Thyroidectomy without Radioiodine in Patients with Low-Risk Thyroid Cancer

Trial Oversight

The ESTIMABL2 trial was performed within the Tumeurs de la Thyroïde Refractaires Network with support from the French Ministry of Health through a grant from the National Cancer Institute. All the patients provided written informed consent. The trial was conducted in accordance with the protocol (available with the full text of this article at, which was approved by a central ethics committee and by national authorities. Data were collected by all the authors and were analyzed by the first and last authors. The manuscript was reviewed by all authors before submission for publication. All the authors assume responsibility for the accuracy and completeness of the data and for the fidelity of the trial to the protocol.


Patients were adults (≥18 years of age) with a differentiated thyroid carcinoma (papillary, follicular, or oncocytic [Hürthle-cell cancer]), with a multifocal pT1a tumor (a diameter of each lesion of ≤1 cm and a sum of the longest diameters of the lesions of ≤2 cm) or a pT1b tumor (> 1 cm and ≤2 cm); with both tumor stages, patients had a nodal status of N0 (no evidence of regional node involvement) or Nx (regional lymph nodes cannot be assessed in the absence of neck dissection), in the absence of extrathyroidal extension (Table S1 in the Supplementary Appendix , available at Patients who had aggressive histologic subtypes (tall-cell, clear-cell, columnar-cell, and diffuse sclerosing variants of papillary thyroid cancer, poorly differentiated) were excluded from the trial.16

Two to five months before randomization, eligible patients had undergone total thyroidectomy with or without dissection of cervical lymph nodes (neck dissection), according to ongoing protocols, with complete tumor resection. In addition, all the patients had undergone postoperative neck ultrasonography without the detection of suspicious abnormalities.


Patients were randomly assigned to undergo postoperative administration of radioiodine (radioiodine group) or not to undergo such administration (no-radioiodine group), with stratification according to the trial site and lymph-node status (N0 or Nx). In the radioiodine group, while patients were receiving thyroid hormone treatment, 1.1 GBq (30 mCi) of radioiodine was administered 24 hours after the second intramuscular injection of recombinant human thyrotropin (Thyrogen, Sanofi), which was given at a dose of 0.9 mg on 2 consecutive days. Whole-body scanning and single-photon-emission computed tomography (SPECT) of the neck were performed 2 to 5 days after radioiodine administration.

The follow-up protocol was consistent with the standard of care in France and consisted of the measurement of thyroglobulin and thyroglobulin antibodies in all patients at 10 months and annually thereafter. Thyroglobulin was measured while the patient was receiving thyroid hormone treatment, except for the measurement at 10 months after randomization in the radioiodine group, in which the measurement was performed after stimulation with recombinant human thyrotropin. Ultrasonography of the neck was performed in all patients 10 months and 3 years after randomization. No diagnostic radioiodine scan was performed after the whole-body scan that was performed after therapy.

Primary and Secondary End Points

The primary objective was to assess noninferiority in the no-radioiodine group as compared with the radioiodine group with respect to the percentage of patients without a functional, structural, or biological event during 3 years after randomization. An event was a composite end point that consisted of several criteria.

In the radioiodine group only, events included the presence of foci of radioiodine uptake outside the thyroid bed on postablation whole-body scanning or on SPECT (functional event) that resulted in additional treatment (radioiodine administration or surgery). In both groups, events included abnormal findings on ultrasonography of the neck (structural events), which were defined as a suspicious lymph node or thyroid mass associated with abnormal cytologic findings or with a serum thyroglobulin level in the aspirate fluid of more than 10 ng per milliliter or an elevated level of thyroglobulin or thyroglobulin antibodies (biological events). In the radioiodine group, in the absence of thyroglobulin antibodies, an elevated thyroglobulin level was defined as a value of more than 5 ng per milliliter (after the receipt of recombinant human thyrotropin or during thyroid hormone treatment) or a value of more than 1 ng per milliliter during thyroid hormone treatment on two consecutive measurements obtained 6 months apart. In the no-radioiodine group, in the absence of thyroglobulin antibodies, an elevated thyroglobulin level was defined as a value of more than 5 ng per milliliter during thyroid hormone treatment or a value of more than 2 ng per milliliter during thyroid hormone treatment on two consecutive measurements obtained 6 months apart. In the two groups, the detection of a thyroglobulin antibody titer above the upper limit of the normal range or an increase in the thyroglobulin antibody titer by more than 50% between two measurements performed 6 months apart was also considered to be part of the composite criteria . Serum samples for the measurement of thyroglobulin and thyroglobulin antibodies that were obtained at randomization and at the 10-month and 3-year follow-up visits were assessed both locally and in a central laboratory (Table S2).17

Secondary end points were quality of life, anxiety, fear of recurrence, and dysfunction of lacrimal and salivary glands. Questionnaires were administered to all the patients at the time of randomization, after radioiodine administration during hospitalization in the radioiodine group or 2 months after randomization in the no-radioiodine group, and at 10 months and 3 years after randomization in the two groups.

Prognostic Factors

Prognostic factors for an event were evaluated with the use of univariate logistic regression. A post hoc analysis comparing the percentages of patients with no evidence of disease (as defined by the 2015 guidelines of the American Thyroid Association as an “excellent response” to treatment) was performed in the two groups at 10 months and at 3 years. An excellent response at 10 months was defined as normal findings on neck ultrasonography with a thyroglobulin level of less than 1 ng per milliliter after the administration of recombinant human thyrotropin (in the radioiodine group) or during thyroid hormone treatment (in the no-radioiodine group ) in the absence of an elevated level of thyroglobulin antibodies. An excellent response at 3 years was defined as normal findings on neck ultrasonography with a thyroglobulin level of less than 0.2 ng per milliliter (in the radioiodine group) and a level of less than 1 ng per milliliter (in the no-radioiodine group) during thyroid hormone treatment in the absence of an elevated level of thyroglobulin antibodies. Results were assessed after review of findings from both local and central laboratories.4

Molecular Analysis

The objective was to describe the type and number of molecular alterations, according to the occurrence of an event (Table S2). For this purpose, we designed a nested case – control study in which cases were patients with an event and controls were patients without an event, regardless of which randomized treatment had been received. Each case patient was paired with two controls on the basis of five criteria: pT classification (pT1a or pT1b), histologic results (papillary or follicular), sex, age group (≤55 years or> 55 years), neck dissection (yes or no), and treatment group (radioiodine or no radioiodine).

Statistical Analysis

We designed our trial as a noninferiority study to answer the question of whether the percentage of patients without an event in the no-radioiodine group would be noninferior to that in the radioiodine group, with a between-group difference of less than 5 percentage points at 3 years. We calculated that the enrollment of 750 patients would provide the trial with 90% power to determine the noninferiority margin, assuming that 95% of the patients in the radioiodine group would not have an event at 3 years and including the potential loss to follow-up ; a P value of less than 0.05 was considered to indicate statistical significance.

Descriptive quantitative data were expressed as means and standard deviations; qualitative data were expressed as percentages and 95% confidence intervals. We calculated the difference in the observed percentages of patients without an event and its 95% one-sided confidence interval, which was equivalent to a two-sided 90% confidence interval, because the trial was designed to use one-sided hypothesis testing at an alpha level of 0.05. We used logistic regression to perform an analysis of the primary end point after adjustment for stratification factors as a test of robustness.

The primary analysis included all the patients who could be evaluated in the per-protocol population (ie, all the patients whose treatment and 3-year follow-up had adhered to the study protocol); patients who had been followed for fewer than 3 years were excluded from the primary population. The percentage of patients without an event during the trial period was calculated without consideration for the timing of the event during that period. A sensitivity analysis was performed in the intention-to-treat population that included the results for all the patients until their last participation in the trial in order to consider those who could not be evaluated at 3 years. We report the between-group difference in event-free survival at 3 years using a time-to-event analysis and its two-sided 90% confidence interval.

For secondary end points, results are reported as point estimates and 95% confidence intervals, since the statistical analysis plan did not include a provision for correcting for multiple comparisons. Thus, the widths of the confidence intervals have not been adjusted for multiplicity, so they should not be used to infer definitive treatment effects. All the analyzes were performed with the use of SAS software, version 9.4 (SAS Institute).

Leave a Reply

Your email address will not be published.