Hyperthyroidism is a clinical syndrome characterized by excessive thyroid hormones levels in the bloodstream due to various causes (1). Graves’ disease (GD) is the most common etiology of hyperthyroidism (2). Treatment options for GD include iodine-131 (¹³¹I), antithyroid drugs (ATD) and surgery (3). Although ATDs are the first-line treatment, relapse occurs in approximately 50% of patients following treatment discontinuation (4). ¹³¹I treatment is widely favored by clinicians due to its well-established safety profile, particularly in cases where ATDs are contraindicated or when patients fail to achieve euthyroidism with ATD therapy (5).

Studies have shown that multiple factors influence the efficacy of ¹³¹I treatment in GD, including gender, age, prior use of ATDs, free thyroxine (FT₄) levels, thyroid mass, and the ¹³¹I dose (6). Among these, thyroid mass is a crucial parameter in determining the appropriate ¹³¹I dose for GD patients (7). Multiple studies have emphasized its influence on the treatment response of GD patients (8–10). Currently, thyroid mass estimation methods include palpation, ultrasound (US), computer tomography (CT), and radionuclide imaging, with US being the most commonly used in clinical practice. The standard US measurement follows the ellipsoid volume formula, V = π/6 × L × W × T. However, a cadaveric study (11) suggested that adjusting the correction factor from 0.524 to 0.479 could improve measurement accuracy. A multi-method thyroid measurement study (12) indicates that compared with CT measurements, US estimates are on average 20.06 ± 8.31 g smaller than CT methods. Research (13) has also highlighted significant discrepancies between thyroid volumes measured by US and those determined intraoperatively, with measurement errors increasing as thyroid volume enlarges. In contrast, CT is a well-established technique for thyroid volume assessment, offering a higher degree of accuracy (14). One phantom study demonstrated exceptional agreement between CT-measured and actual thyroid volumes, with a mean error of just 0.27 ± 1.53% (15). The aim of this study is to evaluate the accuracy of US in estimating thyroid mass, using CT as the reference standard and to identify the critical threshold for significant differences between the two methods. To visualize the differences between CT and US thyroid mass measurements, Sankey diagrams were used to display the distribution patterns across both methods. The findings will enable nuclear medicine practitioners to more accurately prescribe ¹³¹I dosages for GD treatment, better understand the impact of thyroid mass on ¹³¹I treatment efficacy, and develop a prognostic model for predicting treatment outcomes.

¹³¹I treatment is widely recognized as a safe and effective treatment for GD, offering distinct advantages such as ease of administration, high safety profile, short treatment duration and low recurrence rate. In recent years, an increasing number of physicians and patients have opted for ¹³¹I treatment for GD (18). Accurate assessment of thyroid mass is fundamental to determining the appropriate ¹³¹I dosage, which is critical optimizing treatment efficacy and ensuring patient safety.

US imaging is widely recognized for its accuracy in assessing thyroid gland mass within the normal range. However, studies (19) have shown that when thyroid volumes exceed 40 ml, the measurement error of US increases significantly. Certain study (11) evaluated thyroid lobes ranging from 8 to 70 ml and reported an average inaccuracy rate of 16%. In contrast, multiple studies (14, 20) have established CT-based thyroid volumetry as a clinically reliable diagnostic method, indicating excellent agreement with actual thyroid volumes. This technique shows high accuracy in complex thyroid cases, such as multinodular goiters or substernal extensions, as it allows for three-dimensional visualization to enable accurate volume calculations. Consistent with these findings, our study demonstrated that the relative error between CT and the US measurements was 0.19 ± 11.65%. To provide a more intuitive visualization of the relationships and proportional differences between CT and US measurement methods, we employed a Sankey diagram. In this diagram, the width of streamlines, derived from US and CT data nodes, represents the distribution of thyroid mass measurements between the two modalities. Wider streamlines indicate a greater proportion of corresponding mass values within the sample, offering a clearer depiction of the measurement discrepancies between the two techniques (21). Our study demonstrated that in patients with GD, thyroid mass is predominantly distributed in the higher CT cohorts 7 (60–70 g), 8 (70–80 g), and 10 (> 90 g). Furthermore, there is a noticeable tendency for thyroid masses classified under US cohort 8-10 (m > 70 g) to align with CT cohort 10 (m > 90 g), suggesting a systematic underestimation of larger thyroid masses by US. Thyroid mass measurements obtained via CT were consistently higher than those derived from US, highlighting a tendency for US to underestimate larger thyroid mass. This difference is underscored by poor agreement (ICC = 0.179) and significant bias (16.65 ± 15.60 g), indicating that the two methods are not interchangeable in clinical practice. While previous studies have noted this underestimation, a definitive critical threshold had not been established. Using CT as the reference standard, our study identified a significant discrepancy between US and CT measurements when thyroid mass exceeded 20 g. It is demonstrated that for thyroid masses ≤ 20g, US remains a reliable method for assessment, whereas CT calibration is advisable for larger thyroid glands to enhance measurement accuracy.

Multivariate analysis revealed that multiple factors influence treatment outcome, among which disease duration emerged as significant determinant of ¹³¹I treatment efficacy. Our study demonstrated that patients with longer disease duration tended exhibited poorer prognoses, aligning with findings from previous studies (22, 23). A possible explanation is that prolonged course of GD, often accompanied by extended ATD therapy and recurrent exacerbations, may contribute to autoimmune dysfunction. Persistent TRAb binding to TSH receptors on thyroid cells continuously activates the cAMP signaling pathway, promoting hyperplasia in follicular epithelial cells and lymphoid tissue (24). Such result in the depletion or absence of colloid within thyroid follicles, increasing thyroid stiffness and potentially hindering the therapeutic efficacy of β-radiation in ¹³¹I treatment. Additionally, our comparison of thyroid function before ¹³¹I treatment revealed that patients with unsuccessful treatment outcomes exhibited elevated FT₄ levels. Previous studies have reported that GD patients with elevated FT₄ levels exhibit more pronounced disease severity. This heightened metabolic activity may accelerate the catabolism of internal radiation, thereby diminishing therapeutic responsiveness (25). However, some studies (26, 27) have demonstrated that FT₄ levels do not significantly influence the success rate of ¹³¹I treatment. A previous study (28) reported that a higher dose of ¹³¹I was associated with an increased likelihood of therapeutic failure, which aligns with our findings. Conversely, most studies (29, 30) the ¹³¹I dose was higher in the responsive group compared to the non-responsive group. This discrepancy may be attributed to the fact that the maximum initial treatment dose in our study was only 111 MBq (30 mCi), which was considerably lower than that in other studies. Additionally, as thyroid mass increases, US measurement errors become more pronounced. Larger masses often exceed the ultrasound probe’s optimal field of view, causing boundary visualization issues and signal distortions from internal structures like calcifications and substernal extensions (13, 31). These errors lead to underestimating the required ¹³¹I dose, potentially reducing treatment efficacy. Consequently, the administered dose may fall below the therapeutic threshold necessary for achieving a cure, resulting in inconsistent treatment outcomes. Another critical factor influencing the efficacy of ¹³¹I treatment is thyroid gland uptake. A lower thyroid 24-hour RAIU implies a reduced capacity for iodine retention, leading to decreased ¹³¹I absorption and a shorter effective duration in vivo, ultimately compromising therapeutic success. Moreover, individual radiosensitivity may play a key role in determining of the outcome of ¹³¹I treatment (32). Additionally, although the impact of TRAb on iodine therapy was not prominent in this study, we determined the optimal cut-off value for TRAb to be 38.26 based on the Youden index, which only achieved an AUC value of 0.548. Nevertheless, detection of TRAb changes remains of significant value in the diagnosis of Graves’ disease, as well as in evaluating disease course and recurrence (25).

In the context of the ¹³¹I treatment dosage calculation, thyroid mass serves as a critical determinant in establishing the appropriate therapeutic dosage. Our findings align with the early observations (33), which indicated that larger thyroid mass associated with an increased risk of ¹³¹I treatment failure. This correlation has been further validated by multiple studies. For instance, a retrospective study (34) reported that the 1-year cure rates of the groups with gland weight <30g, 30-60g and >60g were 60.0%, 46.7% and 36.1%, respectively, underscoring the inverse correlation between treatment success and thyroid mass. Similarly, another analysis (35) showed that thyroid mass was the sole determinant of treatment success, with a median mass of 44.6 g in patients who achieved remission. In our study, the identified cut-off value was 35.6 g. The underlying mechanism may involve two key factors. First, an increased thyroid mass can result in inconsistent gland thickness and autoimmune-mediated fibrosis, disrupting the uniform distribution of β-radiation. Second, the incomplete visualization of larger thyroid glands during US may lead to an underestimation of thyroid diameters and overall mass, ultimately causing a miscalculation of the required ¹³¹I dose (36). When the thyroid mass is less than 35.6g, the cure rate of ¹³¹I treatment is relatively high. Given this, a differentiated approach is warranted for patients with GD undergoing initial ¹³¹I treatment. For those with a normal or mildly enlarged thyroid, the US can serve as the primary assessment modality, effectively minimizing unnecessary radiation exposure from CT. However, in case of larger thyroid glands, CT calibration is recommended to ensure precise ¹³¹I dose calculation and achieve the intended therapeutic outcome.

This study has several limitations. First, its retrospective design makes it susceptible to selection and statistical biases. Second, the follow - up period was limited to six months, focusing solely on factors affecting the efficacy of a single ¹³¹I treatment. Additionally, the relatively small AUC value for thyroid mass and the moderate accuracy highlights the need for further refinement. Future research should aim to increase sample size, extend the follow – up duration, or explore advanced machine - learning methods to improve predictive accuracy.

For patients with normal or mildly enlarged thyroids, US is sufficient for routine assessment. However, for larger thyroids, CT provides a more precise evaluation, ensuring accurate calculation of the appropriate ¹³¹I dose. Identifying patients at high - risk clinical factors of non-cure before ¹³¹I treatment is crucial, particularly those with larger thyroid mass. Adjusting the ¹³¹I dose accordingly, with CT calibration when necessary, may enhance the cure rate.