Clinical Relevance of Thyroglobulin Doubling Time in the Management of Patients with Differentiated Thyroid Cancer

T he last several years have seen a renewed interest in risk stratification in thyroid cancer with an increased emphasis on how clinical data obtained during follow-up can be used to modify our standard initial risk estimates of recurrence and disease-specific mortality (1). Using this approach, recommendations regarding initial treatment, adjuvant therapy, and early follow-up are based on risk estimates obtained using one of many clinico-pathology staging systems that integrate standard clinical–pathological features (age, primary tumor size, extrathyroidal extension, completeness of resection, lymph node status, presence of distant metastases) to predict either the risk of recurrence (American Thyroid Association [ATA] staging system) or risk of death (American Joint Committee on Cancer [AJCC], MACIS [distant metastases, age, completeness of resection, local invasion, tumor size], etc.) (2). These initial risk estimates are then actively modified (either increased or decreased) during follow-up as additional clinical information becomes available. For example, in two independent studies (3,4), our groups have recently demonstrated that patients who achieve an excellent response to therapy (clinical remission) as defined by stimulated thyroglobulin (Tg) values less than 1 ng/mL with a negative neck ultrasound (US) should be reclassified as low-risk patients, with appropriate reduction in the intensity of follow-up and degree of thyroid hormone suppression. Furthermore, patients failing to achieve an excellent response to therapy (remission) had higher rates of recurrence/persistent disease than predicted by the initial static staging system, and therefore such patients should be followed more closely with continued thyroid hormone–suppressive therapy and often additional treatments.

In this issue of Thyroid, Miyauchi et al. (5) provide an important quantitative tool to assist in dynamic risk assessment by demonstrating that the doubling time of serum Tg measurements during follow-up of patients with persistent disease is a more powerful predictor of overall survival, disease recurrence, and development of distant metastases than the static initial risk estimates obtained using standard clinico-pathologic factors. From a cohort of 1431 patients treated at their center between 1998 and 2004, the authors identified 426 differentiated thyroid cancer patients who had at least four serum Tg measurements obtained with a thyrotropin (TSH) <0.1 mIU/L in the absence of anti-Tg antibodies following total thyroidectomy (167 also received radioactive iodine [RAI] remnant ablation). In most patients, serum Tg levels were obtained 1 and 3 months postoperatively, then twice a year for high-risk patients or once a year for low-risk patients over a median follow-up period of 87 months.

Of the 426 eligible patients, 137 had four or more persistently detectable Tg levels obtained during follow-up, allowing for determination of the Tg doubling time. An additional 88 patients had three detectable Tg levels over time and 201 had undetectable Tg values during follow-up. During follow up, six patients died of thyroid cancer, 25 developed newly identified distant metastases, and 58 had loco-regional recurrence.

In a multivariate analysis that included the standard clinico-pathologic features, and tumor size, node metastases, distant metastases (TNM) staging, only the Tg doubling time was an independent predictor of disease-specific mortality, loco-regional recurrence, and development of newly identified distant metastases. For example, the 10-year cause-specific survival was only 50% in the 17 patients with a Tg doubling time of less than 1 year; however, it was 95% when the doubling time was 1–3 years (21 patients), and 100% when the doubling time was greater than 3 years (30 patients), the Tg levels declined over time (69 patients), or the Tg remained undetectable on suppression during follow-up. Furthermore, these same Tg doubling time cut-offs (<1, 1–3, and >3 years) also effectively risk stratified patients with regard to both the risk of loco-regional recurrence and the risk of developing distant metastases. Therefore, just as the authors have previously demonstrated with regard to calcitonin doubling time in the management of medullary thyroid cancer (6), the current data clearly support the clinical utility of Tg doubling times in the dynamic risk assessment of patients with persistent thyroglobulinemia after total thyroidectomy.

In addition to using all Tg data points available over the entire course of follow-up, the authors also demonstrated that the Tg doubling time calculated using the first four Tg values obtained after thyroidectomy carried the same prognostic importance. This important observation allows for an early restratification of risk within the first 2 years of follow-up that can serve as a guide to frequency and intensity of follow-up by identifying both those patients with an excellent response to therapy who are likely cured of their disease (in remission) and those with persistent disease most likely to develop clinically significant outcomes (short Tg doubling times).

It is also important to point out that the Tg doubling times appear to be an effective tool even if RAI remnant ablation is not routinely performed. In this series, only 167 of 426 eligible patients had RAI ablation in addition to total thyroidectomy. While both persistent disease and normal thyroid remnants can be associated with low-level persistent thyroglobulinemia following total thyroidectomy, rising Tg values in the face of ongoing TSH suppression are far more likely to be an indicator of progressive disease than regrowth of normal thyroid tissue. Therefore, Tg doubling time may serve as a valuable predictor of recurrent disease even in patients treated with total thyroidectomy without RAI remnant ablation.

As is the case with any risk stratification system that relies on serum Tg, the clinical utility is far less in patients with anti-Tg antibodies and is also dependent on the consistency of the assay used and the degree of TSH suppression. However, since the primary finding of this study relates to the use of Tg doubling time in patients with persistent hyperthyroglobulinemia, the degree of TSH suppression is likely to be fairly consistent within an individual patient over several years of follow-up. As the authors point out, a simple web-based doubling time calculator is available on the ATA web site (http://www.thyroid.org/professionals/calculators/CDTC.php May 31, 2011).

The clinical implications of using Tg doubling times are particularly important when one considers that only 34% of our patients treated with total thyroidectomy and RAI achieved an undetectable suppressed Tg and stimulated Tg less than 1 ng/mL in the first 2 years of follow-up (4). In the absence of structurally identifiable disease, the best course of action in patients with persistent low-level Tg values is usually continued TSH suppression and observation over 1–2 years to evaluate the trend in Tg values over time. Oftentimes, detectable serum Tg values in the absence of structurally identifiable disease slowly decline over time, sometimes reaching undetectable levels (7). Conversely, rising Tg values over time have been shown to be associated with an increased risk of disease recurrence. We can now add the specific calculation of Tg doubling time to our ongoing assessment to better predict which patients with persistent thyroglobulinemia are at risk of developing loco-regional recurrence, distant metastases, and death.

While we do not currently recommend initiating systemic chemotherapy or a therapeutic drug trial on the basis of Tg doubling time in the absence of structurally identifiable disease, the Tg doubling time would guide the frequency, extent, and intensity of our evaluations since short doubling times are expected to result in identification of structural disease progression over months to years. Conversely, very long (or nonexistent) Tg doubling times are considered to be very reassuring even in the setting of structurally identifiable disease and likely predict relatively slow disease progression in all but the most poorly differentiated thyroid cancers. For example, a long Tg doubling time, particularly if associated with a negative 2-deoxy-2[¹⁸F]fluoro-d-glucose positron emission tomography scan, would likely warrant cautious observation rather than systemic therapy in patients with structurally identifiable RAI refractory pulmonary metastases that are not causing local compressive symptoms.

In conclusion, the Tg doubling time will serve as a useful tool that can identify those patients with persistent thyroglobulinemia after total thyroidectomy that are at highest risk for developing loco-regional recurrence, distant metastases, and disease-specific mortality from differentiated thyroid cancer. These data provide additional support to the concept of dynamic risk assessment in which data obtained during routine clinical follow-up is used to modify our initial risk estimates in order to tailor the intensity and frequency of follow-up, degree of TSH suppression, and need for additional therapies to a realistic, personalized, and actively updated risk estimate.