The American Thyroid Association (ATA) published management guidelines for children with thyroid nodules and differentiated thyroid cancer (DTC) in 2015.1 Prior to the release of these guidelines, children often were evaluated, treated, and had follow-up recommendations extrapolated from the previously released ATA adult guidelines.2 Over the last decade, it has become increasingly clear that pediatric thyroid cancer has distinct clinical, molecular, and pathological differences compared to their adult counterparts.1 One notable difference is that pediatric thyroid nodules, while less common than in adults, are more likely to be malignant (22 percent to 26 percent versus 5 percent to 10 percent).1 Furthermore, children are more likely to have advanced disease upon presentation with extrathyroidal extension, lymph node involvement, and pulmonary metastasis.1 The molecular and genetic landscape of childhood papillary thyroid cancer (PTC) also is a burgeoning area of research, which has shown Rearranged during Transfection (RET) proto-oncogene/PTC rearrangements to be more common in children, but BRAF mutations, the most common molecular abnormality in adults, are infrequently found.1
Thyroid cancer is the most common endocrine cancer in the pediatric population and represents 2 percent—4 percent of all pediatric malignancies.3-6 Multiple studies have shown that the incidence of pediatric DTC has been increasing over the past several decades.3,5,7 One study used 39 U.S. cancer registries to evaluate the incidence rate from 1998 to 2013 and showed an annual percent change of 4.43 percent per year (from 4.77 per million in 1998 to 8.82 per million in 2013).5 This incidence increase was seen in localized, aggressive, and small (<1 cm) tumors, in addition to those greater than 2 cm. The majority of cases were in the 15- to 19-year-old group, and approximately half of the cases were reported in the last six years of the study.5 Another study examined the Surveillance, Epidemiology, and End Results (SEER) database from 1973 to 2013 and showed a gradual increase in incidence (1.1 percent per year) of pediatric DTC from 1973 to 2006, but then noted a significant uptick in the incidence at 9.6 percent per year after 2006.3 Similar to Bernier et al,5 this investigation also found incidence increases in all tumor sizes.3 Recently Vaccarella et al evaluated pediatric cancer databases that covered 49 countries and territories and found that pediatric thyroid cancer incidence rates are increasing in almost all countries, which also strongly correlated with increased rates in adults.7
Debate exists over whether the increase in incidence of pediatric DTC is real versus simply due to overdiagnosis. The 30-year cancer-specific survival for pediatric PTC is approximately 99 percent to 100 percent, so comparing the incidence trends with mortality trends to determine a true increase is near impossible given that the number of deaths is so small.8 One argument in support of overdiagnosis would be if the increase in pediatric DTC incidences were small indolent tumors, though Qian et al3 and Bernier et al5 both found there to be increases in larger sized tumors and extent of disease. It is also thought that overdiagnosis via incidental findings on imaging would not apply to children as they are less likely than adults to have thyroid nodule screening, and medical imaging of the neck is not routinely performed for other clinical purposes.5,9,10 This argument is countered by the fact that computed tomography (CT) imaging for pediatric patients has more than doubled between 1996 and 2010.11 and there has been an established linear relationship between CT imaging and the increasing incidence of thyroid cancer.8,12
Environmental risk factors also may be contributing to the actual increase in pediatric DTC incidences, in combination with some level of overdiagnosis. Radiation exposure is a known risk factor for developing pediatric thyroid cancer in a dose-dependent relationship. One source of increased radiation exposure may be related to the increased use of CT imaging, as the advances and availability of this modality resulted in 62 million CT scans obtained in 2006 compared with 3 million in 1980 in the United States.3,13 Another potential risk factor may be the increased rise in childhood obesity, via estrogen-related pathways or insulin-resistance mechanisms, which has been shown to have positive association with adult DTC and many other cancers.5,14
Thyroid Nodule Classification Systems
Guidelines for managing thyroid nodules have historically been tailored to the adult population due to a lower prevalence of thyroid cancer in the pediatric population. As recently as the 2009 ATA adult guidelines, pediatric and adult nodules were classified in a similar manner.15 In the evaluation of adult patients with thyroid nodules, two primary nodule classification systems exist: the 2015 ATA Management Guidelines for Adult Patients with Thyroid Nodules and DTC and the American College of Radiology (ACR) Thyroid Imaging Reporting and Data System (TI-RADS).2,16 Given the increased rates of malignancy and more aggressive features in pediatric versus adult thyroid nodules, experts have advised caution in directly applying either of these systems to the pediatric population. In 2015, the ATA spearheaded a task force to create guidelines unique to pediatric patients less than 18 years old.
While the 2015 ATA Adult Thyroid Nodule and DTC guidelines do provide an atlas of nodule sonographic patterns with associated risk of malignancy, the imaging features also are also combined with nodule size to guide fine needle aspiration (FNA) decision-making.2 Utilizing size criteria in the management of pediatric thyroid nodules can be problematic due to differing thyroid volume by age.1 As the size of a nodule alone does not predict the likelihood of malignant histology in children, the 2015 ATA Management Guidelines for Children with Thyroid Nodules and DTC recommend that pediatric nodules should be evaluated using the same pattern-based system with the exception that ultrasound characteristics and clinical context should be used rather than size alone to identify nodules that warrant FNA.1 Concerning ultrasound characteristics in children include hypoechogenicity, irregular margins, increased intranodular blood flow, microcalcifications, and abnormal cervical lymph nodes.1
Given that some nodules do not follow the criteria for sonographic patterns and are unable to be classified using the ATA system, ACR TI-RADS was developed with the intent to classify all nodules based on ultrasound features with a standardized lexicon.16,17 Based on a scoring system in which more suspicious features are awarded higher points, ACR TI-RADS designates levels of relative probability of malignancy and provides management recommendations (that is, no FNA versus follow versus FNA) based on TI-RADS level and nodule size.16 ACR TI-RADS was designed with the intent of decreasing unnecessary biopsies of benign and indolent malignant nodules in the adult population, which is problematic in the evaluation of pediatric thyroid nodules given known differences in malignancy rates and aggressive disease features in this population. Current TI-RADS recommendations have yet to be extensively validated in children and, thus far, results have been somewhat conflicting. A study from Richman et al included 77 pediatric thyroid cancers, of which 17 would not have undergone FNA if TI-RADS management guidelines were applied.18 Yet TI-RADS has performed well in other, though smaller, cohorts. A small study performed by Lim-Dunham et al included only 20 cancers, with only one cancer missed.19
While the TI-RADS classification system is based on ultrasound features (composition, echogenicity, shape, margin, and echogenic foci), it also employs size criteria for FNA recommendations. Recent studies propose employing a more stringent criteria for FNA in the pediatric population. Due to the risk of downgrading concerning nodules, some authors propose performing FNA on all thyroid nodules <1 cm with suspicious features and decreasing size criteria for FNA in the TI-RADS 3 through 5 categories in pediatric patients.18,20 Additionally, some centers advocate forgoing the TI-RADS informed FNA thresholds and instead considering surgery for all lesions > 4 cm due to concerns for false-negative cytologic findings.19,21 More research is needed on how to apply and potentially modify both the ATA and TI-RADS nodule classification systems for the pediatric population to optimize surgical outcomes.
Molecular Genetics of Pediatric Thyroid Cancer
As would be expected given the differences in clinical behavior of pediatric and adult thyroid cancer, research suggests that the genetic landscape is distinct in these populations. A sentinel investigation that greatly advanced our understanding of the molecular genetics of thyroid cancer in adults was the Cancer Genome Atlas Project (TCAP), which characterized the primary driver events of tumorigenesis in PTC. This study introduced a new paradigm of molecular classification of adult PTC into two subgroups—BRAF-like PTC and RAS-like PTC, based on findings that mutually exclusive genetic alterations in the BRAF and RAS genes result in distinct signaling and differentiation properties, with activation of the mitogen-activated protein kinase (MAPK) and phosphatidylinositol 3-kinase (PI3K) pathways.22
Unique differences in the spectrum of PTC genetic alterations have emerged from several smaller pediatric cohort studies. In contrast with adult PTC, oncogenic gene fusions occur with greater frequency in the pediatric population (50 percent to 60 percent in children compared to only 15 percent in adults), and point mutations are less prevalent, found in 30 percent of pediatric thyroid cancers compared to 70 percent in adults.23-25
Gene fusions involving RET and the neurotrophic tyrosine receptor kinase (NTRK) genes (NTRK1 and NTRK3) are the most common oncogenic drivers in pediatric PTC.23,26,27 RET fusions appear to occur in approximately 25 percent to 30 percent of sporadic pediatric PTC (range 14 percent to55 percent). This proportion increases to 45 percent (range 14 percent to 87 percent) in patients with radiation exposure.24 While several RET/PTC rearrangements continue to be identified, the most common are RET/PTC1 and RET/PTC3, which are present in sporadic and radiation-induced PTCs.25 Less commonly observed oncogenic fusions include NTRK (1, 2 and 3), BRAF, and ALK. These rearrangements have been identified in approximately 10 percent (range: 0 percent to 26 percent), 10 percent (range: 0 percent to 18 percent), and 5 percent (range: 0 percent to 21 percent) of sporadic pediatric PTC, respectively24.
BRAF point mutations are regularly found in adult PTC. The BRAF gene is the most common location for a point mutation in pediatric PTC,23,25,28 reportedly occurring in 25 percent to 30 percent of lesions (range: 0 percent to 63 percent).24 While more than 40 mutations have been identified in BRAF to date, more than 95 percent are V600E,23 though this estimate may reflect detection bias since many studies' methodology only assessed this single mutation. The BRAF V600E point mutation also is prevalent in childhood PTC, although it is identified at a much lower rate than in adult PTC.29 BRAF V600EPTCs are detected mostly in adolescents, whereas fusion oncogene PTCs are predominantly found in younger patients.30,31 A striking difference in the clinical correlation of this mutation in children, compared to adults, is that it does not portend worse prognosis or a more aggressive phenotype.32,33 There may be, however, a higher incidence of cervical lymph node metastases at presentation, which may affect the extent of the index surgery.25,34 Alternatively, it has been shown that gene rearrangements and fusions, such as RET/PTC1 and NTRK3-ETV6, may be correlated with more invasive disease in children,34,35 and are associated with more advanced stage, extra-thyroidal extension, infiltrative growth, multifocality, locoregional and distant metastases, as well as higher risk of persistent and recurrent disease.23,30,34-36
DICER 1 mutations are also commonly identified in pediatric thyroid nodules. While the majority of these are associated with benign multinodular goiter, recent publications suggest that DICER 1 mutations can be seen in childhood and adolescent poorly differentiated thyroid carcinomas.37 The minority of these cases are identified in patients with DICER1 syndrome; however, it is important to recognize the potential need for genetic counseling in these patients. In fact, pediatric thyroid carcinomas, particularly those identified in younger children, may be associated with hereditary syndromes. Physicians should be aware of syndromes associated with thyroid tumors and provide children and their families with appropriate genetic counseling resources.
The use of molecular testing has radically transformed diagnostic algorithms in adults with cytologically indeterminate thyroid nodules. Because these molecular diagnostic studies had not been amply validated in children, the 2015 ATA guidelines for the management of DTC in children did not endorse their routine use in clinical practice to augment the diagnostic utility of FNA1. Furthermore, the pediatric ATA guidelines recommended definitive surgery (lobectomy plus isthmusectomy) over repeat FNA for most nodules with indeterminate cytology. This recommendation was based on observed increased rates of The Bethesda System for Reporting Thyroid Cytopathology (TBSRTC) indeterminate categories in children compared to adults, as well as an apparent increased likelihood of malignancy among these indeterminate categories in children.1,38 Others have shown that the rates of indeterminate cytology may be lower than previously reported (about 19 percent, rather than 35 percent), and the risk of malignancy greatly varies across indeterminate categories.39 As such, some experts advocate for developing category-specific recommendations even for the individual indeterminate groups, whereby molecular testing and serial FNA may be the most beneficial approach for AUS/FLUS or FN/SFN thyroid nodules. A recent report supports this repeat biopsy strategy, given that nearly one third of AUS/FLUS nodules were benign on subsequent FNA examinations.40 We urge the readers to carefully consider local institutional performance indices of FNA accuracy when interpreting FNA results to inform treatment plans.
The use of oncogene panels can aid in operative planning and guide shared medical decision-making discussions with families. Studies from pediatric cohorts suggest that nearly 17 percent of FNAs may be positive for a mutation or rearrangement, which correlates with malignancy in 100 percent of cases.41 In a recent review, Christison-Lagay and Baertschiger summarized specific surgical treatment recommendations based on the findings of molecular testing of FNA aspirates.36 For Bethesda III and IV category nodules, should a BRAFmutation or gene rearrangement/fusion (RET/PTC or NTRK3-ETV6) be identified on molecular genetics testing, they suggest total thyroidectomy based on the high risk of malignancy. Lobectomy may be considered if no concerning genetic alteration is identified; for example, a RAS mutation or PAX8-PPARG rearrangement.23,36 Bauer recommended more extensive surgical treatment with total thyroidectomy plus central neck dissection, and consideration of therapeutic lateral neck dissection when clinically warranted, for RET/PTC rearrangement, NTRK1/3 fusions, and BRAF mutations based on increased risk of DTC with invasive disease.23
Molecular genetics also play a major role in guiding adjuvant therapy plans in pediatric patients with PTC. Tyrosine kinase inhibitors such as sorafenib and lenvatinib have shown promise in adult patients with iodine-refractory disease. Similar reports of favorable response to lenvatinib have been documented in children with refractory disease.42 Additionally, selpercatinib was approved by the FDA for use in children older than 12 years with RET fusion-positive (and RAI-refractory) thyroid cancer.34 Lee et al recently reported that selective RET fusion-directed therapy with larotrectinib and selpercatinib led to tumor response and restored radioiodine avidity in pediatric patients with fusion oncogene PTC and 131I-refractory advanced disease.31
Genetic testing is paramount to the diagnosis, staging, surgical treatment, and adjuvant therapy strategies for pediatric patients with PTC. Ongoing studies, such as the NCI-COG Pediatric MATCH (Molecular Analysis for Therapy CHoice) Trial, in which children with refractory or recurrent cancer receive targeted therapies directed to molecular changes in their tumors,43 as well as many other ongoing research efforts, will continue to advance the field of precision medicine and ultimately improve the care of pediatric patients with thyroid cancer.
