Abstract

Objective

International guidelines recommend different approaches for the management of pediatric thyroid nodules with a finding of atypia of undetermined significance (AUS) on cytology. The American Thyroid Association (ATA) pediatric guidelines recommend surgery whereas the European Thyroid Association (ETA) guidelines recommend repeat fine-needle aspiration biopsy after six months. Our objective was to identify markers of malignancy in AUS cases and to discuss the management of pediatric AUS nodules.

Methods

Specimens from pediatric patients who underwent surgery due to AUS cytology were re-evaluated and subcategorized according to the 2023 Bethesda classification.

Results

Of the 20 cases included, 11 (55%) were histologically benign, while 9 (45%) were malignant. On the subcategorization of AUS, nuclear atypia was present in 14 patients (70%) and other atypia in 6 patients (30%). Of the cases with nuclear atypia, 64.3% were malignant (n=9), whereas no malignancy was detected in cases with other atypia (p=0.012). Among the cytopathological features, chromatin clearing, nuclear enlargement, and irregular margins were significantly associated with malignancy (p=0.035, p=0.003, and p=0.012, respectively). Adhering to ETA recommendations would delay diagnosis by at least 6 months in 45% of our malignant cases. Conversely, performing lobectomy according to ATA recommendations may lead to unnecessary surgery in 55% of our cases.

Conclusion

Based on our findings, lobectomy appears to be a more appropriate approach in AUS cases but only when nuclear atypia is present, to avoid diagnostic delay and unnecessary surgery. Guidelines should be updated according to the latest Bethesda classification.

Introduction

Thyroid nodules in children and adolescents are less common than in adults (0.2-5% vs. 19-35%), although the likelihood of malignancy is higher in children and adolescents than adults (up to 20-26% vs. 7-15%) (1, 2, 3, 4, 5, 6). Key ultrasonographic features of thyroid nodules—such as hypoechogenicity, solid composition, microcalcifications, irregular margins, increased intranodular vascularity, and a taller-than-wide shape—are significant indicators of thyroid malignancy (7, 8, 9). The American Thyroid Association (ATA) adult guidelines recommend ultrasound (US)-guided fine-needle aspiration biopsy (FNAB) for solid or partially cystic thyroid nodules ≥1 cm or nodules with suspicious ultrasonographic findings in a patient without risk factors for thyroid malignancy, which is similar to the European Thyroid Association (ETA) guidelines that recommend FNAB for suspicious nodules based on multiple US (5, 6). FNAB results are reported according to the Bethesda System for Reporting Thyroid Cytopathology (TBSRTC). The TBSRTC is a category-based reporting system, comprising six categories, that has provided a uniform reporting scheme for thyroid FNAB since the publication of the first edition in 2009 (10). TBSRTC was updated in 2023. The latest edition simplifies the terminology of the six categories and recommends dropping the terms “unsatisfactory,” “follicular lesion of undetermined significance,” and “suspicious for a follicular neoplasm” for TBSRTC categories 1, 3, and 4, respectively. The new categories are: (1) non-diagnostic; (2) benign; (3) atypia of undetermined significance (AUS); (4) follicular neoplasm; (5) suspicious for malignancy; and (6) malignant (11). Furthermore, in the 2023 TBSRTC, the AUS subcategories are newly divided into two groups—nuclear and other (including architectural, oncocytic, and lymphocytic atypia)—whereas in the 2017 TBSRTC they were divided into five subcategories: cytologic, architectural, cytologic and architectural atypia, Hürthle cell AUS, and atypia not otherwise specified (11).

AUS is the most indeterminate category, as it is neither benign nor clearly malignant and can progress to malignancy or to a “suspicious for malignancy” category. The rate of malignancy (ROM) of AUS nodules in adults is approximately 22% (range, 13–30%), whereas it is higher—28% (range, 11–54%)—in the pediatric population (11). The management of adult thyroid nodules with AUS cytology includes some combination of repeat FNAB, molecular testing, and surgery, depending on risk factors, patient history, clinical features, sonographic patterns, and patient preference. Guidelines provide different recommendations for the management of thyroid nodules with AUS cytology in children. In the management of the pediatric thyroid nodules with AUS cytology, the ATA pediatric guidelines recommend surgery (mostly lobectomy and isthmusectomy) whereas the ETA guidelines recommend repeat FNAB after six months (5, 6). This discrepancy in the guidelines is attributed to the lack of data and limited research on this subject.

The aim of this study was therefore to determine the ROM and markers of malignancy in pediatric AUS cases and to discuss the approach to managing pediatric thyroid nodules with AUS cytology.

Methods

Records of children and adolescents who were followed up for thyroid nodules with AUS cytology in the Pediatric Endocrinology Division of the University of Health Sciences Türkiye, Ankara Dr. Sami Ulus Child Health and Diseases Training and Research Hospital, were retrospectively evaluated. Since the first edition of TBSRTC was published in 2009, patients were screened from 2009 to 2023. Age at diagnosis, gender, thyroid stimulating hormone (TSH) levels at diagnosis (in μIU/mL), patient history, thyroid US findings, FNAB cytopathology, and histopathological results were recorded.

Experienced radiologists performed US and FNAB procedures on the patients using high-frequency linear-array transducers (Toshiba Aplio 500, Toshiba Medical Systems, Tokyo, Japan). The US results were evaluated for composition (solid, semisolid, cystic), echogenicity (isoechoic, hypoechoic), margin characteristics (regular, irregular), and the presence of microcalcifications. Sampling was performed with 22-gauge needles attached to 10-mL syringes. Five or six samples were prepared on slides for each patient. For patients with multiple nodules, the most suspicious nodule was selected. Surgery decisions for children with AUS cytology were made by an experienced multidisciplinary thyroid team, including experts in pediatric endocrinology, radiology, pathology, oncology, and surgery, following ATA guidelines.

Within the scope of this study, FNAB samples were re-evaluated by the same experienced pathologist and subcategorized into two groups by type of atypia, nuclear and other, as described in the new TBSRTC criteria. Nuclear atypia was defined as focal nuclear changes such as chromatin clearing, nuclear enlargement, and irregular nuclear contour.

In the literature, the terms “malignancy rate” and “risk of malignancy” are used interchangeably to describe the percentage of malignant cases in each TBSRTC category. In most studies, the number of malignant cases is divided by the number of cases with a histopathological diagnosis. However, due to variations in surgical selection rates, determining the actual ROM is challenging. In the present study, two different methods were used to estimate the most accurate ROM. The first method (ROM-H: ROM of nodules with AUS cytology based on histopathological diagnosis) was calculated by dividing the number of malignant cases by the number of AUS cases that underwent surgery and had a histopathological diagnosis. The second method (ROM-O: Overall ROM) was calculated by dividing the number of malignant cases by the overall number of AUS cases.

Approval was obtained from the Ankara Atatürk Sanatorium Training and Research Hospital Clinical Research Ethics Committee (decision no: 2012-KAEK-15/2767, date: 23.08.2023).

Statistical Analysis

Statistical Package for the Social Sciences, version 22 (IBM Inc., Chicago, IL, USA) software was used for statistical analysis. In descriptive statistics, qualitative variables are expressed as frequency (n), and percentage (%); quantitative variables as mean±standard deviation for normally distributed data and as median (minimum-maximum) for non-normally distributed data. Normality was assessed using the Shapiro-Wilk test. The χ² test was used in the analysis of categorical variables. The Student’s t-test was used for comparisons of normally distributed continuous variables, and Mann-Whitney U test for non-normally distributed variables. A p<0.05 was considered statistically significant.

Results

Among 312 children and adolescents who underwent FNAB for thyroid nodule(s) in our outpatient clinic, 28 (9%) had AUS cytology. These patients were recommended to undergo lobectomy in accordance with ATA guidelines. Five patients (17.8%) refused surgery, and three patients (10.7%) were transferred to adult clinics upon turning 18 years old and were subsequently followed according to adult guidelines. Twenty patients who underwent surgical evaluation were included.

The mean age of cases was 14.3±3 years (range: 7.7-18 years); the female/male ratio was 4:1. Backgrounds, complaints on admission, and final diagnoses of the patients are presented in Table 1. Six patients (30%) had Hashimoto thyroiditis, two (10%) had congenital hypothyroidism, and one patient each (5%) had a thyroglossal duct cyst, Turner syndrome, insulin-dependent diabetes mellitus, precocious puberty, or epilepsy. On admission, 45% of the patients (n=9) had no symptoms. Among the remaining patients, 25% (n=5) had swelling in the neck, 25% (n=5) had abnormal thyroid function tests, and 5% (n=1) had breathing difficulty. Three patients (15%) had a DICER1 mutation (Table 1).

In terms of thyroid US findings, 25% of the patients (n=5) had parenchymal heterogeneity. Sixty-five percent of the nodules (n=13) were solid, 10% (n=2) were semisolid, and 25% (n=5) were cystic. Thirty-five percent of the nodules (n=7) were hypoechoic, and 65% of the nodules (n=13) were isoechoic. Microcalcifications were observed in 35% of the nodules (n=7), irregular margins in 15% (n=3), and increased intranodular vascularity in 15% (n=3).

When the FNAB slides were re-evaluated by the same pathologist, nuclear atypia findings revealed chromatin clearing in 50% (n=10), nuclear enlargement in 50% (n=10), irregular nuclear contours in 70% (n=14), and histiocytoid cells in 5% (n=1). Among other atypia findings, three-dimensional groups were present in 10% (n=2) and microfollicles in 30% (n=6). Using the newly recommended subcategorization of AUS, nuclear atypia was found in 14 (70%), and other atypia in 6 (30%).

Histological diagnosis included neoplasia in 18 patients (90%). Of these, 11 (55%) were benign [follicular adenoma 9.1% (n=1), nodular hyperplasia in 72.7% (n=8), and chronic lymphocytic thyroiditis in 18.2% (n=2)], while nine were malignant (45%) [papillary thyroid carcinoma (PTC) in 44.5% (n=4), follicular variant-PTC in 22.2% (n=2), and papillary microcarcinoma in 33.3% (n=3)]. The ROM-H of nodules with AUS cytology was 45% (9/20), and when all aspirated AUS nodules were included, the ROM-O was 32.1% (9/28). In the benign and malignant groups, no significant differences were found in terms of age, gender, pubertal status, genetic predisposition, or serum TSH levels (Table 2). There were no significant difference in the US findings between malignant and benign nodules (Table 3) with the exception of the benign nodules being significantly larger in diameter (p=0.036) (Table 3). The mean diameter of cystic nodules was 25.8±22.2 mm, while the non-cystic nodules had a mean diameter of 13.4±5.5 mm (p=0.003). Eighty percent (n=4) of the cystic nodules were benign, while 20% (n=1) were malignant. The ROM-H was higher in patients with nuclear atypia features, such as chromatin clearing, nuclear enlargement, and irregular nuclear contours (p=0.035, p=0.003, p=0.012, respectively) (Table 4). All nodules with microfollicles were benign (p=0.012) (Table 4). Of the cases with nuclear atypia, 64.3% were malignant (9/14) whereas no malignancy was detected in the “other atypia” group (p=0.012) (Table 5).

In the present study, patients with DICER1 mutations were examined separately due to their predisposition to malignancy. We found that only one of the patients with a DICER1 mutation had malignant histology (PTC), whereas the other two had nodular hyperplasia. So, the ROM-H of the patients with DICER1 mutation was found to be 33.3% whereas it was 47% for those without the mutation (p=0.57) (Table 2).

Discussion

In the literature, there are varying data on the ROM of pediatric AUS cases. The ROM-H was 45% while ROM-O was 32.1% in our cohort. The true ROM lies between the ROM-H and the ROM-O. Considering resection rates ranging from 52% to 100%, the ROM-H of thyroid nodules with AUS cytology in the literature has been estimated to be between 20% and 75% (12, 13, 14, 15, 16, 17). In our clinic, three of the 28 AUS cases were transferred to adult clinics due to being over 18 years old, and five patients refused surgery. It was noted that the three patients transferred to adult clinics were not operated on but were followed up clinically according to adult guidelines. Therefore, since our surgery rate (80%) is relatively high compared with the literature, we believe that the ROM-H found in our study reflects a true ROM.

DICER1 syndrome is an autosomal-dominant, pleiotropic tumor-predisposition disorder, increasing the risk of non-toxic multinodular goiter, adenoma, and thyroid cancer, caused by pathogenic variants in DICER1 (18, 19, 20). In our cohort, only one of the patients with DICER1 mutation had differentiated thyroid cancer while the other two had nodular hyperplasia. In our study, the ROM-H of the patients with DICER1 mutation was 33.3% compared to 47% for those without mutation. Contrary to the literature, DICER1 mutation did not increase the ROM in our cohort but the low number of patients with DICER1 mutations, which is one of the limitations of the study, may have led to this unusual finding.

In earlier studies, US findings, such as solid structure, hypoechogenicity, presence of microcalcifications, irregular margins, increased intranodular vascularity, and abnormal cervical lymph nodes, have been found to be associated with malignancy (21, 22). Lee et al. (23) also found that microcalcifications and irregular margins on US were associated with malignancy in AUS cases. However, no significant relationship was found between malignancy and US findings in our study, again likely due to the small sample size. In our cohort, however, benign nodules were found to be significantly larger. Most of the cystic nodules were benign and larger, probably due to the cystic component.

The latest edition of TBSRTC divides AUS into two subcategories: AUS-nuclear and AUS-other. In our cohort, the ROM-H of cases with nuclear atypia was found to be 64.3% while ROM-H of the cases with other atypia was found to be 0%. In the literature, both in adult and pediatric studies, the ROM in nuclear atypia has been reported to be 2-4 times higher than in other atypia (24, 25, 26, 27, 28, 29). Similar to our study, Smith et al. (24) in a study including 43 pediatric AUS cases, reported that the ROM-H of nuclear atypia (formerly known as cytological atypia) was 50%. Another recent pediatric study by Jin et al. (13) found ROM-H to be 52%. According to the study of Cherella et al. (16) that included 68 pediatric thyroid nodules with AUS cytology, ROM-O in the presence of nuclear atypia increased to 59%, which was 10 times higher than in the absence of nuclear atypia.

Guidelines recommend different management strategies for thyroid nodules with AUS cytology in children. The ATA pediatric guidelines recommend surgery while the ETA guidelines recommend repeating FNAB after six months (5, 6). When lobectomy is performed based on ATA recommendations for AUS cases, the likelihood of missed diagnosis is minimized, yet some patients undergo unnecessary surgery. Following ETA recommendations with repeat FNAB after six months reduces the risk of unnecessary surgery but may delay diagnosis in malignant cases. Our data suggest that adhering to ETA recommendations would delay diagnosis of malignancy by at least six months in 45% of our cases, which is a considerable risk. However, when we performed lobectomy according to the ATA recommendations for our cases, 55% of children underwent unnecessary surgery. Due to the significantly higher ROM-H in cases with nuclear atypia in our study, performing lobectomy for these cases would reduce the risk of unnecessary surgery from 55% to 35.7%, while the risk of misdiagnosis remains at 0%. Based on clinical and cytological evaluations, we suggest that surgical decisions should be individualized for each patient, even in the absence of nuclear atypia. AUS cases without nuclear atypia may carry a lower risk, but surgical decisions should also take other clinical factors into consideration. Based on our findings, performing surgery only in cases with nuclear atypia and carrying out repeat FNAB after six months for cases with other atypia would be more appropriate than the current recommendations of both the ATA and ETA guidelines.

Study Limitations

One of the limitations of our study is the small sample size. Furthermore, since not all AUS cases were operated on, we cannot calculate the true ROM. We only know that the true ROM falls between ROM-H and ROM-O. However, given our high operation rates, the ROM-H we report most likely closely reflects the true ROM.

Conclusion

The literature indicates that the ROM in AUS in children is higher than in adults. Recommending FNAB six months later for AUS cases may lead to a significant number of delayed malignancy diagnoses. However, performing surgery on every AUS case often results in unnecessary surgery. Based on our findings, it is essential to consider the subgroups in the latest TBSRTC classification for the management of AUS cases. The association between nuclear atypia and malignancy has been clearly demonstrated, and lobectomy appears to be a more appropriate approach in AUS cases with nuclear atypia to avoid diagnostic delay and unnecessary surgery. Therefore, management decisions for AUS cases should be based on the presence of nuclear atypia, and guidelines should be updated according to the latest Bethesda classification to avoid diagnostic delays or leaps and subsequent unnecessary surgery.