The Management Guidelines for Children with Thyroid Nodules and Differentiated Thyroid Cancer from The American Thyroid Association recommends fine-needle aspiration (FNA) as an essential tool for evaluation and management of pediatric thyroid nodules, and The Bethesda System for Reporting Thyroid Cytopathology (TBSRTC) as a guideline to categorize FNA cytologic findings. A few publications have focused on TBSRTC categorization in children.
To evaluate our institutional experience of using TBSRTC in the pediatric population.
We conducted a retrospective data search for thyroid FNA specimens from patients younger than 21 years who were assessed by using TBSRTC categorization between January 1, 2011, and September 30, 2017. Cytologic and histologic diagnoses, rate of surgical follow-up, rate of histology-proven malignancy, and cytologic-histologic concordance were assessed in comparison with our previously published adult experience.
Among a total of 201 specimens, benign category accounted for 103 (51.2%), followed by 35 (17.4%) malignant, 30 (14.9%) atypia of undetermined significance/follicular lesion of undetermined significance (AUS/FLUS), 14 (7.0%) nondiagnostic, 10 (5.0%) follicular neoplasm/suspicious for follicular neoplasm (FN/SFN), and 9 (4.5%) suspicious for malignancy (SFM) categories. Surgeries were subsequently performed in 100 of the 201 cases (49.8%). All surgically removed nodules categorized as malignant and SFM by FNA showed malignancy, while malignancy was evident in various proportions of specimens classified as AUS/FLUS (5 of 16; 31.3%), FN/SFN (1 of 9; 11.1%), and nondiagnostic (1 of 7; 14.2%). No malignancy was identified in the benign category. Cytologic-histologic concordance achieved 100% in malignant, SFM, and FN/SFN categories, and exceeded beyond 85% and 50% for benign and AUS/FLUS categories, respectively.
TBSRTC is effective for appropriate categorization of pediatric thyroid nodules. Risk of malignancy in thyroid nodules is higher for the pediatric than adult population. Understanding the difference in TBSRTC categorization between children and adults may aid in achieving more appropriate evaluation and management of pediatric thyroid nodules.
For adult patients with thyroid nodules, fine-needle aspiration (FNA) has been considered an important triage tool for distinguishing thyroid lesions that require surgical interventions from lesions that may be conservatively managed with clinical and/or radiologic follow-up. The Bethesda System for Reporting Thyroid Cytopathology (TBSRTC),1 a uniform, 6-tiered reporting system with standardized nomenclature and definitions, has been widely used for the assessment and interpretation of thyroid specimens obtained from adult patients. In comparison with adults, thyroid nodules in the pediatric population are less common with an estimated prevalence of 1.6% to 1.8%.2,3 However, a greater cancer risk has been reported in children than adults and thyroid cancers account for 4% of all malignancies in children and adolescents.4 The American Thyroid Association (ATA) published Management Guidelines for Children with Thyroid Nodules and Differentiated Thyroid Cancer5 in 2015. Accordingly, FNA is recommended as an essential tool for the evaluation and management of pediatric thyroid nodules. Further, ultrasound characteristics and clinical context rather than size alone are assessed to identify the nodules that need FNA, and all FNAs in children should be performed under ultrasound guidance. TBSRTC, used for adults, is recommended to categorize the cytologic findings in children. Many studies have investigated cytologic categorizations of thyroid nodules in adults, using TBSRTC as guidelines. On the contrary, fewer published data have been focused on the usage of TBSRTC categorization for analysis of pediatric thyroid nodules.6–11
TBSRTC has been implemented in our institution for assessment of thyroid nodule in both adults and the pediatric population since January 2011. The current study was conducted to evaluate our institutional experience of using TBSRTC categorization in pediatric thyroid nodules in comparison with our previously published adult experience.12
MATERIALS AND METHODS
The study was approved by the Institutional Review Board of the University of Michigan, Ann Arbor. A retrospective search of the electronic medical record systems of the University of Michigan–Michigan Medicine was performed by spanning a period of 6 years, 9 months from January 1, 2011, to September 30, 2017. All thyroid FNA specimens obtained from patients younger than 21 years were retrieved, including mainly in-house cases, and a subset of cases from transfer of care and expert consultation service. For each case, age and sex of the patient, and number of targeted nodules were recorded. For in-house cases, interventional radiologists performed ultrasound-guided FNAs with cytopathologist-assisted rapid on-site evaluation. Briefly, 2 conventional smears were made for each pass. One smear was air-dried and stained with Diff-Quik protocol to be evaluated immediately for specimen adequacy, while the other smear was fixed with Sprayfix (Leica Biosystems Richmond Inc, Richmond, Illinois) and later stained with Papanicolaou stain. The needle was then rinsed in Cytolyt solution for a ThinPrep (Hologic Inc, Marlborough, Massachusetts) and/or a cell block. For the transfer and consulting cases, the specimens were processed as conventional smears and/or ThinPrep slides at the referring institutions.
Individual FNA specimens were originally interpreted by using TBSRTC and were categorized into 1 of the following 6 categories: nondiagnostic, benign, atypia of undetermined significance/follicular lesion of undetermined significance (AUS/FLUS), follicular neoplasm/suspicious for follicular neoplasm (FN/SFN), suspicious for malignancy (SFM), and malignant. Cytologic diagnoses and the corresponding follow-up histologic diagnoses, where available, were recorded. Using the histologic diagnosis as a gold standard, rates of malignancy and neoplasia were assessed for each of TBSRTC categories. Furthermore, re-review of slides prepared from the surgical specimens with a histologic diagnosis of malignancy, particularly papillary thyroid carcinoma, was performed to identify the cases that fulfilled the diagnostic criteria of noninvasive follicular thyroid neoplasm with papillary-like nuclear features (NIFTP).13 Cytologic-histologic correlation was then performed on applicable cases. Correlations were deemed concordant in the FN/SFN and SFM categories if there was histologic evidence of neoplasia or malignancy. FLUS cases were somewhat difficult to correlate. However, we considered histology-proven neoplasia or malignancy as concordant, accounting for the fact that atypical cytologic findings favor to represent atypical proliferation including neoplastic or malignant processes. Additionally, we performed comparison of our pediatric study findings to our published data on adult patients.12
A total of 201 FNA specimens were retrieved from 148 patients. Patient ages ranged from 5 to 20 years with a female predominance (female to male ratio, 4.9:1). Fine-needle aspiration was performed on 1 to 4 nodules per patient. Of these patients, 112 had 1 nodule with 1 FNA performed. Thirteen patients had 1 nodule biopsied on 2 separate occasions. Fourteen patients had 2 nodules with 1 FNA performed on each. One patient had 2 nodules with subsequent repeated FNA on both. Five patients had 3 nodules, each with 1 FNA performed. One patient had 3 nodules with 1 repeated FNA on 1 nodule and 2 repeated FNAs on the other 2 nodules. One patient had 4 nodules, each with 1 FNA. One patient had 4 nodules, with 2 nodules having been sampled first followed by repeated sampling of the first 2 nodules, with FNA of the additional 2 nodules.
Table 1 summarizes the distribution of the FNA specimens across TBSRTC categories, the rate of surgical follow-up, the rate of histology-proven malignancy and neoplasia for each individual category of TBSRTC, as well as cytologic-histologic concordance. Among the total of 201 specimens, benign category accounted for 103 cases (51.2%), followed by 35 (17.4%) malignant, 30 (14.9%) AUS/FLUS, 14 (7.0%) nondiagnostic, 10 (5.0%) FN/SFN, and 9 (4.5%) SFM categories. Nearly half (n = 100; 49.8%) of the total 201 specimens were followed by surgical interventions. The surgical rate showed a wide range across TBSRTC categories with the greatest rate for FN/SFN (9 of 10 cases; 90%), followed by malignant (31 of 35 cases; 88.6%), SFM (6 of 9 cases; 66.7%), AUS/FLUS (16 of 30 cases; 53.3%), nondiagnostic (7 of 14 case; 50.0%), and lastly benign (31 of 103 cases; 30.1%). Hemithyroidectomies were performed in 2 nondiagnostic, 23 benign, 10 AUS/FLUS, and 7 FN/SFN cases. Total thyroidectomies were performed in 5 nondiagnostic, 8 benign, 6 AUS/FLUS, 2 FN/SFN, 5 SFM, and 31 malignant cases. In 1 SFM case, the patient underwent lobectomy followed by completion thyroidectomy. All surgically removed nodules categorized as malignant or SFM by cytologic evaluation showed malignancy on excisions. Histology-proven malignancy was evident in various proportions of specimens classified as AUS/FLUS (5 of 16 cases; 31.3%), FN/SFN (1 of 9 cases; 11.1%), and nondiagnostic (1 of 7 cases; 14.2%). The overall rate of malignancy reached 44.0%. Most malignancies were papillary thyroid carcinoma, mainly conventional type and rarely follicular variants. In addition, 2 follicular carcinomas and 3 medullary thyroid carcinomas were identified. No malignancy was identified upon histologic examination of nodules categorized as benign. NIFTP was not identified upon the re-review of the surgical specimens with a histologic diagnosis of malignancy or neoplasia, particularly papillary thyroid carcinoma. All cases within the FN/SFN category had neoplastic outcome on excision, including 6 follicular adenomas, 2 Hürthle-cell adenomas, and 1 papillary thyroid carcinoma, follicular variant. Cytologic-histologic concordance achieved 100% in malignant, SFM, and FN/SFN categories. In addition, the concordant rate was 87.1% and 56.3% for benign and AUS/FLUS categories, respectively.
Table 2 compares the pediatric data from the current study with the previously published data from adult patients in our institution.12 The current study demonstrated a lower nondiagnostic rate and markedly higher proportion across almost all of TBSRTC diagnostic categories, except benign and nondiagnostic. The rate of malignant category was 3.6 times higher, SFM 6.4 times higher, SFN 2.3 higher, and FLUS 2 times higher in our pediatric study. More surgical interventions were performed in pediatric thyroid nodules classified as nondiagnostic, benign, AUS/FLUS, and FN/SFN. Overall surgical rate was also greater in children (100 of 201 cases; 49.8%) than adults (390 of 2159 cases; 18.1%). The rate of histology-proven malignancy was greater in children (44 of 100 cases; 44.0%) than adults (144 of 390 cases; 36.9%). Lastly, a better cytologic-histologic concordance was observed in children than adults for malignant, SFM, FN/SFN, and AUS/FLUS categories.
Our study emphasizes the utility of TBSRTC for cytologic evaluation of pediatric thyroid nodules preoperatively. Table 3 compares data from the current study with those previously reported by other institutions.6–11 As can be seen, a greater proportion of pediatric thyroid nodules evaluated in our institution were categorized as malignant or AUS/FLUS, while a lesser proportion fell into the benign or nondiagnostic category, in comparison to the data of most if not all reported studies. In addition, our data revealed the second highest overall rate of surgical follow-up and rate of histology-proven malignancy. With regard to the rate of histology-proven malignancy in AUS/FLUS and FN/SFN categories, a wide range is evident among current and the aforementioned studies, from 8% to 50% and from 10% to 100% for AUS/FLUS and FN/SFN categories, respectively. However, all of these studies contained limited case numbers with surgical follow-up in both AUS/FLUS (n = 4–25) and FN/SFN (n = 4–19). It is noteworthy to mention that variations exist among different institutions with regard to many factors, including but not limited to the tertiary care referral aspect, prevalence of benign and malignant pediatric thyroid nodules, as well as diagnostic thresholds for assessing cytologic and/or histologic specimens for each diagnostic category.
We further looked into the relatively high surgical follow-up in the benign category within our cohort. The patients received surgical treatments for large multinodular goiter, lymphocytic thyroiditis, clinical concerns for a neoplastic process, or family history of thyroid malignancy. Eight cases were managed by total thyroidectomy and the remaining 23 cases had hemithyroidectomy. Five cases with clinical concerns for neoplasm had histology-proven follicular adenoma or single adenomatoid nodule and 3 cases had a malignant nodule on the contralateral side. It is possible that the increased risk of malignancy in pediatric thyroid nodules and the referral nature of our hospital may have contributed to this higher surgical follow-up. However, potential overtreatment of benign thyroid in the pediatric population may also be suggested.
Because surgeons may want to minimize the pediatric reoperative surgical rate in the suspicious categories, we retrospectively investigated the partial and total thyroidectomy cases following SFN and SFM diagnoses. Of 10 SFN cases, 7 patients underwent partial thyroidectomy and 2 patients underwent total thyroidectomy with malignancy identified in 1 of these 2 cases. In only 1 case the patient underwent lobectomy followed by completion thyroidectomy for the SFM category, while the remaining 5 cases involved total resection after the SFM cytologic interpretation. It is crucial for pathologists, endocrinologists, and surgeons to have a clear understanding of practice patterns in their own institutions in order to provide management appropriate to children with thyroid nodules.
In regard to the comparison with our adult study,12 the current study demonstrated a lower nondiagnostic rate, likely resulting from more frequent utilization of cytologist-assisted rapid on-site evaluation of the FNA specimens. Additionally, a higher proportion of pediatric nodules were categorized into malignant, SFM, FN/SFN, and AUS/FLUS categories. A greater overall surgical rate, a higher rate of histology-proven malignancy, as well as better cytologic-histologic concordance for malignant, SFM, FN/SFN, and AUS/FLUS categories, were observed in children than in adults. While interpreting these differences, one should note possible selection or sampling bias due to the limited and relatively smaller pediatric case cohort contained in the current study. Multiple factors may be related to the performance of more surgical interventions for pediatric thyroid nodules classified as nondiagnostic or indeterminate. One explanation for this difference could be the increased vigilance in children from both parents and treating clinicians, as there is a higher incidence of malignancy for thyroid nodules in this age group as previously reported by the ATA Task Force.5 Another plausible explanation is the tertiary care and the referral nature of our pediatric hospital.
Following the implementation of new nomenclature for NIFTP,13 the second edition of TBSRTC has modified the implied risk of malignancy for the indeterminate categories as follows: 6% to 8% for AUS/FLUS, 10% to 40% for FN/SFN, and 45% to 60% for SFM,14 which included most findings in the adult population. All the papillary thyroid carcinoma cases and follicular adenomas in this study were re-reviewed again and none of our surgical follow-up samples fulfilled the criteria for NIFTP diagnosis. Histology-proven risk of malignancy rate revealed by the current study for AUS/FLUS (31.3%) and SFM (100%) categories exceeded the upper limit implied by TBSRTC. Also, the malignant rate for AUS/FLUS specimens contained in the current study appeared to be slightly higher than the previously reported percentage of 28% for AUS/FLUS lesions in the pediatric population.5 Again, the limitation of the smaller case cohort and availability of surgical follow-up for every FNA case, especially within AUS/FLUS category, may account for these differences. Nonetheless, these findings may support the notion by ATA Guidelines Task Force on Pediatric Thyroid Cancer that the pediatric thyroid nodules carry a greater risk of malignancy, and surgery rather than repeated aspiration may be favored for nodules interpreted as indeterminate,5 that is, AUS/FLUS. On the other hand, molecular testing has been considered a useful tool for further identification of malignant nodules contained in the indeterminate categories. This approach is helpful in effective management of indeterminate thyroid nodules in terms of expediting surgery and minimizing unnecessary procedures.6 Unfortunately, utility and determining factors of this ancillary modality have not been well delineated for the pediatric population in our practice.
In summary, TBSRTC is effective for appropriate categorization of pediatric thyroid nodules. Risk of malignancy in thyroid nodules is higher in the pediatric than adult population. Understanding the difference in TBSRTC categorization, particularly the implied risk of malignancy between children and adults, may aid in achieving more appropriate evaluation and management of pediatric thyroid nodules.
The authors have no relevant financial interest in the products or companies described in this article.
The abstract was initially presented at the United States and Canadian Academy of Pathology (USCAP) 107th Annual Meeting; March 21, 2018; Vancouver, British Columbia, Canada.