Warning: fopen(/home/virtual/enm-kes/journal/upload/ip_log/ip_log_2025-03.txt): failed to open stream: Permission denied in /home/virtual/lib/view_data.php on line 100 Warning: fwrite() expects parameter 1 to be resource, boolean given in /home/virtual/lib/view_data.php on line 101 Microvascular Ultrasonography Vascularity Index as a Rapid and Simplified Assessment Tool for Differentiating Graves’ Disease from Destructive Thyroiditis and Managing Thyrotoxicosis
Skip Navigation
Skip to contents

Endocrinol Metab : Endocrinology and Metabolism

clarivate
OPEN ACCESS
SEARCH
Search
Advanced Search
mobile menu button

Articles

Page Path
HOME > Endocrinol Metab > Ahead-of print > Article
Original Article
Microvascular Ultrasonography Vascularity Index as a Rapid and Simplified Assessment Tool for Differentiating Graves’ Disease from Destructive Thyroiditis and Managing Thyrotoxicosis
Han-Sang Baek1orcid, Chaiho Jeong1, Jeonghoon Ha2, Dong-Jun Lim2orcid

DOI: https://doi.org/10.3803/EnM.2024.2206
Published online: February 25, 2025

1Division of Endocrinology and Metabolism, Department of Internal Medicine, Uijeongbu St. Mary’s Hospital, College of Medicine, The Catholic University of Korea, Uijeongbu, Korea

2Division of Endocrinology and Metabolism, Department of Internal Medicine, Seoul St. Mary’s Hospital, College of Medicine, The Catholic University of Korea, Seoul, Korea

Corresponding author: Dong-Jun Lim Division of Endocrinology and Metabolism, Department of Internal Medicine, Seoul St. Mary’s Hospital, College of Medicine, The Catholic University of Korea, 222 Banpo-daero, Seocho-gu, Seoul 06591, Korea Tel: +82-2-2258-6009, Fax: +82-2-599-3589, E-mail: ldj6026@catholic.ac.kr
• Received: October 12, 2024   • Revised: November 28, 2024   • Accepted: December 2, 2024

Copyright © 2025 Korean Endocrine Society

This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

  • 171 Views
  • 15 Download
Full Article

Download PDF Download PDF

  • Background
    Thyrotoxicosis presents significant diagnostic challenges in distinguishing Graves’ disease (GD) from destruction-induced thyrotoxicosis (DT) using ultrasound imaging. We evaluated a new technology, microvascular ultrasonography (MVUS) to effectively differentiate between GD and DT, and observe the MVUS changes during follow-up.
  • Methods
    A total of 264 consecutive patients were prospectively enrolled into two cohorts from August 2022 to March 2024 at one tertiary referral hospital: cohort 1 comprised patients initially presenting with thyrotoxicosis (n=185; 98 with GD and 87 with DT). Cohort 2 included patients either with GD considering antithyroid drug discontinuation or with DT in the follow-up phase after treatment (n=77). Ultrasound imaging was conducted using the MVUS technique, and the vascularity index (MVUS-VI) was automatically calculated as the percentage ratio of color pixels to total grayscale pixels within a specified region of interest.
  • Results
    Diagnostic accuracy highlighted MVUS-VI as the most accurate diagnostic tool, achieving a sensitivity of 79.6%, specificity of 84.3%, with an area under the curve of 0.856 (95% confidence interval, 0.800 to 0.911). Presence of thyroid peroxidase antibody or thyroglobulin antibody affected MVUS-VI’s performance, requiring a higher cut-off value for specificity in this subgroup. Follow-up in cohort 2 (n=77) demonstrated significant normalization in thyroid function and reductions in MVUS-VI from an initial 32.6%±23.4% to 20.8%±13.5% at follow-up (P<0.001).
  • Conclusion
    MVUS-VI provides a rapid, non-invasive diagnostic alternative to traditional methods in differentiating GD from DT, thus aiding in the management of patients with thyrotoxicosis.
  • Keywords: Thyrotoxicosis; Graves disease; Thyroiditis; Receptors, thyrotropin; Immunoglobulins, thyroid-stimulating; Ultrasonography, Doppler, color; Microvascular imaging
Thyrotoxicosis refers to the clinical symptoms and signs attributed to increased thyroid hormone levels in the body [1]. The etiology of thyrotoxicosis could be divided into two causes: thyroid hormone overproduction and destruction-induced thyrotoxicosis (DT) due to inflammation [2]. A typical example of thyroid hormone overproduction associated thyrotoxicosis is Graves’ disease (GD), which is often treated with antithyroid drugs (ATDs) or definite treatments, such as surgery or radioiodine therapy [3]. In contrast, DT is often caused by conditions such as autoimmune thyroiditis (AIT), subacute thyroiditis (SAT), postpartum thyroiditis, and amiodarone-induced thyrotoxicosis. It is typically managed conservatively, without ATDs DT is typically managed conservatively, without ATDs [1]. Thus, prompt and accurate differentiation between GD and DT is crucial.
Current diagnostic methodologies, such as thyroid scintigraphy and thyroid-stimulating hormone (TSH) receptor antibody (TSH-R-Ab) measurements, provide high sensitivity and specificity but are often delayed or unavailable in less equipped settings [4]. In contrast, thyroid ultrasound (US), including newer techniques such as microvascular ultrasonography (MVUS), offers a rapid, non-invasive diagnostic alternative [5,6]. However, conventional Doppler US struggles with quantification challenges, often resulting in subjective interpretations that may not effectively differentiate GD from DT [7-10]. MVUS is an advanced imaging technique that detects slow-velocity blood flow in small vessels without intravenous contrast, offering superior intralesional flow visualization compared to conventional color and power Doppler (PD) imaging, and is increasingly being incorporated into mid-level scanners [11]. The ability to automatically generate a vascularity index (VI) is typically provided on machines equipped with MVUS, enhancing the reproducibility and accuracy of the measurements [12].
A previous study demonstrated that MVUS-VI could differentiate GD from DT with 80.8% and 85.2% accuracy, and positively correlated with the TSH-R-Ab [5]. However, the use of MVUS can be complicated by conditions, such as AIT, where inflammatory processes may induce an increased microvascular flow, potentially mimicking the vascular patterns observed in GD [5,6]. This phenomenon underscores the importance of distinguishing the inflammatory vascular enhancements typical of AIT from those indicated in GD [8]. Furthermore, studies on the effects of thyroid peroxidase antibody (TPO-Ab) and thyroglobulin antibody (Tg-Ab) on MVUS parameters are lacking, raising questions about the impact of these antibodies on the interpretation of MVUS results. Moreover, studies on the post-treatment microvascular changes across different thyroid diseases are notably lacking. Accordingly, we aimed to determine whether the flow changes induced by AIT can be distinguished from those induced by GD, assess the impact of TPO-Ab and Tg-Ab on MVUS parameters, and evaluate how MVUS changes during treatment and correlates with antibody levels. This study was designed to confirm the usefulness of MVUS according to the clinical course of thyrotoxicosis and to evaluate the effect of TPO-Ab and Tg-Ab on MVUS results, incorporating follow-up data and a broader range of thyroid conditions based on our preliminary work.
Study design and recruitment
In a previous primary study, we prospectively enrolled consecutive patients with thyrotoxicosis who visited one tertiary referral hospital between October 2020 and November 2021 [5]. For expansion of the initial study, we prospectively recruited more patients defining two cohorts with extended periods, from August 2022 to March 2024: cohort 1 included an increased number of study participants similar to those who initially present with thyrotoxicosis. Cohort 2 comprised follow-up patients at the consideration point for stopping ATDs in GD or with a history of DT approximately 1-year post-diagnosis (https://clinicaltrials.gov/ct2/show/NCT04879173).
Participant selection
Adults aged ≥18 years with thyrotoxicosis were included in the study. In the previous primary study, 114 patients underwent initial screening. Of these, 17 patients, including eight normal controls, were excluded based on the exclusion criteria, resulting in 97 patients being included in cohort 1 (Fig. 1). For the expansion study, an additional 88 patients added to cohort 1 for initial work-up. Cohort 2 includes 77 patients (71 diagnosed with GD): 14 patients who underwent follow-up MVUS in the primary study and 63 patients who underwent MVUS at the follow-up stage in the expansion study.
In both the primary and expansion studies, patients were excluded if they had toxic adenoma, thyroid nodules, or thyroid cancer, or if they were lost to follow-up, rendering their clinical course unclear. The primary study excluded patients who had taken ATDs within 2 weeks prior to the study, whereas the expansion study included patients who had taken ATD within up to 4 weeks prior to the study. For detailed information, refer to Fig. 1 and its legend.
Ultrasound imaging protocol
US assessments, including grayscale US, color Doppler (CD), PD, and MV-FLOW (Samsung Medison Co. Ltd., Seoul, Korea), were conducted using an RS85 equipped with MV-Flow. The two observers in this study were both endocrinologists. One had over 10 years of experience in thyroid US, while the other had approximately 4 years of experience. In the expansion study, only one observer scanned the newly added patients according to the same protocol of data acquisition.
Volumetric measurements of each thyroid lobe were calculated using grayscale US with the formula π/6×length×width× depth. CD and PD images were captured in a transverse orientation at the central portion of both thyroid lobes. For MVUS, imaging captured the most prominent vascular signals during the systolic phase to accommodate cardiac cycle variations. VI was defined as the ratio of colored to total grayscale pixels within a manually outlined region of interest (ROI) and used for reducing the subjectivity in assessing the vascularity. Thyroid nodules and other architectural abnormalities were excluded based on the exclusion criteria, ensuring that the ROI was representative of the entire gland. The VI measurements were performed on the US scanner in real-time while the patient was present. For a detailed method of a semi-quantitative scale, with four patterns from minimal to ‘thyroid inferno’ states, MVUS and MVUS-VI, kindly refer to our previous primary study [5].
In the US examination protocol, nodules that occupied an entire lobe were excluded. For smaller nodules, ROI was carefully selected to exclude these nodules, ensuring they did not affect VI calculations.
Data collection for clinical information
Confirmation of thyrotoxicosis in enrolled participants was established through biochemical assays demonstrating elevated levels of free thyroxine (fT4) and suppressed TSH levels. Comprehensive laboratory evaluations included measurements of TSH, fT4, TPO-Ab, Tg-Ab, and TSH-R-Ab assessed using the competitive thyrotropin-binding inhibitory immunoglobulin (TBII) assay and, when clinically indicated, thyroid-stimulating immunoglobulin (TSI) bioassay [13]. The reference intervals were; TSH 0.55–4.78 μIU/mL, fT4 0.89–1.76 ng/mL, triiodothyronine 0.6–1.81 ng/mL, TPO-Ab and Tg-Ab <60 U/mL, TBII <1.75 IU/L, and TSI bioassay <140%.
GD diagnosis was primarily through positive TSH-R-Ab tests, with considerations of clinical history and thyroid scans. AIT was identified by elevated TPO-Ab or Tg-Ab levels and US features indicative of Hashimoto’s thyroiditis. SAT diagnosis included painful thyroid symptoms, elevated erythrocyte sedimentation rate and C-reactive protein, and US detection of hypoechoic areas. Drug-induced or postpartum thyroiditis was diagnosed through patient history. All participants had a minimum follow-up of 3 months.
For the patients with GD in cohort 2, US imaging was performed when considering discontinuing ATD as the thyroid function test (TFT) results demonstrated euthyroid status after sufficient use of ATD, and the physician was planning to stop the ATD. For the patients with DT in cohort 2, US was performed for follow-up approximately 12 months after the initial diagnosis of the disease, when the TFT results were normal.
Sample size calculation and statistical analysis
We estimated the required sample size for evaluating the sensitivity and specificity of MVUS-VI in distinguishing GD within a thyrotoxicosis cohort. Based on previously observed sensitivity (87%) and specificity (80.9%) [5], and assuming a GD prevalence of 40% in our cohort, we aimed for a 95% confidence level with a precision of ±10 percentage points (0.1) and accounted for a 10% dropout rate, resulting in a final sample size of 122 participants [14,15]. We assumed the GD prevalence of 40% because in iodine-sufficient areas such as Korea, the prevalence of AIT, including Hashimoto’s thyroiditis, is higher compared to iodine-deficient regions [16,17].
Data were analyzed using R version 4.1.1 (R Project for Statistical Computing, Vienna, Austria). Graphs were produced using Prism version 8.02 (GraphPad Software Inc., La Jolla, CA, USA). Baseline characteristics are depicted as mean±standard deviation, median (interquartile range), or frequency (%). Continuous variables were compared through t test, while categorical variables were assessed using the chi-square test. The diagnostic performance of MVUS in distinguishing between GD and DT was evaluated through receiver operating characteristic (ROC) curve analysis, calculating sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and the area under the curve (AUC). The optimal cut-off value was defined as the point where the sum of sensitivity and 1-specificity was maximized [18]. With an optimal cut-off value, the accuracy for the diagnosis of GD using CD, PD, MVUS, and MVUS-VI was calculated [19]. The AUC is described with 95% confidence interval (CI). Delong’s test was used for comparing two correlated ROC curves [20].
Ethical considerations
The study was approved by the Institutional Review Board of Seoul St. Mary’s Hospital, and all the participants provided written informed consent (KC19DCSI0744). The study adhered to the ethical guidelines of the Declaration of Helsinki for medical research involving human participants.
Baseline clinical characteristics

Initial presentation (cohort 1): differences between GD and DT

At initial presentation, the study included 98 and 87 participants with GD and DT, respectively. The age and female sex distribution between the GD and DT groups were similar, with mean ages of 47.8±14.7 and 47.1±13.4 years (P=0.755) and 81.6% vs. 82.8% female participants (P=0.988), respectively.
The TFT and thyroid autoantibody levels showed significant difference between the two groups. The GD group had a higher mean fT4 level (3.6±1.9 ng/dL vs. 2.9±1.7 ng/dL, P=0.007) and significantly elevated TBII (11.6±10.1 μ/L vs. 1.6±4.7 μ/L, P<0.001) and TSI bioassay (352.1%±182.6% vs. 81.5%±81.0%, P<0.001) levels compared to those of the DT group.
The US parameters in the GD group reflected higher vascularity, with CD, PD, MVUS, and MVUS-VI at 1.7%±0.9%, 1.9%±0.9%, 2.4%±0.8%, and 45.8%±22.2%, respectively, all significantly higher than those in the DT group (P<0.001 for all comparisons) (Table 1).

Follow-up (cohort 2): changes from cohort 1

Upon follow-up, which included evaluations of 77 patients, normalization in the TFT results and reduced vascularity were observed. The mean fT4 levels decreased significantly to 1.7±1.3 ng/dL (P<0.001 compared to initial), and the TSH levels normalized to 2.325±1.623 μIU/mL (P<0.001 compared to initial). The TBII and TSI bioassay also demonstrated a significant decrease, indicating an overall reduction in thyroid autoimmunity. US parameters further substantiated these findings, with reductions in the CD, PD, MVUS, and MVUS-VI, indicating diminished thyroid gland vascularity and activity. The MVUS-VI decreased from an average of 32.8%±23.5% at the initial diagnosis (cohort 1) to 21.9%±14.0% at follow-up (cohort 2) with statistical significance (P<0.001) (Table 2).
Diagnostic accuracy of MVUS through ROC curve analysis

Diagnostic performance of ultrasound parameters in the entire study population (cohort 1)

Table 3 shows the performance matrix of CD, PD, MVUS, and MVUS-VI for differentiating GD and DT. MVUS by semi-quantitative measurement further improved the sensitivity to 84.0% and had a specificity of 72.1%, with an AUC of 0.823 (95% CI, 0.762 to 0.884), and accuracy of 74.4% with optimal cut-off of 2, compared to CD or PD. However, the ROC curve comparison demonstrated no significant difference with PD (P=0.997) and CD (P=0.469). MVUS-VI achieved the highest diagnostic accuracy, with a sensitivity of 79.6%, specificity of 83.9%, and an AUC of 0.855 (95% CI, 0.799 to 0.910), reaching an accuracy of 81.1%. The optimal cut-off value was 27.35% for differentiating the GD from DT. The Delong’s tests indicated a significant improvement using CD (P=0.011) and a non-significant improvement using PD (P=0.161) (Table 3).

Diagnostic performance of ultrasound parameters in the study subjects for TPO-Ab or Tg-Ab

Further analysis targeted patients positive for TPO-Ab or Tg-Ab. The results indicate that while the sensitivity of both MVUS and MVUS-VI increases in the presence of TPO-Ab or Tg-Ab, their specificity decreases, leading to decreased overall accuracy. MVUS-VI demonstrated a sensitivity of 87.5%, specificity of 72.3%, PPV of 79.0%, NPV of 82.9%, and an AUC 0.831 (95% CI, 0.750 to 0.913), achieving an accuracy of 80.6% (Table 4).
Fig. 2 illustrates the sensitivity and specificity according to cut-off of MVUS-VI through the ROC curve analysis. Fig. 2A is for the entire study population, and Fig. 2B is for population with positive TPO-Ab or Tg-Ab. Although the optimal cut-off value of MVUS-VI for differentiating GD from DT is similar for the entire study population or patients with positive TPO-Ab or Tg-Ab (27.35%), the specificity achieved with the optimal cut-off value was lower in patients with positive TPO-Ab or Tg-Ab. Therefore, a higher cut-off value was needed to obtain higher specificity in this population. For example, for obtaining 91% specificity in the entire cohort 1 population, MVUS-VI should be 35.15% but in the TPO-Ab or Tg-Ab positive population, MVUS-VI was 53.60% for the same specificity. However, in this case, the sensitivity was lowered, and was not adopted as an optimal cut-off.
VI in three groups: GD, AIT, and SAT
For further analysis, we divided the DT group into AIT and SAT groups (Fig. 3). The differences of MVUS-VI values among three groups demonstrated statistical significance (45.75%±22.2% in GD, 11.30%±5.6% in SAT, and 26.36%±17.7% in AIT, P<0.001). The multiple comparisons also showed statistical significance between each group (P<0.001 for GD and SAT, P<0.001 for GD and AIT, and P=0.002 for SAT and AIT). Despite this difference, the MVUS-VI values overlapped between GD and AIT. The 13 AIT patients were classified as GD using MVUS-VI (false positive), and 20 of the GD patients were classified as not GD using MVUS-VI (false negative).
Analysis of MVUS-VI in patients with GD: influence of TPO-Ab or Tg-Ab on treatment stage
To evaluate the influence of thyroiditis, which may increase MVUS-VI resulting in false positive results for GD, we further analyzed TPO-Ab or Tg-Ab and MVUS-VI in patients with GD (Fig. 4). The MVUS-VI of this population demonstrated variations according to TPO-Ab or Tg-Ab positivity. At initial diagnosis, the MVUS-VI was significantly higher in patients with positive TPO-Ab or Tg-Ab, compared to that in patients with negative TPO-Ab and Tg-Ab (51.7%±22.4% vs. 32.1%±17.9%, P= 0.002). However, at the time of stable status following ATD therapy, considering ATD discontinuation, the differences of MVUS-VI values between two groups showed no statistical significance (25.8%±16.1% vs. 22.4%±10.0%, P=0.662).
At the time of GD diagnosis, the MVUS-VI and TBII showed positive correlation (r=0.422, P<0.001) (Fig. 5). The TSI bioassay also demonstrated positive correlation with the MVUS-VI (r=0.582, P<0.001). At the time before ATD discontinuation, the TBII and TSI bioassay also demonstrated positive correlation with the MVUS-VI (r=0.427, P<0.001 for TBII and r=0.473, P=0.001).
Our study demonstrated that MVUS and MVUS-VI could differentiate GD from DT with high accuracy compared to conventional CD and PD and revealing a positive correlation between MVUS-VI values and the TSH-R-Ab at the time of diagnosis, as well as follow-up. Furthermore, our findings highlighted the dynamic nature of real-time quantitative MVUS-VI values in relation to the clinical course of thyrotoxicosis. In addition, TPO-Ab or Tg-Ab positivity impacts on diagnostic accuracy of MVUS-VI. Our findings suggest that MVUS and MVUS-VI can provide real-time diagnostic confirmation for physicians, particularly during the time gap when waiting for TSH-R-Ab results. This rapid assessment capability is crucial for timely decision-making in the management of thyrotoxicosis [21]. Additionally, MVUS-VI can be useful during follow-up, especially in settings where facilities for measuring TSI are limited, as it helps in monitoring the disease course effectively.
In our previous primary study, MVUS-VI distinguished GD from DT with 87.0% and 80.9% of sensitivity and specificity, respectively (AUC 0.852) when the optimal cut-off is 24.95% [5]. In this study, with a larger study population, MVUS demonstrated a similar AUC of 0.856. Although the sensitivity slightly decreased to 79.6%, the specificity improved to 84.3%, which supports the consistency in diagnostic performance across studies with slightly increased optimal cut-off value (27.35% in this study). A meta-analysis revealed that the AUC of Doppler US for distinguishing between GD and DT was 0.94, although the study measured the peak velocity of the thyroid artery using Doppler US [2]. Considering that MVUS-VI is more intuitive than measuring the velocity of the thyroid artery, MVUS would be useful in differentiating between GD and DT.
We observed the normalization of thyroid function and a substantial reduction in the vascularity upon follow-up. This underscores the importance of comprehensive initial assessment and ongoing monitoring of patients with thyrotoxicosis to tailor treatment approaches effectively. Hiromatsu et al. [22] demonstrated that CD US could effectively identify acute-stage SAT lesions lacking increased vascularity, as well as lesions with slightly enhanced vascularity during the recovery phase. In the study conducted by Baldini et al. [23], by comparing three groups of patients with untreated GD, patients with euthyroid during ATD treatment, and patients with euthyroid 1 to 2 years after discontinuing ATD, they found thyroid hypervascularization to be related to the activity of underlying autoimmune processes, not circulating thyroid hormone. Our findings are also consistent with previous studies and furthermore quantified thyroid vascularity, and demonstrated a positive correlation with the TSH-R-Ab. As the TSH-R-Ab is related to GD activity and prognosis, MVUS-VI could be useful for predicting the disease prognosis, and diagnosing GD [3,24].
While the differential diagnosis between GD and SAT is generally not difficult based on clinical or US findings with vascularity, the differentiation between GD and AIT is now challenging. In this study, the effects of TPO-Ab and Tg-Ab on thyroid vascularity were analyzed. The MVUS-VI was significantly elevated in GD, indicating its hyperactive state and extensive autoimmunity. However, there was a noticeable overlap in MVUS-VI values between GD and AIT, suggesting that while GD generally shows increased vascularity, a subset of AIT patients exhibits similar MVUS-VI levels, highlighting that MVUS-VI is not exclusively definitive for GD. This overlap was also seen in our previous study and in studies using MVUS in pediatric patients with thyrotoxicosis [5,6], emphasizing the heterogeneity within autoimmune thyroid diseases and potential variations in disease activity or pathophysiological mechanisms [25]. Thus, the specificity of MVUS-VI seemed to be lower in TPO-Ab or Tg-Ab positive patients in our study. Therefore, the cut-off of MVUS-VI can be set slightly higher in actual clinical situations, if TPO-Ab and Tg-Ab are positive: this way could be the first way to overcome the limitation. Another way to overcome this limitation is implementing a multi-step method: Stoian et al. [26] proposed a model that differentiates thyrotoxicosis using a multi-step method by additionally using sheer wave elastography on gray scale images and the peak systolic velocity of the Doppler images. MVUS-VI application to this multi-step diagnostic approach would further increase accuracy, warranting further studies for validation. Further investigations are necessary to enhance the diagnostic utility of MVUS-VI. Additionally, emerging technologies such as machine learning may offer potential benefits in addressing these diagnostic challenges [21]. Despite the mentioned limitations, MVUS-VI could help simplify thyroid vascularity quantification, thereby correlating with the disease course.
In our previous primary study, the intra- and inter-observer variability of MVUS-VI was low and showed good agreement [5]. However, the measurement of intra-observer variability was only confirmed during certain stages of obtaining US images, indicating that it does not necessarily reflect all potential biases [10]. Indeed, many biases are already embedded in the process, and the MVUS-VI figures do not account for these biases. In cases of hyperthyroidism, the entire cross-section of the thyroid is the target, so depending on where or which cross-section is chosen, operator bias can possibly occur [5]. Furthermore, even the image cross-section captured can differ based on the systolic or diastolic phase [27]. Therefore, stating that the MVUS-VI calculated from these captured images has good intra-observer variability is limited owing to the biases already incorporated into the capture. To address these limitations, research has been directed into enhancing US diagnostics and reducing intra- and inter-observer variation using machine learning [28]. Machine learning using Doppler US could be used for differentiating the etiology of thyrotoxicosis [21].
Our study has several limitations. First, the clinical symptoms of the patients were not uniformly measurable at the time of US scanning. Additionally, the dates of blood tests did not coincide with the dates of US examinations in some cases, which could have resulted in inconsistencies in correlation of the clinical data with the imaging results. Secondly, in our study, we assessed the association between the MVUS-VI and TSH-R-Ab, noting a significant decrease in MVUS-VI prior to medication discontinuation, suggesting clinical improvement. However, we did not follow patients long-term after ATD discontinuation to observe potential relapses in GD, limiting our understanding of the prognostic value of MVUS-VI over time. Additionally, while this study tracks changes in MVUS-VI over time, aligning with improvements in thyroid function, we have not established a direct correlation between these changes and specific clinical outcomes due to the lack of comprehensive clinical symptoms and follow-up data on patient recurrence rates. This limitation underscores the need for future research to define the clinical thresholds of MVUS-VI changes that correlate with meaningful clinical improvement. Lastly, we excluded thyroid nodules when selecting the ROI for MVUS-VI measurement based on prior studies that focused on differentiating thyroid nodules from thyroid cancer using superb microvascular flow imaging [29,30]. This exclusion was intended to reduce potential variability in MVUS-VI readings that could arise from nodule presence. However, considering the high prevalence of thyroid nodules in patients with GD [31], future studies should aim to include these patients to assess the applicability and accuracy of MVUS-VI across a broader spectrum of the GD population.
Despite its limitations, MVUS-VI has demonstrated considerable effectiveness in rapidly and relatively accurately differentiating between GD and DT in thyrotoxicosis. This technology offers a quantitative measure of vascularity, which is crucial for assessing the disease’s activity. Essentially, MVUS-VI shows a strong correlation with the course of the disease, reinforcing its utility in clinical settings. This correlation implies MVUS-VI’s potential not only as a diagnostic tool but also as a valuable monitor of disease progression and response to treatment.

CONFLICTS OF INTEREST

This research was financially supported by Samsung Medison Co., Ltd. (Seoul, Korea). The funding organization played no role in the design of the study, collection, analysis, and interpretation of data, writing of the manuscript, or the decision to submit the manuscript for publication.

ACKNOWLEDGMENTS

We are profoundly grateful to our research nurse, Jeongeun Lee, for her invaluable contribution and unwavering commitment to our study, particularly in the data collection and organization tasks, and Mr. Byung-so Park, Associate Director of Clinical Solution Development at Samsung Medison, for his significant support and expert guidance in our research.

Parts of this work were presented as an abstract at the 45th Annual Meeting of the European Thyroid Association, 9 to 12 September, 2023, Milan, Italy.

The authors acknowledge that the initial draft of this manuscript was written by us. We utilized ChatGPT 4.0 (OpenAI, https://chat.openai.com) for English editing and summarizing our notes, followed by further English editing by a professional expert from Editage (www.editage.co.kr).

AUTHOR CONTRIBUTIONS

Conception or design: H.S.B., C.J., J.H., D.J.L. Acquisition, analysis, or interpretation of data: H.S.B., D.J.L. Drafting the work or revising: H.S.B., D.J.L. Final approval of the manuscript: H.S.B., C.J., J.H., D.J.L.

Fig. 1.
Summary of study design. The figure depicts the participant flow and exclusion criteria for a research study and composition of cohorts. In the primary study, 114 participants initially underwent an exam, from which 17 were excluded for various reasons: eight as normal controls, two due to toxic adenomas, one due to thyroid cancer, and six due to loss of follow-up; thus, resulting in 97 participants who were analyzed in cohort 1. In the expansion study, 107 participants underwent an exam at initial thyrotoxicosis presentation. From these, 19 were excluded: two for lack of ultrasound (US) data, two due to thyroid nodules or cancer, seven for absence of thyrotoxicosis, and eight due to loss of follow-up. This left 88 participants were also analyzed in cohort 1. On the other hand, 63 participants underwent US for follow-up purposes. In the primary study, 14 individuals underwent US for follow-up, and 77 patients were finally analyzed. These patients were grouped into cohort 2 in this study.
enm-2024-2206f1.jpg
Fig. 2.
Sensitivity and specificity according to the vascularity index of microvascular ultrasound (MVUS-VI) cut-off (A) for entire study population, and (B) is for population with positive thyroid peroxidase antibody (TPO-Ab) or thyroglobulin antibody (Tg-Ab). Although the optimal cut-off value of VI for differentiating the Graves’ disease from destructive thyroiditis was the same in the entire study population or patients with positive TPO-Ab or Tg-Ab (27.35%), the specificity achieved with the optimal cut-off value was lower in patients with positive TPO-Ab or Tg-Ab. Therefore, a higher cut-off value was needed to obtain higher specificity in this population.
enm-2024-2206f2.jpg
Fig. 3.
Distribution of the vascularity index among patients diagnosed with Graves’ disease (GD), autoimmune thyroiditis (AIT), and subacute thyroiditis (SAT). The differences in the vascularity index of microvascular ultrasound (MVUS-VI) values among the three groups demonstrated statistical significance (45.75±22.2 in GD, 11.30±5.6 in SAT, and 26.36±17.7 in AIT, P<0.001). The multiple comparisons also showed statistical significance between each group (P<0.001 for GD and SAT, P<0.001 for GD and AIT, and P=0.002 for SAT and AIT). Despite this difference, the MVUS-VI values overlapped between GD and AIT. Thirteen of the patients with AIT were classified as GD based on the MVUS-VI (false positive), and 20 of the patients with GD were classified as not GD based on the MVUS-VI (false negative).
enm-2024-2206f3.jpg
Fig. 4.
Thyroid peroxidase antibody (TPO-Ab) or thyroglobulin antibody (Tg-Ab) and vascularity index of microvascular ultrasound (MVUS-VI) in patients with Graves’ disease (GD). The MVUS-VI of patients with GD demonstrated variations according to the TPO-Ab or Tg-Ab positivity. At initial diagnosis, the MVUS-VI was significant higher in patients with positive TPO-Ab or Tg-Ab, compared to that in patients with negative TPO-Ab and Tg-Ab (51.7±22.4 vs. 32.1±17.9, P=0.002). However, at the time of stable status after antithyroid drug (ATD) therapy, considering ATD discontinuation, the differences of MVUS-VI values between the two groups showed no statistical significance (25.8±16.1 vs. 22.4±10.0, P=0.662). NS, not significant, indicating no statistical significance (P≥0.05). aP<0.05 indicates statistical significance.
enm-2024-2206f4.jpg
Fig. 5.
Correlation between vascularity index of microvascular ultrasound (MVUS-VI) and thyroid-stimulating hormone (TSH) receptor antibodies. (A) At the time of Graves’ disease (GD) diagnosis, the MVUS-VI and thyrotropin-binding inhibitory immunoglobulin (TBII) showed positive correlation (r=0.422, P<0.001). (C) The thyroid-stimulating immunoglobulin (TSI) bioassay showed also positive correlation with the MVUS-VI (r=0.582, P<0.001). (B, D) The TBII and TSI bioassay also showed positive correlation with the MVUS-VI at the time before antithyroid drug discontinuation (r=0.427, P<0.001 for TBII and r=0.473, P=0.001).
enm-2024-2206f5.jpg
Table 1.
Cohort 1 Characteristics (Initial Diagnosis)
Characteristic GD (n=98) DT (n=87) P value
Age, yr 47.8±14.7 47.1±13.4 0.755
Female sex 80 (81.6) 72 (82.8) 0.988
BMI, kg/m2 21.6±3.3 22.0±2.9 0.427
fT4, ng/dL 3.6±1.9 2.9±1.7 0.007
TSH, μIU/mL 0.029±0.157 0.048±0.242 0.537
TBII, μ/L 11.6±10.1 1.6±4.7 <0.001
TSI bioassay, %a 352.1±182.6 81.5±81.0 <0.001
Thyroid volume, mL 17.8±9.8 17.2±8.5 0.656
CD 1.7±0.9 0.7±0.7 <0.001
PD 1.9±0.9 0.8±0.7 <0.001
MVUS 2.4±0.8 1.2±0.9 <0.001
MVUS-VI, % 45.8 ±22.2 18.2±14.7 <0.001

Values are expressed as mean±standard deviation or number (%).

GD, Graves’ disease; DT, destructive thyroiditis; BMI, body mass index; fT4, free thyroxine; TSH, thyroid-stimulating hormone; TBII, the competitive thyrotropin-binding inhibitory immunoglobulin assay; TSI, thyroid-stimulating immunoglobulin; CD, color Doppler; PD, power Doppler; MVUS, microvascular ultrasound; MVUS-VI, vascularity index of MVUS.

a TSI bioassay was measured in 52 individuals.

Table 2.
Cohort 2 (Follow-up) Characteristics and Comparison with Cohort 1
Characteristic Cohort 1 (n=185) Cohort 2 (n=77) P value
Age, yr 47.5±14.3 49.4±13.9 0.296
Female sex 150 (81.1) 55 (71.4) 0.119
BMI, kg/m2 21.8±3.1 22.3±3.2 0.353
fT4, ng/dL 3.3±1.9 1.7±1.3 <0.001
TSH, μIU/mL 0.038±0.200 2.325±1.623 <0.001
TBII, μ/L 7.1±9.5 2.4±5.3 <0.001
TSI bioassay, % 300.1±198.9 93.6±116.3a <0.001
Thyroid volume, mL 17.5±9.2 20.1±9.5 0.037
CD 1.2±1.0 0.8±0.6 <0.001
PD 1.4±1.0 1.0±0.7 0.002
MVUS 1.8±1.0 1.5±0.7 0.006
MVUS-VI, % 32.8±23.5 21.9±14.0 <0.001

Values are expressed as mean±standard deviation or number (%).

BMI, body mass index; fT4, free thyroxine; TSH, thyroid-stimulating hormone; TBII, the competitive thyrotropin-binding inhibitory immunoglobulin assay; TSI, thyroid-stimulating immunoglobulin; CD, color Doppler; PD, power Doppler; MVUS, microvascular ultrasound; MVUS-VI, vascularity index of MVUS.

a TSI bioassay was measured in 40 individuals in cohort 2, only measured and followed up in the Graves’ disease group.

Table 3.
Diagnostic Accuracy of Ultrasound Parameters in Cohort 1 (Initial Diagnosis)
Variable Optimal cut-off Sensitivity, % Specificity, % PPV, % NPV, % AUC 95% CI Accuracy, % P value for ROC curve power Delong’s test for two correlated ROC curves
CD 1.25 60.2 85.1 81.9 65.5 0.791 0.728–0.853 71.4 <0.001
PD 1.5 74.5 80.5 81.1 73.7 0.823 0.764–0.883 75.7 <0.001
MVUS 2 84.0 72.1 76.7 80.5 0.823 0.762–0.884 74.4 <0.001 0.469 (vs. CD)
0.997 (vs. PD)
MVUS-VI 27.35 79.6 83.9 84.8 78.5 0.855 0.799–0.910 81.1 <0.001 0.011 (vs. CD)
0.161 (vs. PD)

PPV, positive predictive value; NPV, negative predictive value; AUC, area under the curve; CI, confidence interval; ROC, receiver operating characteristic; CD, color Doppler; PD, power Doppler; MVUS, microvascular ultrasound; MVUS-VI, vascularity index of MVUS.

Table 4.
Diagnostic Performance of Ultrasound Parameters in Patients with Positive for TPO-Ab or Tg-Ab (n=103)
Variable Optimal cut-off Sensitivity, % Specificity, % PPV, % NPV, % AUC 95% CI Accuracy, % P value for ROC curve power Delong’s test for two correlated ROC curves
CD 2 64.3 80.9 80.0 65.5 0.770 0.682–0.859 71.8 <0.001
PD 2 76.8 78.7 81.1 74.0 0.827 0.746–0.907 77.7 <0.001
MVUS 2.5 74.5 80.9 82.0 73.1 0.807 0.722–0.892 77.7 <0.001 0.559 (vs. CD)
0.740 (vs. PD)
MVUS-VI 27.35 87.5 72.3 79.0 82.9 0.831 0.750–0.913 80.6 <0.001 0.074 (vs. CD)
0.877 (vs. PD)

TPO-Ab, thyroid peroxidase antibody; Tg-Ab, thyroglobulin antibody; PPV, positive predictive value; NPV, negative predictive value; AUC, area under the curve; CI, confidence interval; ROC, receiver operating characteristic; CD, color Doppler; PD, power Doppler; MVUS, microvascular ultrasound; MVUS-VI, vascularity index of MVUS.

  • 1. Badiu C. Williams textbook of endocrinology: 14th revised edition. Acta Endocrinol (Buchar) 2019;15:416.ArticlePMC
  • 2. Peng X, Wu S, Bie C, Tang H, Xiong Z, Tang S. Mean peak systolic velocity of superior thyroid artery for the differential diagnosis of thyrotoxicosis: a diagnostic meta-analysis. BMC Endocr Disord 2019;19:56.ArticlePubMedPMCPDF
  • 3. Baek HS, Lee J, Jeong CH, Lee J, Ha J, Jo K, et al. The prediction model using thyroid-stimulating immunoglobulin bioassay for relapse of Graves’ disease. J Endocr Soc 2022;6:bvac023.ArticlePubMedPMCPDF
  • 4. Kahaly GJ. Management of Graves thyroidal and extrathyroidal disease: an update. J Clin Endocrinol Metab 2020;105:3704–20.ArticlePubMedPMCPDF
  • 5. Baek HS, Park JY, Jeong CH, Ha J, Kang MI, Lim DJ. Usefulness of real-time quantitative microvascular ultrasonography for differentiation of Graves’ disease from destructive thyroiditis in thyrotoxic patients. Endocrinol Metab (Seoul) 2022;37:323–32.ArticlePubMedPMCPDF
  • 6. Bayramoglu Z, Kandemirli SG, Akyol Sari ZN, Kardelen AD, Poyrazoglu S, Bas F, et al. Superb microvascular imaging in the evaluation of pediatric Graves disease and Hashimoto thyroiditis. J Ultrasound Med 2020;39:901–9.ArticlePubMedPDF
  • 7. Donkol RH, Nada AM, Boughattas S. Role of color Doppler in differentiation of Graves’ disease and thyroiditis in thyrotoxicosis. World J Radiol 2013;5:178–83.ArticlePubMedPMC
  • 8. Rosario PW, Santos JB, Nunes NS, da Silva AL, Calsolari MR. Color flow Doppler sonography for the etiologic diagnosis of thyrotoxicosis. Horm Metab Res 2014;46:505–9.ArticlePubMed
  • 9. Alzahrani AS, Ceresini G, Aldasouqi SA. Role of ultrasonography in the differential diagnosis of thyrotoxicosis: a noninvasive, cost-effective, and widely available but underutilized diagnostic tool. Endocr Pract 2012;18:567–78.ArticlePubMed
  • 10. Lim DJ, Luo S, Kim MH, Ko SH, Kim Y. Interobserver agreement and intraobserver reproducibility in thyroid ultrasound elastography. AJR Am J Roentgenol 2012;198:896–901.ArticlePubMed
  • 11. Cannella R, Pilato G, Mazzola M, Bartolotta TV. New microvascular ultrasound techniques: abdominal applications. Radiol Med 2023;128:1023–34.ArticlePubMedPMCPDF
  • 12. Jiang ZZ, Huang YH, Shen HL, Liu XT. Clinical applications of superb microvascular imaging in the liver, breast, thyroid, skeletal muscle, and carotid plaques. J Ultrasound Med 2019;38:2811–20.ArticlePubMedPDF
  • 13. Yu J, Baek HS, Jeong C, Jo K, Lee J, Ha J, et al. The early changes in thyroid-stimulating immunoglobulin bioassay over anti-thyroid drug treatment could predict prognosis of Graves’ disease. Endocrinol Metab (Seoul) 2023;38:338–46.ArticlePubMedPMCPDF
  • 14. Arifin WN. Sample size calculator [Internet]. wnarifin.github.io; 2024 [cited 2024 Dec 30]. Available from: http://wnarifin.github.io.
  • 15. Buderer NM. Statistical methodology: I. Incorporating the prevalence of disease into the sample size calculation for sensitivity and specificity. Acad Emerg Med 1996;3:895–900.ArticlePubMed
  • 16. Vitti P, Hegedus L. Thyroid diseases: pathogenesis, diagnosis, and treatment; Cham: Springer International Publishing; 2018. Chapter 7, Hashimoto’s thyroiditis. p. 205–47.
  • 17. Luo Y, Kawashima A, Ishido Y, Yoshihara A, Oda K, Hiroi N, et al. Iodine excess as an environmental risk factor for autoimmune thyroid disease. Int J Mol Sci 2014;15:12895–912.ArticlePubMedPMC
  • 18. Nahm FS. Receiver operating characteristic curve: overview and practical use for clinicians. Korean J Anesthesiol 2022;75:25–36.ArticlePubMedPMCPDF
  • 19. Baratloo A, Hosseini M, Negida A, El Ashal G. Part 1: Simple definition and calculation of accuracy, sensitivity and specificity. Emerg (Tehran) 2015;3:48–9.PubMedPMC
  • 20. DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics 1988;44:837–45.ArticlePubMed
  • 21. Baek HS, Kim J, Jeong C, Lee J, Ha J, Jo K, et al. Deep learning analysis with Gray scale and Doppler ultrasonography images to differentiate Graves’ disease. J Clin Endocrinol Metab 2024;109:2872–81.ArticlePubMedPDF
  • 22. Hiromatsu Y, Ishibashi M, Miyake I, Soyejima E, Yamashita K, Koike N, et al. Color Doppler ultrasonography in patients with subacute thyroiditis. Thyroid 1999;9:1189–93.ArticlePubMed
  • 23. Baldini M, Castagnone D, Rivolta R, Meroni L, Pappalettera M, Cantalamessa L. Thyroid vascularization by color doppler ultrasonography in Graves’ disease: changes related to different phases and to the long-term outcome of the disease. Thyroid 1997;7:823–8.ArticlePubMed
  • 24. Kahaly GJ, Diana T, Kanitz M, Frommer L, Olivo PD. Prospective trial of functional thyrotropin receptor antibodies in Graves disease. J Clin Endocrinol Metab 2020;105:e1006–14.ArticlePubMedPDF
  • 25. Corona G, Biagini C, Rotondi M, Bonamano A, Cremonini N, Petrone L, et al. Correlation between, clinical, biochemical, color Doppler ultrasound thyroid parameters, and CXCL-10 in autoimmune thyroid diseases. Endocr J 2008;55:345–50.ArticlePubMed
  • 26. Stoian D, Borlea A, Moisa-Luca L, Paul C. Multiparametric ultrasound-based assessment of overt hyperthyroid diffuse thyroid disease. Front Endocrinol (Lausanne) 2023;14:1300447.ArticlePubMedPMC
  • 27. Lan Y, Li N, Song Q, Zhang MB, Luo YK, Zhang Y. Correlation and agreement between superb micro-vascular imaging and contrast-enhanced ultrasound for assessing radiofrequency ablation treatment of thyroid nodules: a preliminary study. BMC Med Imaging 2021;21:175.ArticlePubMedPMCPDF
  • 28. Zhou B, Yang X, Curran WJ, Liu T. Artificial intelligence in quantitative ultrasound imaging: a survey. J Ultrasound Med 2022;41:1329–42.ArticlePubMedPDF
  • 29. Negroni D, Maddalena G, Bono R, Abruzzese F, Cesano S, Conte P, et al. Superb microvascular imaging (SMI) and elastosonography in thyroid nodule: diagnostic value in a real-time cohort. J Ultrasound 2024;27:613–20.ArticlePubMedPMCPDF
  • 30. Ma G, Chen L, Wang Y, Luo Z, Zeng Y, Wang X, et al. Application of microvascular ultrasound-assisted thyroid imaging report and data system in thyroid nodule risk stratification. Insights Imaging 2024;15:230.ArticlePubMedPMCPDF
  • 31. Kraimps JL, Bouin-Pineau MH, Mathonnet M, De Calan L, Ronceray J, Visset J, et al. Multicentre study of thyroid nodules in patients with Graves’ disease. Br J Surg 2000;87:1111–3.ArticlePubMedPDF

Figure & Data

References

    Citations

    Citations to this article as recorded by  

      Figure
      • 0
      • 1
      • 2
      • 3
      • 4
      Microvascular Ultrasonography Vascularity Index as a Rapid and Simplified Assessment Tool for Differentiating Graves’ Disease from Destructive Thyroiditis and Managing Thyrotoxicosis
      Image Image Image Image Image
      Fig. 1. Summary of study design. The figure depicts the participant flow and exclusion criteria for a research study and composition of cohorts. In the primary study, 114 participants initially underwent an exam, from which 17 were excluded for various reasons: eight as normal controls, two due to toxic adenomas, one due to thyroid cancer, and six due to loss of follow-up; thus, resulting in 97 participants who were analyzed in cohort 1. In the expansion study, 107 participants underwent an exam at initial thyrotoxicosis presentation. From these, 19 were excluded: two for lack of ultrasound (US) data, two due to thyroid nodules or cancer, seven for absence of thyrotoxicosis, and eight due to loss of follow-up. This left 88 participants were also analyzed in cohort 1. On the other hand, 63 participants underwent US for follow-up purposes. In the primary study, 14 individuals underwent US for follow-up, and 77 patients were finally analyzed. These patients were grouped into cohort 2 in this study.
      Fig. 2. Sensitivity and specificity according to the vascularity index of microvascular ultrasound (MVUS-VI) cut-off (A) for entire study population, and (B) is for population with positive thyroid peroxidase antibody (TPO-Ab) or thyroglobulin antibody (Tg-Ab). Although the optimal cut-off value of VI for differentiating the Graves’ disease from destructive thyroiditis was the same in the entire study population or patients with positive TPO-Ab or Tg-Ab (27.35%), the specificity achieved with the optimal cut-off value was lower in patients with positive TPO-Ab or Tg-Ab. Therefore, a higher cut-off value was needed to obtain higher specificity in this population.
      Fig. 3. Distribution of the vascularity index among patients diagnosed with Graves’ disease (GD), autoimmune thyroiditis (AIT), and subacute thyroiditis (SAT). The differences in the vascularity index of microvascular ultrasound (MVUS-VI) values among the three groups demonstrated statistical significance (45.75±22.2 in GD, 11.30±5.6 in SAT, and 26.36±17.7 in AIT, P<0.001). The multiple comparisons also showed statistical significance between each group (P<0.001 for GD and SAT, P<0.001 for GD and AIT, and P=0.002 for SAT and AIT). Despite this difference, the MVUS-VI values overlapped between GD and AIT. Thirteen of the patients with AIT were classified as GD based on the MVUS-VI (false positive), and 20 of the patients with GD were classified as not GD based on the MVUS-VI (false negative).
      Fig. 4. Thyroid peroxidase antibody (TPO-Ab) or thyroglobulin antibody (Tg-Ab) and vascularity index of microvascular ultrasound (MVUS-VI) in patients with Graves’ disease (GD). The MVUS-VI of patients with GD demonstrated variations according to the TPO-Ab or Tg-Ab positivity. At initial diagnosis, the MVUS-VI was significant higher in patients with positive TPO-Ab or Tg-Ab, compared to that in patients with negative TPO-Ab and Tg-Ab (51.7±22.4 vs. 32.1±17.9, P=0.002). However, at the time of stable status after antithyroid drug (ATD) therapy, considering ATD discontinuation, the differences of MVUS-VI values between the two groups showed no statistical significance (25.8±16.1 vs. 22.4±10.0, P=0.662). NS, not significant, indicating no statistical significance (P≥0.05). aP<0.05 indicates statistical significance.
      Fig. 5. Correlation between vascularity index of microvascular ultrasound (MVUS-VI) and thyroid-stimulating hormone (TSH) receptor antibodies. (A) At the time of Graves’ disease (GD) diagnosis, the MVUS-VI and thyrotropin-binding inhibitory immunoglobulin (TBII) showed positive correlation (r=0.422, P<0.001). (C) The thyroid-stimulating immunoglobulin (TSI) bioassay showed also positive correlation with the MVUS-VI (r=0.582, P<0.001). (B, D) The TBII and TSI bioassay also showed positive correlation with the MVUS-VI at the time before antithyroid drug discontinuation (r=0.427, P<0.001 for TBII and r=0.473, P=0.001).

      Fig. 1.

      Fig. 2.

      Fig. 3.

      Fig. 4.

      Fig. 5.

      Microvascular Ultrasonography Vascularity Index as a Rapid and Simplified Assessment Tool for Differentiating Graves’ Disease from Destructive Thyroiditis and Managing Thyrotoxicosis
      Characteristic GD (n=98) DT (n=87) P value
      Age, yr 47.8±14.7 47.1±13.4 0.755
      Female sex 80 (81.6) 72 (82.8) 0.988
      BMI, kg/m2 21.6±3.3 22.0±2.9 0.427
      fT4, ng/dL 3.6±1.9 2.9±1.7 0.007
      TSH, μIU/mL 0.029±0.157 0.048±0.242 0.537
      TBII, μ/L 11.6±10.1 1.6±4.7 <0.001
      TSI bioassay, %a 352.1±182.6 81.5±81.0 <0.001
      Thyroid volume, mL 17.8±9.8 17.2±8.5 0.656
      CD 1.7±0.9 0.7±0.7 <0.001
      PD 1.9±0.9 0.8±0.7 <0.001
      MVUS 2.4±0.8 1.2±0.9 <0.001
      MVUS-VI, % 45.8 ±22.2 18.2±14.7 <0.001
      Characteristic Cohort 1 (n=185) Cohort 2 (n=77) P value
      Age, yr 47.5±14.3 49.4±13.9 0.296
      Female sex 150 (81.1) 55 (71.4) 0.119
      BMI, kg/m2 21.8±3.1 22.3±3.2 0.353
      fT4, ng/dL 3.3±1.9 1.7±1.3 <0.001
      TSH, μIU/mL 0.038±0.200 2.325±1.623 <0.001
      TBII, μ/L 7.1±9.5 2.4±5.3 <0.001
      TSI bioassay, % 300.1±198.9 93.6±116.3a <0.001
      Thyroid volume, mL 17.5±9.2 20.1±9.5 0.037
      CD 1.2±1.0 0.8±0.6 <0.001
      PD 1.4±1.0 1.0±0.7 0.002
      MVUS 1.8±1.0 1.5±0.7 0.006
      MVUS-VI, % 32.8±23.5 21.9±14.0 <0.001
      Variable Optimal cut-off Sensitivity, % Specificity, % PPV, % NPV, % AUC 95% CI Accuracy, % P value for ROC curve power Delong’s test for two correlated ROC curves
      CD 1.25 60.2 85.1 81.9 65.5 0.791 0.728–0.853 71.4 <0.001
      PD 1.5 74.5 80.5 81.1 73.7 0.823 0.764–0.883 75.7 <0.001
      MVUS 2 84.0 72.1 76.7 80.5 0.823 0.762–0.884 74.4 <0.001 0.469 (vs. CD)
      0.997 (vs. PD)
      MVUS-VI 27.35 79.6 83.9 84.8 78.5 0.855 0.799–0.910 81.1 <0.001 0.011 (vs. CD)
      0.161 (vs. PD)
      Variable Optimal cut-off Sensitivity, % Specificity, % PPV, % NPV, % AUC 95% CI Accuracy, % P value for ROC curve power Delong’s test for two correlated ROC curves
      CD 2 64.3 80.9 80.0 65.5 0.770 0.682–0.859 71.8 <0.001
      PD 2 76.8 78.7 81.1 74.0 0.827 0.746–0.907 77.7 <0.001
      MVUS 2.5 74.5 80.9 82.0 73.1 0.807 0.722–0.892 77.7 <0.001 0.559 (vs. CD)
      0.740 (vs. PD)
      MVUS-VI 27.35 87.5 72.3 79.0 82.9 0.831 0.750–0.913 80.6 <0.001 0.074 (vs. CD)
      0.877 (vs. PD)
      Table 1. Cohort 1 Characteristics (Initial Diagnosis)

      Values are expressed as mean±standard deviation or number (%).

      GD, Graves’ disease; DT, destructive thyroiditis; BMI, body mass index; fT4, free thyroxine; TSH, thyroid-stimulating hormone; TBII, the competitive thyrotropin-binding inhibitory immunoglobulin assay; TSI, thyroid-stimulating immunoglobulin; CD, color Doppler; PD, power Doppler; MVUS, microvascular ultrasound; MVUS-VI, vascularity index of MVUS.

      TSI bioassay was measured in 52 individuals.

      Table 2. Cohort 2 (Follow-up) Characteristics and Comparison with Cohort 1

      Values are expressed as mean±standard deviation or number (%).

      BMI, body mass index; fT4, free thyroxine; TSH, thyroid-stimulating hormone; TBII, the competitive thyrotropin-binding inhibitory immunoglobulin assay; TSI, thyroid-stimulating immunoglobulin; CD, color Doppler; PD, power Doppler; MVUS, microvascular ultrasound; MVUS-VI, vascularity index of MVUS.

      TSI bioassay was measured in 40 individuals in cohort 2, only measured and followed up in the Graves’ disease group.

      Table 3. Diagnostic Accuracy of Ultrasound Parameters in Cohort 1 (Initial Diagnosis)

      PPV, positive predictive value; NPV, negative predictive value; AUC, area under the curve; CI, confidence interval; ROC, receiver operating characteristic; CD, color Doppler; PD, power Doppler; MVUS, microvascular ultrasound; MVUS-VI, vascularity index of MVUS.

      Table 4. Diagnostic Performance of Ultrasound Parameters in Patients with Positive for TPO-Ab or Tg-Ab (n=103)

      TPO-Ab, thyroid peroxidase antibody; Tg-Ab, thyroglobulin antibody; PPV, positive predictive value; NPV, negative predictive value; AUC, area under the curve; CI, confidence interval; ROC, receiver operating characteristic; CD, color Doppler; PD, power Doppler; MVUS, microvascular ultrasound; MVUS-VI, vascularity index of MVUS.

      Table 1.

      Table 2.

      Table 3.

      Table 4.

      Endocrinol Metab : Endocrinology and Metabolism
      © Korean Endocrine Society.
      TOP