the TgAb value went from negative with the previous assay and to positive with the new assay. When this happens, the thoughts from clinicians are: (1) something is wrong with the assay or (2) the clinical status of the patient has change. However, it is important to recognize that these differences might be solely due to the different configuration of the TgAb immunoassays. Tg is a very large molecule (molecular weight = 660 kDa) and only a fraction of the molecule is used to detect the presence of TgAb in the various immunoassays. As a consequence, TgAb measured by any of the available immunoassays represent a subset of the entire TgAb spectrum and no assay will be capable of detecting all antibodies. It is important to emphasize that if TgAb has been detected in one assay the patient should be considered TgAb positive and the Tg level should be carefully correlated with the clinical presentation of the patient.
Another important issue concerning TgAb interpretation in the context of Tg assay interference is that all FDA approved TgAb assays are aimed for the aid in the diagnosis if thyroid autoimmune diseases and reference ranges are optimized for this indication and not for the detection of interference in Tg assays. Consequently, the majority of assays can only detect relatively high TgAb concentrations and could miss significant TgAb interference in the Tg assay. In addition, values within the normal range could still cause significant interference in the Tg assays. It has been proposed that when using TgAb in the context of Tg assay interference, any detectable TgAb should be considered positive. Ideally each laboratory should establish the value of TgAb that might interfere with their respective Tg assay.
One potential methodology to get away from the TgAb interference problem is liquid chromatography-tandem mass spectrometry (LC-MS/MS). The use of trypsin digestion prior to measuring Tg by LC-MS/MS can potentially overcome the interference from TgAb since both Tg and the antibodies will be digested. Proof of concept for a LC-MS/MS Tg assay has been published for TgAb negative patients and is currently been investigated by various groups, including our institution, in TgAb positive patients. However, at this point the reported limit of detection the Tg LC-MS/MS assay is significantly higher (2.6 ng/mL) than the currently available most sensitive immunoassays (0.1 ng/mL). In addition, the complexity of the LC-MS/MS method and potential IP issues will likely preclude it from been widely adopted. In theory, the cases that will benefit the most of this methodology will be TgAb positives patients with an undetectable or very low Tg level. Additional studies are needed to determine the clinical impact/benefit of measuring Tg by LC-MS/MS in this population.