First reports on thyroid ultrasound imaging in pediatric populations appeared in the 1970s, almost ten years after the first report in adults(1,2). Since then, ultrasonography (US) has frequently proved to be a useful diagnostic method in patients with thyroid diseases. Apart from diagnosing thyroid nodules, US can visualize the echostructure and vascularity pattern of the thyroid tissue, particularly helpful in diagnosing autoimmune thyroid diseases(3,4).
Hashimoto’s lymphocytic thyroiditis (HLT) is an autoimmune chronic inflammatory disease of the thyroid gland. It is one of the most common thyropathies and the cause of acquired hypothyroidism in children(5). The disease can be associated with childhood diabetes, Turner syndrome, Down syndrome, Noonan syndrome or may occur after the treatment of Hodgkin disease(5).
In adult patients, typical sonographic features of HLT include enlargement of the thyroid, fibrotic septa and micronodulation with a reduction in echogenicity of the thyroid parenchyma(6). Similar sonographic findings are observed in pediatric populations. Serres-Créixams et al. (2008) noticed a statistically significant increase in the incidence of enlarged lymph nodes in the paratracheal region in adult patients with Hashimoto’s thyroiditis(7). They suggested that these lymph nodes represent an additional sonographic sign of Hashimoto’s thyroiditis.
In our practice, the presence of lymph nodes adjacent to the lower part of the thyroid gland has been frequently observed in sonographic imaging in children with HLT.
Our study was conducted to assess the ultrasound sensitivity and specificity for lymph nodes adjacent to the lower part of the thyroid gland, as an additional sonographic finding in HLT in children.
A total of 98 children (14 boys (14%) and 84 girls (86%), with a mean age of 12.7 years, range 7–17 years), selected from the register of the endocrinology outpatient department at the Medical University of Gdańsk, were prospectively examined. All subjects met the diagnostic criteria for HLT. The diagnosis of HLT was in all cases based on: clinical data and laboratory findingss, as well as, in 8 cases, on fine-needle aspiration biopsy results.
The following laboratory parameters were measured with chemiluminescent microparticle immunoassay CMIA (Abbott Laboratories, Abbott Park, IL): free triiodothyronine (fT3; normal range 2.63–5.7 pmol/L), free thyroxine (fT4; normal range 9.01–19.05 pmol/L), serum thyrotropin (TSH; normal range 0.35–4.94 mU/L). IgG class serum thyroglobulin antibodies (anti-TGAb; poorly positive >150–250 IU/mL, strongly positive >250 IU/mL) and thyroid peroxidase antibodies (anti-TPO; positive >100 U/mL) and/or anti-TSH receptor antibody (positive >0.4 U/mL) were assayed using the enzyme immunoassay microplate test (BioSystems S.A., Barcelona, Spain) according to the manufacturer’s instructions.
For the purpose of this study, two control groups were enrolled and evaluated prospectively. Control group A included 102 healthy volunteers (18 boys (18%) / 84 girls (82%) – mean age 12.8 years, range 7–16 years) and a control group B included 94 children with cervical lymphadenopathy (20 boys (21%) and 74 girls (79%) – mean age 12.5 years, range 6–16 years) from the registry of the oncology outpatient department. There was no difference in age and sex between the control groups (p > 0.05).
Sonographic examination of the thyroid was performed by two operators (radiologist and pediatrician) with a Logiq 7 scanner with a convex (2.0–5.0 MHz) and a linear transducer (4.5–13.0 MHz) (GE Medical Systems, Waukesha, WI, USA). All children were examined using the same technique and equipment.
Informed written consent was obtained from all volunteers and the study was approved by the ethical committee on human experimentation of our institution.
Statistical analysis was performed with STATA 12.1 software (StataCorp, College Station, Texas, USA). The Student’s t-test and chi-square test or Fisher’s exact test were used. A p value of <0.05 was considered statistically significant.
Children were examined in the supine position with a hyperextended neck. A grey-scale ultrasound was performed in each patient to evaluate thyroid echogenicity, echostructure and volume. The volume of each lobe was calculated separately, using a formula for elliptical shape volume [(Π/6) × length × width × depth]. The thyroid volume was a sum of the volumes of both lobes. Thyroid echogenicity was compared with the echogenicity of the sternocleidomastoid muscle. Power Doppler (PD) and color Doppler (CD) sonography settings were optimized for detecting low blood flow in the lymph nodes (high sensitivity, low wall filter, pulse repetition frequency of 1700 Hz).
Special attention was paid to the presence and size of the thyroid gland adjacent to region VI lymph nodes, according to the American Joint Committee on Cancer (AJCC) classification of the cervical lymph nodes(8). They were localized below the lower part of the thyroid gland. They were seen in conjunction with the thyroid tissue on single longitudinal scans (Fig. 1). All lymph nodes in that region which were not seen on the same scan with the thyroid tissue were excluded from our study. All detected lymph nodes were assessed on the basis of the criteria presented by Ying and Solbiati(9,10).
Five different sized lymph nodes (white arrows) localized below the lower part of the thyroid gland (double short white arrow) on the same scan with the thyroid tissue
The size of lymph nodes was graded as small (<5 mm) (Fig. 2) medium (5–10 mm) (Fig. 3) and large (>10 mm) (Fig. 4) based on the longitudinal diameter of the lymph node.
Small lymph nodes <5 mm (white arrows) adjacent to the lower part of the thyroid gland (double short white arrow)
Medium lymph nodes 5–10 mm (white arrows) adjacent to the lower part of the thyroid gland (double short white arrow) with a visible focal lesion (black arrow)
Large lymph node >10 mm (white arrows) adjacent to the lower part of the thyroid gland (double short white arrow)
Thyroid ultrasound revealed typical sonographic signs of autoimmune thyroid disease in all children with Hashimoto’s thyroiditis (HT) and none in the control groups.
Thyroid echogenicity was decreased in all (100%) children with HLT, 8 (7%) children from control group A, and none (0%) of the children in control group B. Fibrotic septa and micronodulation were observed in the HLT group only.
The thyroid volume was statistically higher in the HLT group 18.35 ml ± 12 versus 9.2 ml ± 4.1 in the control group A (p < 0.001) and 8.8 ml ± 3.8 in the control group B (p < 0.001). There was no difference in thyroid volume between control groups (p > 0.05)
At least 2 lymph nodes adjacent to the lower part of the thyroid gland lobes, localized on both sides of the thyroid, were seen in 96 children (98%) from the HLT group (Tab. 1). A total number of 326 lymph nodes were found. Fifty-two out of 326 lymph nodes were less than 5 mm, 186 of 326 lymph nodes were between 5 and 10 mm, and 88 were more than 10 mm in the long axis. The ratio of “short axis diameter to long axis diameter” was less than 0. 6 in 62 lymph nodes, with a maximum size of 6 mm and less (small round-shaped lymph nodes). All lymph nodes had sharp borders and typical, low echogenicity with no signs of cystic necrosis or calcifications. The hilus was visible only in 16 lymph nodes with longitudinal diameter larger than 10 mm. A normal hilar intranodal vascularity in PD, without an increased flow was in 14 lymph nodes from that group. Most of the lymph nodes (312) had no detectable vascularity in CD and PD.
Ultrasound features of lymph nodes in Hashimoto’s lymphocytic thyroiditis group
|Features of lymph nodes||Lymph nodes in HLT group|
|Size||<5.0 mm||52 (16%)|
|5–10 mm||186 (57%)|
|>10 mm||88 (27%)|
|Shape (Solbiati index)||<2.0||62 (19%)|
|not sharp||0 (0%)|
|Cystic necrosis||present||0 (0%)|
|Intranodal distribution of vascularity||absent||312 (96%)|
Ultrasound examination did not show any noticeable lymph nodes adjacent to the lower part of the thyroid lobes in 2 children (2%) from the HLT group.
None of the children in the control groups had lymph nodes in this location.
The calculated value of sensitivity for the diagnosis of HLT for adjacent lymph nodes visualization was 98% with 95% CI varying from 92.8% to 99.8% with respect to total control group. The value of specificity for the HLT diagnosis was equal to 100% with 95% CI varying from 96.2% to 100% for patients with lymphadenopathy (sample prevalence 51%) and 95% CI varying from 98.1% to 100% estimated for total pediatric population (sample prevalence 33%).
In 39 (39%) children with HLT, para-isthmian lymph nodes were noted as well (Fig. 5). Para-isthmian lymph nodes were not seen in both control groups. The sensitivity of this parameter was 40% and the specificity was 100%.
Three para-isthmian lymph nodes (white arrows) and thyroid isthmus (double short white arrow)
Although fine needle aspiration cytology (FNAC) is regarded a golden standard in HLT diagnostics, it is not used in routine pediatric clinical practice, mainly due to its invasive nature. In children, HLT diagnostics is generally based on clinical data and elevated anti-TPO and/or anti-TGAb levels. The sensitivity of antibody markers is about 90% and they can be negative in some histologically proven cases of thyroiditis(11–13).
Ultrasonography (US) is the first-line imaging method in thyroid diseases. This method is noninvasive, inexpensive and suitable for all age groups. Additionally, it provides data on thyroid volume and parenchymal changes. Typical sonographic features of HLT include: increased size, fibrotic septa, micronodulation with reduced thyroid echogenicity and marked intraparenchymal hypervascularity in power Doppler modality. There are similar, well-known US findings in children and adults(14,15–17). The most frequently described ultrasound markers of autoimmune thyroiditis, both in children and adults, are diffuse hypoechogenicity and enlargement of the thyroid gland(6,13,17,18). In the late stage, atrophic glandular parenchyma is observed. However, the different course of HLT in adults and in children and the dissimilarities in ultrasound imagining must be borne in mind. In adults, a positive predictive value of diffuse hypoechogenicity ranges widely from 18.5% to 96%(18,19). In children populations, this range is from 37.5 % to 41.4 %(15,20).
Differential diagnosis should mainly consider diffuse and limited colloid goiter, and subacute (granulomatous) thyroiditis. The colloid goiter can mimic focal or diffuse HLT, especially when small hyperechoic colloid deposits are not present in ultrasound examinations. Although technologically advanced ultrasound systems are currently in use, many investigators have problems with correct ultrasound image description in the case of HLT. Moreover, hypoechogenicity as a single feature of HLT seems to be insufficient.
‘Lymph nodes adjacent to the thyroid gland’ could represent an additional, reliable ultrasound sign of HLT in children. According to the AJCC classification, all of these lymph nodes should belong to the group of anterior cervical lymph nodes (Level VI)(8). However, these are pretracheal, prelaryngeal and paratracheal lymph nodes that are usually noticed in that location(9). To our knowledge, there are no studies describing lymph nodes adjacent to the thyroid capsule.
We showed a strong correlation between the presence of lymph nodes adjacent to the thyroid gland, and HLT. We did not find this surprising because, apart from the commonly used criteria for ultrasound imagining in children with HLT, we used a 2002 feature ‘lymph nodes adjacent to the thyroid gland’. We were astonished to find that these lymph nodes were not present in control group A and B.
We believe that lymphadenopathy and the presence of lymph nodes adjacent to the thyroid are connected with the nature of HLT. Histopathological changes in HLT include diffuse lymphoid and plasma infiltration of thyroid parenchyma, with occasional germinal centers, varying degrees of follicular atrophy and fibrosis(21). Diffuse lymphoid infiltration of thyroid parenchyma is related to the activation of the lymphatic system and response of lymph nodes adjacent to the thyroid gland.
There was no detectable vascularity in CD and PD examinations in nearly all of the lymph nodes. Visible vascularity is a good indicator of lymph nodes and may be connected with a phase of HLT – this requires further investigation.
Preliminary data published by Solivetti at al. 1998 indicate that enlargement of para-isthmian lymph nodes appears to be one of the signs of subacute thyroiditis(22). However, other lymph node groups were not examined, which is the main bias of that study. In 2008, a report was published on the presence of paratracheal lymph nodes in adult autoimmune thyroiditis. The authors estimated the sensitivity of that sign in diagnosing autoimmune thyroiditis at 93.4%, and its specificity at 74.5%(7). Our results suggest higher sensitivity (98%) and specificity (100%) of this parameter for pediatric population.
Although our results are very encouraging, there are also some limitations of the study:
The stage of HLT was not taken into consideration (no available data in children population).
Children with other autoimmune thyroid diseases were not included in the differential diagnosis.
A relatively small number of children had biopsy-confirmed HLT.
Lymph nodes adjacent to the lower part of the thyroid lobes are an additional ultrasound sign of pediatric Hashimoto’s lymphocytic thyroiditis with 98% sensitivity and 100% specificity.