Downloaded from http://journals.lww.com/ultrasound-quarterly by BhDMf5ePHKav1zEoum1tQfN4a+kJLhEZgbsIHo4XMi0hCywCX1AWnYQp/IlQrHD3i3D0OdRyi7TvSFl4Cf3VC1y0abggQZXdgGj2MwlZLeI= on 03/10/2021 Downloadedfrom http://journals.lww.com/ultrasound-quarterlyby BhDMf5ePHKav1zEoum1tQfN4a+kJLhEZgbsIHo4XMi0hCywCX1AWnYQp/IlQrHD3i3D0OdRyi7TvSFl4Cf3VC1y0abggQZXdgGj2MwlZLeI=on 03/10/2021
Quantitative Assessment of Thyroid Gland Vascularization
With Vascularization Index Using Color
Superb Microvascular Imaging in Pediatric Patients
With Hashimoto Thyroiditis
Mehmet Sedat Durmaz, MD,* Nesibe Akyürek, MD,
† Turgay Kara, MD,‡ Fatih Ateş, MD,‡
Bora Özbak
ır, MD,§ Funda Gökgöz Durmaz, MD,∥ Seda Soğukpınar Karaağaç, MD,‡
and Mehmet Öztürk, MD*
Abstract: The study aimed to investigate the effectiveness of the vas-cularization index (VI) obtained using superb microvascular imaging (SMI) technique in the diagnosis of Hashimoto thyroiditis (HT). The thyroid glands of 80 patients with HT and 107 healthy, asymptomatic participants were examined using SMI. The thyroid parenchyma echogenicity was evaluated, and the thyroid gland volume was mea-sured. Vascularization index measurements were performed by manu-ally drawing the contours of the thyroid parenchyma using the free region of interest with color 2-dimensional SMI VI mode. The quanti-tative VI values of the patients and the asymptomatic group were com-pared. Correlations between VI values and thyroid autoantibodies and thyroid hormone levels were analyzed. The mean VI value of the thy-roid gland was 4.74% ± 1.96% in the asymptomatic group and 12.45% ± 5.87% in HT patients with a statistically significant differ-ence (P < 0.001). Hashimoto thyroiditis can be diagnosed with 86.3% sensitivity and 82.2% specificity when 6.00% VI value was designated as the cutoff value. There was a positive significant correlation between the VI value and the thyroid-stimulating hormone, antithyroglobulin antibodies, anti–thyroid peroxidase antibody levels (P < 0.05); how-ever, no significant correlation was found between the VI values and thyroglobulin and free thyroxine levels (P > 0.05). There was a signif-icant negative correlation between the VI values and the parenchyma echogenicity and positive significant correlation between the thyroid
gland volume and the antithyroglobulin antibody and anti–thyroid per-oxidase antibody levels (P < 0.05). The VI obtained using the SMI technique can be effectively used as an imaging method for the diagno-sis of HT because of its high sensitivity and specificity in representing objective, quantitative numerical values.
Key Words: 2-dimensional cSMI vascularization index, Hashimoto thyroiditis, ultrasonography
(Ultrasound Quarterly 2019;35: 281–289)
T
hyroiditis is an inflammation of the thyroid gland.Thyroid-itis is a generic term used to describe an array of clinical en-tities affecting the thyroid gland. There are many possible causes of thyroiditis. Autoimmune thyroid disorders cover a va-riety of conditions that may share some common clinical, bio-chemical, and ultrasonography (US) features that can lead to diagnostic challenges. Hashimoto thyroiditis (HT) is the most common autoimmune thyroid disorder in which antibodies di-rected against the thyroid paranchyma cause chronic inflamma-tion. Hashimoto thyroiditis affects women more than men and can affect people of any age, but it occurs most commonly in
middle-aged women.1–3 Diagnosis can be obtained based on
the presence of circulating antithyroid autoantibodies (anti– thyroid peroxidase antibodies [TPOAbs] and antithyroglobulin
antibodies [TgAbs]), US findings, and clinical symptoms.4,5
Thyroid autoantibodies are responsible for thyroid function im-pairment and morphological damage. Over the years, an in-crease in antithyroid antibody titers has been detected in most
HT patients.5,6 Ultrasonography is widely used in diagnosis
and follow-up of HT. Gray-scale US findings, heterogeneous echotexture, decreased parenchymal echogenicity, presence of hypoechoic micronodules, and lobulated contour have high
specificity, but they have limited sensitivity for diagnosis.7,8
Conventional Doppler imaging (CDI) techniques (color and power Doppler vascular imaging) are widely used to evaluate HT. In many studies, it has been reported that the diagnostic ac-curacy of HT is increased by the combined use of US and CDI
techniques.5,9–11Increased vascularity can be seen during the
early stages of HT; however, in the long term, hypovascularity may appear as a result of CDI techniques examination
espe-cially in atrophic forms of HT.5Conventional Doppler imaging
Received for publication October 23, 2018; accepted January 11, 2019. *Department of Radiology, Medicine Faculty, Selçuk University; and
†Depart-ments of Pediatric Endocrinology and‡Radiology, Training and Research Hospital, Konya Health Sciences University, Konya; §Department of Radi-ology, Gynecology-Obstetrics and Pediatrics Hospital, Isparta; and∥Karatay Community Health Center Family Medicine, Konya, Turkey.
The authors declare no conflict of interest.
The approval of the ethics committee was obtained before the initiation of the study. All the participants were previously informed about the study, and a written informed consent was signed by each participant or parents of children before the ultrasonography examination. All the procedures performed in the studies on humans were in line with the ethical institutional and/or national research committee standards and with the 1964 Helsinki Declaration and its amendments or comparable ethical standards.
Address correspondence to: Mehmet Sedat Durmaz, MD, Department of Radiology, Medicine Faculty, Selçuk University, Necip Fazıl Mahallesi, Fatih Cad. No. 4/1, Meram, Konya, Turkey 42090 (e‐mail: [email protected]).
Copyright © 2019 Wolters Kluwer Health, Inc. All rights reserved. DOI: 10.1097/RUQ.0000000000000430
ity using these CDI techniques12–15can be qualitative (visual as-sessment of thyroid vascularity).
Superb microvascular imaging (SMI) is a new vascular imaging technique that was developed recently to overcome the limitations of CDI techniques. Superb microvascular imag-ing effectively uses a new adaptive algorithm to analyze motion artifacts, separate the flow signals from overlaying tissue mo-tion artifacts, remove tissue momo-tion, and reveal true blood flow. By overreaching this goal, SMI can detect microvascular blood flow as well as and low-velocity blood flow using
high-resolution images.12,13,16 Superb microvascular imaging is a
powerful and smart algorithm that preserves even the subtlest
low-flow components with unmatched detail and definition.12
Superb microvascular imaging can be operated in 2 modes: color SMI (cSMI) and monochrome SMI. The cSMI mode si-multaneously displays a conventional gray-scale US with color-encoded Doppler signals. The monochrome SMI mode improves the visibility of vascular structures by eliminating the background
signals and focusing only on the vasculature signals.12,15
Superb microvascular imaging allows quantitative data analysis using the vascularization index (VI) parameter. The VI reflects the proportion of blood flow within the tissue, which can be calculated by automated application that allows the quan-tification of the signals by establishing a ratio of colored pixels (blood flow) to all of the pixels (total number of both color and gray pixels) within the region of interest (ROI). Its unit is
expressed as a numerical value ranging from 0 to 100.4,17
In our study, the vascularity of thyroid gland was evaluated objectively with quantitative numerical values using the VI values obtained as a result of drawing the whole of the thyroid gland pa-renchyma into the free ROI on cSMI mode (2-dimensional cSMI VI [2DcSMIVI]) on the longitudinal and transverse planes in HT patients and asymptomatic groups. This technique can play an important role in the monitoring of thyroid vascularity with reliable quantitative numerical values. This study aimed to quantitatively assess the thyroid parenchyma vascularity by 2DcSMIVI in patients with HT and correlate the 2DcSMIVI values to the asymptomatic control group with serum hormone and autoantibody levels.
MATERIALS AND METHODS
This study was conducted at our institution between November 2017 and October 2018. The study was approved by the local research ethics committee. All the participants were informed, and written informed consent letters were signed by the participants or their parents before the US examination. All procedures performed in this study involving human partic-ipants were in accordance with the ethical standards specified by the institutional and/or national research committee and with the Helsinki Declaration and its later amendments or compara-ble ethical standards.
ings of thyroiditis patterns in the US examination. All serologi-cal and hormone tests were completed within 1 week in the same laboratory before the measurement of the thyroid gland VI. The normal reference ranges were as follows: 0.70 to 1.74 ng/dL free thyroxine (fT4), 0.57 to 5.6 mIU/L thyroid-stimulating hormone (TSH), 0 to 40 IU/mL TgAbs and 0 to 35 IU/mL TPOAbs, and 1.6 to 60 ng/mL thyroglobulin.
At the time of the study, most of the patients who had either normal thyroid function or hypothyroidism-hyperthyroidism had been receiving medical treatment. Patients with newly diagnosed HT were also included in the study, but these patients were few in number. The patients with Graves disease, thyroid nodules in the US, or radioactive iodine treatment history and who have undergone thyroid gland operation were not included in the study. The control group was composed of patients who were admitted to the endocrinology outpatient clinic for minor acute illnesses, such as for problems other than that of the thyroid gland. They did not have any history of chronic disease or chronic drug use. A control appointment was established to as-sess the thyroid function tests and US findings. The control group was completely asymptomatic with no signs or symptoms of HT. The thyroid US findings and hormone levels (fT4, TSH) were normal, and no history of any thyroid disease was reported in these asymptomatic, euthyroid control group. Biometric data, such as sex, age, weight, and height, were retrieved, and the body mass index (BMI) of each participant was calculated prior to the US and 2DcSMIVI examination.
Ultrasound and 2DcSMIVI Technique
The thyroid glands of all participants were assessed using the B-mode imaging and cSMI. All US and cSMI vascular aminations were performed by a radiologist with 13 years of ex-perience in US and 3 years of exex-perience in cSMI. The US and cSMI examinations of the thyroid gland were performed with a high-frequency (4–14 MHz) linear array transducer Canon Aplio 500 (Canon Medical System Corporation, Tokyo, Japan). In all the participants, the examinations performed in the supine posi-tion with the neck slightly extended began with a standard gray-scale US examination of the thyroid gland. Three
dimen-sions (length [L] width [W] height [H]) of each thyroid lobe
were measured in millimeters. Then the volume of the thyroid gland was automatically calculated by the built-in software of
the US device using the formula L W H 0.523 in the right
and left lobes of the thyroid gland (Fig. 1). The total volume of the thyroid gland was calculated. The total thyroid volume was defined as the sum of the right and left lobe volumes. The vol-ume of the isthmus was ignored.
Typically, a normal thyroid gland presents a slightly higher or similar echogenicity when compared with a submandibular gland on the gray-scale US. The thyroid parenchyma is consid-ered as hypoechogenic in the presence of fewer echogenicity than
the submandibular gland or similar to the prethyroid muscles.18 The degree of the thyroid parenchyma echogenicity was subjec-tively analyzed by 2 independent radiologists who agreed to most of the cases when compared with the echogenicity of the sub-mandibular gland classified in 4 groups as hyperechogenic, isoechogenic, hypoechogenic, or marked hypoechogenic. Dis-crepancies for the judgment of echogenicity were found in only 8 patients, who were further reexamined by the 2 radiologists together. An agreement was then reached for their classification. The presence of echogenic bands in the thyroid parenchyma and lymph node, adjacent to thyroid gland, was also reported.
Following the gray-scale US examination, all participants underwent thyroid gland vascular imaging with the same US system equipped with cSMI function using a pulse repetition frequency set at 150 to 180 Hz. Each participant was examined under standard circumstances. The VI measurements were per-formed on cSMI images. During the cSMI examination, the im-ages of the thyroid gland were obtained by contacting the US probe too lightly without pressure on the skin using an appropri-ate amount of the gel. During the examination, the patient was requested to breathe normally and was asked not to move or swallow. Color SMI vascular examination was performed as much as possible on a magnified view. According to the plane, the thyroid gland was examined using the cSMI for 5 seconds to measure 2DcSMIVI, following which the image was frozen. The VI calculation program in the ultrasound device was acti-vated. Quantitative VI values were calculated as much as possi-ble on this magnified view by manually drawing the margins of the entire thyroid gland using free ROI. The gray-scale pixels were eliminated, the color pixels were calculated automatically, and a quantitative evaluation of the 2DcSMIVI was obtained as a percentage of the number of color pixels to the total number of pixels on the examination plane within the ROI. In order to ob-tain the 2DcSMIVI value, the thyroid gland was manually drawn using free ROI, a 5-second image was examined back-ward, and the highest VI value calculated during this period was noted. The VI measurements were performed in the thyroid gland as 3 separate images, separately in the right and left lobes on the longitudinal plane, following which the VIs of both thy-roid lobes were calculated together in a single image in the transverse plane (Fig. 2). Three VI measurements were calcu-lated in each of the 3 planes, and the average VI value of these 3 measurements was calculated for each plane (Fig. 3). The
average time for the thyroid gland examination (US and calcu-late 2DcSMIVI) was about 8 minutes. The obtained images were transferred to our digital US database and later stored on
computer media (23-inch, full HD, 1920 1080 resolution,
Dell, USA). The 2DcSMIVI numerical values of the right and left lobes on the longitudinal plane were collected and divided into two to obtain a mean VI value for the entire thyroid gland on the longitudinal plane (right mean VI value on longitudinal plane + left mean VI value on longitudinal plane/2). Following this, the mean longitudinal plane VI was summed with the nu-merical value of the VI obtained from the transverse plane of the whole thyroid gland. The sum value was divided into two, and the mean VI value of all thyroid glands was calculated. (The mean VI of thyroid gland = [the mean VI of thyroid gland in longitudinal plane + the mean VI of thyroid gland in transverse plane]/2) (Fig. 4). Thus, we aimed to increase the accuracy of the numerical VI values.
The VI values obtained from the symptomatic and asymptomatic groups were compared. The goal was to identify those VI values that could exclude the diagnosis of HT in the asymptomatic group and reveal the limit of the VI values sug-gesting HT for the symptomatic patients group. The volumes of the thyroid gland obtained from the symptomatic and asymp-tomatic group and VI values were compared. Correlation analy-sis was performed among the laboratory values (TSH, fT4, thyroglobulin, TPOAbs, TgAbs) and the VI values.
Statistical Analysis
Statistical analysis was performed using the Statistical Package for the Social Sciences version 24 (IBM, Armonk, NY) software to evaluate the data. Descriptive statistics were expressed as the mean, SD, minimum-maximum values, per-centile, and frequency. The Kolmogorov-Smirnov test was used to determine the normal distribution of continuous variables.
Definitive statistics related to the variables were evaluated.χ2
And 1-way analysis of variance tests were used to evaluate the differences between the 2 groups. Mann-Whitney U test was used for a comparison of the VI values between patients with HT and the control group. The Pearson correlation analysis was used to evaluate the relationship between the VI values and serum autoantibody and thyroid hormone levels. The Pearson correlation analysis was also used to evaluate the relationship between the thyroid gland volume and serum autoantibody
FIGURE 1. Measurement of the 3 dimensions and the volume of the right and left thyroid gland lobe on the transverse and longitudinal sections of a 14-year-old girl. The total thyroid volume (right lobe volume [7.8 cm3] + left lobe volume [6.7 cm3]) were measured as 14.5cm3.
and thyroglobulin levels. Friedman test was used to evaluate the differences between the VI values of different measurements planes and the calculated mean VI values (right lobe VI in lon-gitudinal plane, left lobe VI in lonlon-gitudinal plane, mean VI in
longitudinal plane, VI of thyroid gland in transverse plane, and mean VI of thyroid gland). P < 0.05 was considered statis-tically significant with a 95% confidence level. The receiver operating characteristic (ROC) curve analysis was used to
FIGURE 3. A 10-year-old symptomatic woman with HT. Quantitative 2DcSMIVI values were measured by manually drawing the contours of the thyroid gland structure, with free ROI, on the cSMI mode, in the longest left lobe longitudinal plane, and a 5-second image was examined backward, and the highest VI value calculated during this period was noted. Three 2DcSMIVI values obtained by this method and the average VI values of these 3 measurements were calculated for this plane. The mean VI value for left lobe in longitudinal plane was 17.3% (16:5þ16:7þ18:73 Þ:
FIGURE 2. A 15-year-old symptomatic woman with HT. Quantitative 2DcSMIVI values were measured by manually drawing the contours of the thyroid gland structure, on the cSMI mode, in the longest transverse plane (A), right lobe (B), and left lobe (C) longitudinal plane. The quantitative VI values were measured as 14.1% in transverse plane (A), 12.0% in right lobe longitudinal plane, 15.6% in left lobe longitudinal plane.
determine the best cutoff value for the VI values in patients with HT and the asymptomatic group. Sensitivity, specificity, posi-tive predicposi-tive value (PPV), negaposi-tive predicposi-tive value (NPV), and accuracy were calculated.
RESULTS
The participants were between the ages of 6 and 18 years (mean, 12.37 ± 3.30 years). The symptomatic group included 62 females (77.5%) and 18 males (22.5%) ranging in age from 6 to 18 years (mean, 12.38 ± 3.32 years), and the asymptomatic group included 76 females (71.0%) and 31 males (29.0%) rang-ing in age from 5 to 17 years (mean, 11.14 ± 3.25 years). The
BMI of the symptomatic group was 20.79 ± 4.72 kg/m2, whereas
the BMI of the asymptomatic group was 20.01 ± 4.82 kg/m2. The
was no statistically significant difference between the 2 groups in terms of sex (P = 0.401), age (P = 0.078), and BMI (P = 0.270). In the asymptomatic group, 4 (3.7%) of the patients had lymph nodes adjacent to thyroid gland with reactional features, and one of them (0.9%) had an echogenic band in the thyroid gland. In the symptomatic group, 60 patients (75.0%) had lymph nodes with reactional features, and 43 of them (53.8%) had an echogenic band. There were statistically significant dif-ferences between the 2 groups, both in terms of reactive lymph nodes adjacent to thyroid glands and echogenic bands in the thyroid gland (P < 0.001).
The mean thyroid volume and VI values obtained from the patient and control group are summarized in Table 1. Thyroid volume and VI values were compared in terms of the significant difference between the patient and the control groups. Because
the data (Kolmogorov-Smirnov test) did not conform to normal distribution, the Mann-Whitney U test was used as a nonparamet-ric test. Patients with HT had significantly higher VI values (P < 0.001) and higher thyroid volumes (P < 0.001). Addition-ally, the mean VI values of the thyroid gland were significantly higher in the symptomatic group when compared with the asymptomatic group (P < 0.001).
Receiver operating characteristic curves were plotted for VI values based on the presence of HT, and the area under the curve was calculated. The cutoff values of VI for HT and sensi-tivity, specificity, PPV, NPV, and the diagnostic accuracy of these cutoff values are shown in Table 2. The ROC curve anal-ysis of the VI values in HT is shown in Figure 5.
Friedman test was used to evaluate the differences be-tween the VI values of different measurements planes and the calculated mean VI values (right lobe VI in longitudinal plane, left lobe VI in longitudinal plane, mean VI in longitudinal plane, VI of thyroid gland in transverse plane, mean VI of thyroid gland). There were no significant differences between these 5
obtained VI values (P = 0.056,χ2= 64.8).
The mean parameters of the thyroid function of the symptomatic patients and asymptomatic group are presented in Table 3.
The correlation analysis between the laboratory values (TSH, fT4, thyroglobulin, TPOAbs, TgAbs) and VI values is summarized in Table 4. There was a positive significant correla-tion between the VI values and TSH, TgAbs, and TPOAbs levels (P < 0.05); however, no significant correlation was found between VI values and fT4 and thyroglobulin levels (P > 0.05).
FIGURE 4. A 12-year-old symptomatic woman with HT. Quantitative 2DcSMIVI values were in the longest transverse plane
(VI = 20.3%) (A), right lobe (VI = 15.7%) (B), and left lobe (VI = 16.1%) (C) longitudinal plane, using the free ROI. The mean quantitative VI values for thyroid lobe were measured as 18.1%(20:3þ 15:7þ16:1½ð 2 Þ=2Þ:
TABLE 1. The Volume and VI Values of Thyroid Gland in Patients With HT and Asymptomatic Group
Symptomatic Group, % Asymptomatic Group, % P Right lobe volume 7.05 ± 5.50 2.85 ± 1.60 <0.001* Left lobe volume 6.02 ± 4.58 2.55 ± 1.39 <0.001* Total volume 12.87 ± 9.07 5.40 ± 2.91 <0.001* Isthmus thickness 3.32 ± 1.48 2.19 ± 0.66 <0.001* Right thyroid lobe mean VI in longitudinal plane 12.43 ± 6.33 4.59 ± 2.34 <0.001* Left thyroid lobe mean VI in longitudinal plane 11.83 ± 6.62 4.23 ± 2.73 <0.001* The mean VI of thyroid gland in longitudinal plane (right long + left long/2) 12.13 ± 6.16 4.41 ± 2.21 <0.001* The mean VI of thyroid gland in transverse plane 12.78 ± 6.14 5.08 ± 2.31 <0.001* The mean VI of thyroid gland 12.45 ± 5.87 (range, 4.08–31.98) 4.74 ± 1.96 (range, 1.03–10.75) <0.001*
There was a positive significant correlation between the total
thyroid volume and TgAbs and TPOAbs levels (P ≤ 0.001);
however, no significant correlation was found between the total thyroid volume and thyroglobulin levels (P > 0.05) (Table 4).
We found decreased thyroid parenchyma echogenicity in patients with HT. There was a statistically significant difference between the asymptomatic group and the patient with HT group in terms of thyroid parenchyma echogenicity (P < 0.001). As the thyroid parenchyma echogenicity decreased, the VI values increased. The mean VI values according to the thyroid gland parenchyma echogenicity are summarized in Table 5.
There was a significant correlation between the thyroid pa-renchymal echogenicity and VI in patients with HT (P = 0.011). There was no significant correlation between the thyroid pa-renchymal echogenicity and VI in the asymptomatic group (P > 0.05). Correlation analysis revealed a negative correla-tion between the thyroid parenchymal echogenicity and the VI in the correlation between both patient and control groups (P < 0.001) (Table 6).
DISCUSSION
There are several studies reporting that CDI techniques have significant limitations with regard to the assessment of
vas-cularity.12,13,15,19In the current literature, the SMI has been
re-ported as an innovative technology that enables tissue vessel imaging without using a contrast medium. The SMI technique reduces and can significantly distinguish motion artifacts and
allows a visualization of low-velocity blood flow in
capil-laries.12,13,15,16,19Superb microvascular imaging is more
sensi-tive compared with the CDI techniques when demonstrating abnormal thyroid gland parenchymal vascularity in HT. The SMI could be promising in the HT diagnosis and management because of a higher sensitivity to the inflammatory processes and the presence of the quantitative evaluation obtained via VI
application.4 In this study, we evaluated the thyroid gland
vascularity by 2DcSMIVI in patients with HT and in an asymptomatic control group. The present study showed that the mean VI values of the thyroid gland in patients with HT (12.45% ± 5.87%) were significantly higher than those in the asymptomatic control group (4.74% ± 1.96%) (P < 0.001). In our study, the cutoff value with the highest diagnostic accuracy for the VI value was found to be 6.350% in the right and left thyroid lobes VI in the longitudinal plane, 6.225% for the mean VI of thyroid gland in the longitudinal plane, 6.350% for the VI of thyroid gland in transverse plane, 6.000% for the mean VI of all thyroid glands. Our study claims that any of these cutoff VI values can be used (>72% sensitivity, specificity, PPV, NPV, and diagnostic accuracy) for predicting the presence of HT. The gray-scale US or CDI techniques may not be able to reflect changes during the early HT diagnostic stages. However, for de-tecting the VI, the above specified cutoff values using SMI and HT can be diagnosed by high sensitivity and specificity.
We found no significant difference between the mean VI of thyroid gland, VI values of the right lobe VI in the longitudi-nal plane, left lobe VI in the longitudilongitudi-nal plane, the mean VI of
The mean VI of thyroid gland 6.000 86.3 82.2 78.4 88.8 83.9 0.925
the thyroid gland in the longitudinal plane, or the VI of the thy-roid gland in the transverse plane (P > 0.05). This finding shows that instead of making multiple measurements to obtain a mean VI numerical value for all thyroid glands, the single VI value obtained from the longitudinal plane from a single lobe can be reliably used in the diagnosis of HT with the specified cutoff values, in the daily polyclinic conditions. Therefore, in pediatric cases where there is difficulty in cooperation, the VI examina-tion time may be shortened by measuring from a single lobe.
There is only 1 study where the thyroid gland was
evalu-ated using the 2DcSMIVI reported in the literature.4This study
was performed among a smaller number of participants (33
asymptomatic and 70 HT patients) when compared with our study. In that study, the margins of each thyroid lobe were man-ually outlined as the ROI on the cSMI images, upper, middle, and inferior zones of the right and left lobes at the transverse section in order to obtain the VI, and a mean value of the VI for each lobe were calculated from these 3 measurements for
each lobe.4We used the mean VI values of the thyroid gland
ob-tained by drawing the whole of the thyroid gland into the free ROI in the longest axis in the longitudinal and transverse planes in this study. Compared with the previous study in the
litera-ture,4we obtained average VI values, and this finding supports
the reliability of our VI measurement technique. We consider
TABLE 3. The Mean Parameters of the Thyroid Function of the Symptomatic Patients and Asymptomatic Group
Symptomatic Group Asymptomatic Control Group
P Mean ± SD (Range) Mean ± SD (Range)
TSH 5.40 ± 8.02 (0.64–6) 2.74 ± 1.27 (0.80–63.42) *0.001 fT4 3.71 ± 20.75 (0.84–187) 1.32 ± 0.16 (1.02–1.73) 0.541 Thyroglobulin 11.82 ± 18.80 (0.20–84.20) — TPOAbs 497.58 ± 380.41 (13.40–1000) — TgAbs 550.01 ± 834.59 (25.50–3000) — *P < 0.05.
TABLE 4. Correlation Analysis Between Laboratory Values (TSH, fT4, Thyroglobulin, TPOAbs, TgAbs) and VI Values and Correlation Analysis Between Thyroid Volume and Thyroglobulin, TPOAbs, and TgAbs
Correlation Analysis Between P Pearson Coefficient TSH and Right thyroid lobe VI in longitudinal plane 0.042* 0.202 (positive correlation)
Left thyroid lobe VI in longitudinal plane <0.001* 0.413 (positive correlation) Mean VI of thyroid gland in longitudinal plane 0.003* 0.326 (positive correlation) VI of thyroid gland in transverse plane 0.014* 0.274 (positive correlation) Mean VI of thyroid gland 0.004* 0.315 (positive correlation) fT4 and Right thyroid lobe VI in longitudinal plane 0.055 0.215
Left thyroid lobe VI in longitudinal plane 0.885 0.160 Mean VI of thyroid gland in longitudinal plane 0.291 0.120 VI of thyroid gland in transverse plane 0.118 0.176 Mean VI of thyroid gland 0.170 0.155 Thyroglobulin and Right thyroid lobe VI in longitudinal plane 0.795 −0.030 Left thyroid lobe VI in longitudinal plane 0.366 −0.102 The mean VI of thyroid gland in longitudinal plane 0.536 −0.070 VI of thyroid gland in a transverse plane 0.940 0.009 Mean VI of thyroid gland 0.776 −0.032 TPOAbs and Right thyroid lobe VI in longitudinal plane 0.046* 0.224 (positive correlation)
Left thyroid lobe VI in longitudinal plane 0.032* 0.240 (positive correlation) The mean VI of thyroid gland in longitudinal plane 0.029* 0.244 (positive correlation) VI of thyroid gland in transverse plane 0.041* 0.229 (positive correlation) Mean VI of thyroid gland 0.027* 0.248 (positive correlation) TgAbs and Right thyroid lobe VI in longitudinal plane <0.001* 0.445 (positive correlation) Left thyroid lobe VI in longitudinal plane 0.001* 0.384 (positive correlation) The mean VI of thyroid gland in longitudinal plane <0.001* 0.435 (positive correlation) VI of thyroid gland in transverse plane 0.001* 0.374 (positive correlation) Mean VI of thyroid gland <0.001* 0.424 (positive correlation) TgAbs and thyroid volume <0.001* 0.511 (positive correlation) TPOAbs 0.001* 0.359 (positive correlation)
Thyroglobulin 0.558 −0.066
this measurement method to be superior, more applicable, more repeatable, and more reliable than the transverse plane technique measurements used before. However, in our study, because of the fact that there is no statistically significant difference between the measurements planes, only the transverse plane VI can be used with our cutoff values. In the study performed by Bayramoglu
et al,4statistically significant differences were found between
the patient group VI and the asymptomatic control group, and their best cutoff value was found to be 10.58%. Although the same device was used with similar parameters in the study
per-formed by Bayramoglu et al,4the reasons for finding a higher
cutoff value than the one in our study might be because the av-erage age of the participants included in the study was higher than that in our study; additionally, patients with fibrotic thyroid glands, echogenic septa, and marked pseudonodular appearance had not been included in the former study. In the late stages of HT, where these US findings were observed, a decrease was al-ready expected in the thyroid vascularity when compared with
early-stage HT.7
High-frequency US is an important and most commonly
used imaging tool for diagnosing thyroid diseases.20 One of
the diagnostic criteria used in our study for HT included low echogenicity and heterogeneous echotexture on US. Millimetric colloidal cysts made a heterogeneous appearance on the gray-scale US examination and may be misdiagnosed as an HT
com-monly in early childhood.4The use of SMI and VI would be useful
for excluding HT in patients with heterogeneous thyroid paren-chyma due to a colloidal cyst, because with SMI, this patient group can demonstrate reliably that there is no increase in vascularity, whereas low VI can quantitatively demonstrate differences.
Serum antibody levels indirectly reflect the degree of
in-flammation in the thyroid gland of patients with HT disease.21
The presence of TgAbs may reflect an early immune response, whereas TPOAbs may be characteristic of a late adaptive
im-mune response.22,23,24Anti–thyroid peroxidase antibodies are
associated with over thyroid dysfunction, and their presence tends to correlate with the degree of lymphocytic infiltration
and thyroid damage.22,23Bayramoglu et al4found a positive
significant correlation between the VI and TgAbs and TPOAbs levels. In our study, we also found significant positive correla-tion between the VI obtained via cSMI and TgAbs and TPOAbs levels (P < 0.05); however, no significant correlation between VI values and thyroglobulin and fT4 levels were found (P > 0.05). Positive significant correlation between autoanti-body levels and the VI values indicates that the VI value of the thyroid gland obtained may be related to the degree of
we have demonstrated a positive correlation between the thy-roid blood flow and TSH using objective numerical data that uses 2DcSMIVI.
Increased lymphoid tissue and fibrosis cause the volume
of the thyroid gland to increase in HT.28In this study, we found
that the isthmus thickness and thyroid right lobe, left lobe, and the total thyroid volume were significantly higher in patients with HT when compared with the asymptomatic group (P < 0.001).
Kandemirli et al1found a significant positive correlation between
total thyroid volume and TPOAbs levels. We found a significant positive correlation between the thyroid volume and TgAbs and
TPOAbs levels (P≤ 0.001). Serum autoantibody levels indirectly
reflected the degree of inflammation and immune response and correlated with the degree of thyroid damage and lymphocytic
in-filtration in the thyroid glands of patients with HT disease.21,23,24
Therefore, in patients with HT, an increased thyroid volume and blood flow (increased VI) were the expected outcomes of the in-creased serum autoantibody levels, and the results of our study support this finding.
Hypoechoic echogenicity that reflects a longer duration of inflammation and echogenicity of the thyroid gland during the
initial stages of HT may be normal.20The degree of lymphocyte
infiltration, lymph follicle formation, fibrosis, and the duration of inflammation in thyroid tissues could cause different degrees
of hypoechogenicity.20,22,28,29The echogenicity of the thyroid
gland in the HT group was mainly reduced. Studies in which the thyroid gland vascularity was evaluated by grading using CDI techniques found out that thyroid vascularity increased as
the thyroid parenchymal echo decreased.11,14In our study, as
the thyroid parenchyma echo decreased, the mean values of the VI increased. We found a significant negative correlation between the parenchymal echogenicity and the VI in patients with HT (P = 0.011). And when the patient and control groups were evaluated together, we found a stronger significant nega-tive correlation between thyroid parenchymal echogenicity and the VI values (P < 0.001). Our results are consistent with lit-erature, but we found similar results with the objective numeri-cal data that use VI differently from previous studies. B-mode scanning and VI could be used to evaluate the degree of disease activity in HT. However, because VI can quantitatively show
Hypoechogenic 0 36 (45.0%) 36 11.9806 ± 4.6909 Marked
hypoechogenic
0 14 (17.5%) 14 16.8429 ± 7.3336
TABLE 6. Correlation Analysis Between Thyroid Gland Echogenicity and VI Values Correlation Between Echogenicity and VI in P Pearson Correlation Coefficient Patients with HT 0.011* −0.282 Asymptomatic group 0.916 0.010 All participants <0.001* −0.506 *P < 0.05.
this condition with numerical data, we think that it is a superior examination method to the B-mode US.
Our study has several limitations. The diagnosis of HT re-lied on the clinical manifestation, US, and laboratory findings instead of pathological findings. We were not able to analyze the possible histopathologic changes in the thyroid gland of HT patients or the control group by biopsy. Most of the patients included in the study were undergoing follow-up with medical treatment. Patients with newly diagnosed HT were also included in the study, but these patients were few in number. If future studies with larger sample sizes evaluate and compare the VI numerical values initiation prior to the medical treatment with a follow-up and pathology findings, the efficacy of the VI in showing the degree of inflammation may be presented in a more reliable way. Autoantibody was not examined in an asymptom-atic group, but in a asymptomasymptom-atic group, the thyroid US was completely normal and the serum hormone levels (fT4, TSH) were within normal limits. Based on the gray-scale US findings, the HT was subdivided into 3 categories as diffuse thyroiditis,
focal thyroiditis, and fibrotic thyroid gland.1In our study, we
did not subdivide the patients according to the US findings. We included only children in this study. Because of the physio-logic changes seen during growth, children may have different VI values. These possible changes in the normal thyroid gland may create difficulties in the use of VI in the thyroid paranchyma pathologies of children. One operator did all the examinations, so we could not evaluate interobserver variability.
In conclusion, patients with HT had higher thyroid VI values when compared with the healthy asymptomatic group. The best cutoff value for VI value was defined as 6.00% for predicting the presence of HT. We found significant correlation between the VI values and autoantibody levels, TSH levels, and thyroid parenchyma echogenicity. Therefore, using these find-ings, we suggest that the 2DcSMIVI technique can effectively be used as an imaging method in the diagnosis of HT with high sensitivity and specificity and predict inflammation and assess severity of inflammation in the follow-up check-ups of HT pa-tients with objective quantitative values. However, further stud-ies with larger numbers of patients are needed to confirm the use of the 2DcSMIVI technique in routine HT evaluation in daily polyclinic conditions.
REFERENCES
1. Kandemirli SG, Bayramoglu Z, Çalişkan E, et al. Quantitative assessment of thyroid gland elasticity with shear-wave elastography in pediatric patients with Hashimoto's thyroiditis. J Med Ultrason. 2018;45:417–423. 2. Ruchala M, Szmyt K, Slawek S, et al. Ultrasound sonoelastography in the
evaluation of thyroiditis and autoimmune thyroid disease. Endokrynol Pol. 2014;65:520–526.
3. McGrogan A, Seaman HE, Wright JW, et al. The incidence of autoimmune thyroid disease: a systematic review of the literature. Clin Endocrinol (Oxf ). 2008;69:687–696.
4. Bayramoglu Z, Kandemirli SG, Caliskan E, et al. Assessment of paediatric Hashimoto's thyroiditis using superb microvascular imaging. Clin Radiol. 2018;73:1059.e9–1059.e15.
5. Willms A, Bieler D, Wieler H, et al. Correlation between sonography and antibody activity in patients with Hashimoto thyroiditis. J Ultrasound Med. 2013;32:1979–1986.
6. Sinclair D. Clinical and laboratory aspects of thyroid autoantibodies. Ann Clin Biochem. 2006;43:173–183.
7. Yamashiro I, Saito OC, Chammas MC, et al. Ultrasound findings in thyroiditis. Radiol Bras. 2007;40:75–79.
8. Raber W, Gessl A, Nowotny P, et al. Thyroid ultrasound versus antithyroid peroxidase antibody determination: a cohort study of four hundred fifty-one subjects. Thyroid. 2002;12:725–731.
9. Bogazzi F, Bartalena L, Brogioni S, et al. Thyroid vascularity and blood flow are not dependent on serum thyroid hormone levels: studies in vivo by color flow doppler sonography. Eur J Endocrinol. 1999;140:452–456. 10. Ralls PW, Mayekawa DS, Lee KP, et al. Color-flow Doppler sonography in
Graves disease:“thyroid inferno.”. AJR Am J Roentgenol. 1988;150: 781–784.
11. Hari Kumar KV, Pasupuleti V, Jayaraman M, et al. Role of thyroid Doppler in differential diagnosis of thyrotoxicosis. Endocr Pract. 2009;15(1):6–9. 12. Hata J. Toshiba Medical Systems Corporation; 2014. Seeing the unseen new
techniques in vascular imaging; superb micro-vascular imaging. Available at: http://www.toshibamedicalsystems.com/library/ us/index.html. Accessed November 29, 2017.
13. Durmaz MS, Sivri M. Comparison of superb micro-vascular imaging (SMI) and conventional Doppler imaging techniques for evaluating testicular blood flow. J Med Ultrason (2001). 2018;45:443–452.
14. Vitti P, Rago T, Mazzeo S, et al. Thyroid blood flow evaluation by color-flow doppler sonography distinguishes Graves' disease from Hashimoto's thyroiditis. J Endocrinol. 1995;18:857–861.
15. Machado P, Forsberg F. Toshiba Medical Systems Corporation; 2014. Initial experience with a novel microvascular flow imaging technique. Medical review. Available at: http://www.toshibamedicalsystems. com/library/us/index.html. Accessed November 29, 2017.
16. Durmaz MS, Özbakır B, Cebeci H, et al. Superb microvascular imaging in the visualization of recanalization in deep vein thrombosis. Acta Phlebol. 2018;19:20–27.
17. Stępniak A, Czuczwar P. 3D power Doppler vascular indices as a novel technique in assessing the outcome of minimally invasive techniques in uterine fibroids treatment. Prz Menopauzalny. 2017;16:118–121. 18. Höfling DB, Cerri GG, Juliano AG, et al. Value of thyroid echogenicity in
the diagnosis of chronic autoimmune thyroiditis. Radiol Bras. 2008;41: 409–417.
19. Arslan S, Karahan AY, Oncu F, et al. Diagnostic performance of superb microvascular imaging and other sonographic modalities in the assessment of lateral epicondylosis. J Ultrasound Med. 2018 Mar;37(3):585–593. 20. Wang Y, Ruan Z, Zhang J, et al. Hashimoto's thyroiditis shows
characteristics of increased thyroid volume, decreased echo and increased blood flow signals using ultrasound gray-scale intensity technology. Biomed Res. 2017;28:6365–6370.
21. Jianxue L, Yaoren Z, Yonghao J, et al. The value of shear wave elastography in diffuse thyroid disease. Clin Imaging. 2018;49:187–192.
22. Fukuhara T, Matsuda E, Izawa S, et al. Utility of shear wave elastography for diagnosing chronic autoimmune thyroiditis. J Thyroid Res. 2015;2015: 164548.
23. Pearce EN, Farwell AP, Braverman LE. Thyroiditis. N Engl J Med. 2003; 348:2646–2655.
24. Schiemann U, Avenhaus W, Konturek JW, et al. Relationship of clinical features and laboratory parameters to thyroid echogenicity measured by standardized grey scale ultrasonography in patients with Hashimoto's thyroiditis. Med Sci Monit. 2003;9:MT13–MT17.
25. Anderson L, Middleton WD, Teefey SA, et al. Hashimoto thyroiditis: part 1, sonographic analysis of the nodular form of Hashimoto thyroiditis. AJR Am J Roentgenol. 2010;195:208–215.
26. Sarikaya B, Demirbilek H, Akata D, et al. The role of the resistive index in Hashimoto's thyroiditis: a sonographic pilot study in children. Clinics (Sao Paulo). 2012;67:1253–1257.
27. Onoda N, Kato Y, Seki T, et al. Increased thyroid blood flow in the hypoechoic lesions in patients with recurrent, painful Hashimoto's thyroiditis at the time of acute exacerbation. Endocr J. 2009;56:65–72. 28. Li Y, Zhou G, Ozaki T, et al. Distinct histopathological features of
Hashimoto's thyroiditis with respect to IgG4-related disease. Mod Pathol. 2012;25:1086–1097.
29. Caturegli P, de Remigis A, Rose NR. Hashimoto thyroiditis: clinical and diagnostic criteria. Autoimmun Rev. 2014;13:391–397.
