Introduction

Poorly differentiated thyroid carcinomas (PDTC) are uncommon high-grade follicular cell-derived thyroid carcinomas that are morphologically and biologically intermediate between differentiated thyroid carcinoma (DTC) and anaplastic thyroid carcinoma (ATC) [1]. Poorly differentiated thyroid carcinomas and high-grade differentiated thyroid carcinomas belong to a group of neoplasms that are classified as high-grade follicular cell-derived non-anaplastic thyroid carcinomas [1,2,3]. The 2022 WHO classification of thyroid tumors carried over the endorsement of the Turin morphologic consensus criteria to distinguish PDTCs from other high-grade differentiated thyroid carcinomas [1, 2]. Accordingly, the diagnosis of PDTC is restricted to an invasive follicular cell-derived thyroid carcinoma with solid/trabecular/insular growth pattern unassociated with diagnostic nuclear features of papillary thyroid carcinoma but have one of the following features: (i) tumor necrosis, (ii) ≥ 3 mitoses per 2 mm2 and (iii) convoluted nuclei [1, 2].

Most PDTCs occur in adults with a mean age of 60 years, and pediatric manifestations are even rarer [2,3,4,5,6,7,8,9,10,11,12,13]. The knowledge about PDTC mostly stems from studies in adult-onset disease which may arise either de novo or from an existing differentiated thyroid carcinoma component [1]. At a molecular level, adult PDTCs typically harbor ‘early’ genetic driver mutations in BRAF (BRAF p. V600E) and RAS genes [3, 14]. They also carry aggressive secondary mutations — ‘late’ changes, such as TERT promoter and in some cases other alterations including but not limited to PIK3CA and TP53 [3,4,5,6,7,8,9,10,11,12,13,14,15,16]. Distinct clinicopathologic features of pediatric thyroid carcinomas have been an area of interest in thyroid pathology [17]. To date, there have been a few studies focusing on molecular and/or clinicopathological characteristics of pediatric PDTCs [18,19,20].

Recent discoveries in the field of pediatric thyroid pathology have linked DICER1 mutations to benign follicular cell-derived tumors (e.g., follicular adenoma with papillary architecture, follicular nodular disease) [20,21,22], low-risk follicular cell-derived differentiated thyroid carcinomas [20,21,22,23,24] and a subset of pediatric PDTCs [18, 20]. DICER1 is located on chromosome 14q32.13 and encodes a ribonuclease (RNase) IIIb protein which plays a central regulator role in the miRNA processing [25]. Heterozygous germline pathogenic DICER1 mutations cause DICER1 syndrome, predisposing to a variety of neoplasms of mostly pediatric onset [26]. Despite the widely recognized prognostic role of TERT promoter mutations in follicular cell-derived thyroid carcinomas [27, 28], DICER1 mutations in pediatric and adult-onset differentiated thyroid carcinomas were consistently seen in the setting of follicular patterned low-risk disease [21, 22, 24]. The dismal outcome of DICER1-harboring pediatric PDTCs, which is enriched in fatal or recurrent/progressive disease, is an area of interest that stems from a limited number of reported patients’ data given the rarity of pediatric PDTCs [18, 20]. In light of the former observations on DICER1-mutant pediatric PDTCs, we aimed to assess clinicopathological variables of a series of 5 pediatric (≤ 18 years old) PDTCs by analyzing the status of DICER1 and TERT promoter mutations.

Material and Methods

Patient Selection

Files of the Pathology Department of Istanbul University, Istanbul Faculty of Medicine were searched for patients aged \(\le\) 18 years at the time of diagnosis, who were diagnosed with PDTC, from 2000 to 2019. Five cases that met the Turin criteria were included for data collection and further analyses.

Clinicopathological Variables

Clinicopathologic and follow-up data including age, gender, tumor size, presence or absence of associated well-differentiated thyroid carcinoma component, previous history of radiation to the neck region, mitotic activity per 2 mm2 [based on a mitotic count per 10 mm2 from high mitotic density regions], tumor necrosis, immunohistochemical p53 (Leica DO-7) expression status, and the status of local or distant metastasis were gathered from the medical records. The follow-up time was calculated as the first visit after the surgery and the last visit in 2022. Disease-free survival was estimated from the date of surgery to the date of the first clinical event or last visit.

Molecular Studies

DICER1 and TERT promoter mutation analyses were performed on the best representative formalin-fixed paraffin-embedded tumor samples chosen by one of the study pathologists. When present, the accompanying well-differentiated thyroid carcinoma component was also analyzed. For DICER1, non-tumoral tissues were also analyzed to find out whether the mutation was germline or somatic.

DICER1 Mutation Analysis

Hotspot regions for pathogenic mutations in exons 26 and 27 of the DICER1 gene were analyzed by Polymerase Chain Reaction (PCR) based direct sequencing. Tumor targets were manually micro-dissected from 5-μm thick unstained histological sections. After deparaffinization and rehydration, DNA was isolated from each target using QIAamp DNA FFPE Tissue Kit (50) (catalog #: 56,404) (QIAGEN, Hilden, Germany) in accordance with the manufacturer’s instructions. The primers used were as follows: Exon 26-Forward: 5′ TGGGGATCAGTTGCTATGTG′3, exon 26-Reverse: 5′ CGGGTCTTCATAAAGGTGCT′3, exon 27-Forward: 5′ TGGACTGCCTGTAAAAGTGG′3, exon 27-Reverse: 5′ ATGTAAATGGCACCAGCAAG′3. The purified samples were submitted to direct sequencing in both directions (forward and reverse) by applying reagents from the Big Dye Terminator v3.1 Cycle Sequencing kit (ABI, Applied Biosystems, USA) in accordance with the manufacturer’s protocol. After ethanol precipitation, subsequent products were run on the ABI-3730 (48 capillary) automatic sequencer (Applied Biosystems, USA). Bidirectional sequence traces were analyzed with SeqScape® Software v3.0 (Applied Biosystems, USA) and manually reviewed with the reference sequence of the DICER1 gene (NM_177438.2).

TERT Promoter Mutation Analysis

To identify mutations in the promoter region of the TERT gene (chr5, 1,295,228C > T and 1,295,250C > T), extracted DNA was analyzed using PCR-based direct sequencing. PCR amplifications were performed in a Thermal Cycler (ABI, Applied Biosystems, USA) using the HotStarTaq DNA Polymerase kit (catalog no.: 203205) (QIAGEN, Hilden, Germany) and appropriate primers (Forward: 5′ CAGCGCTGCCTGAAACTC′ 3 and Reverse: 5′ GTCCTGCCCCTTCACCTT′3). PCR reactions were run as a total volume of 50 μL reaction mixtures consisting of nuclease-free water, 5 μL 10 × PCR Buffer, 10 μL Q solution, 1.5 μL of 10 mM dNTP mix (ABI, Applied Biosystems, USA), 5 μL of each primer (4 pmol/μL), 0.25 μL of Hot Start Taq DNA polymerase, and 50 ng of DNA from each tumor. After an initial denaturation at 95 °C for 15 min, 42 cycles were performed of 30-s denaturation at 95 °C, 30 s annealing at 55 °C, and 45 s extension at 72 °C, followed by a final extension of 10 min at 72 °C. The intensity of PCR products was checked by running 5 μL of each PCR reaction with 2 μL of loading dye on a 2% agarose gel. Reagent contamination control was achieved by examining the lane for ‘No DNA’ blank tube. All successful PCR products were then purified using a PureLink PCR Purification Kit (catalog no.: K3100-01) (Invitrogen Life Technologies, USA) in accordance with the manufacturer’s instructions. The purified amplicons were submitted to direct sequencing in both directions (forward and reverse) using reagents from the Big Dye Terminator v3.1 Cycle Sequencing kit (ABI, Applied Biosystems, USA) in accordance with the manufacturer’s protocol. After ethanol precipitation, subsequent products were run on the ABI-3730 (48 capillary) automatic sequencer (Applied Biosystems, USA). Bidirectional sequence traces were analyzed with SeqScape Software v3.0 (Applied Biosystems, USA) and manually reviewed.

Results

Clinicopathologic Features and Follow-Up Data

Five PDTCs lacking a family history of thyroid carcinoma were included in the study. The clinicopathological features were summarized in Table 1. The mean age at the time of diagnosis was 15.4 years (range: 11–18 years). Most of the patients were male (M) comprising 3 of the 5 patients (60%). One patient had a previous history of radiation to the cervical region for Hodgkin’s lymphoma. No patients had a history of DICER1 syndrome-related tumors or other clinicopathological diagnostic features of DICER1 syndrome.

Table 1 Clinicopathological features of 5 pediatric poorly differentiated thyroid carcinomas
Full size table

All patients underwent total thyroidectomy (including one patient that required completion thyroidectomy after the initial partial thyroidectomy). The mean tumor size was 3.9 cm (range: 2–7 cm). Mitotic activity ranged from 3 to 10 mitoses per 2mm2 (Fig. 1). Tumor necrosis (comedo and punctate) was present in 2 cases (Fig. 1). All tumors were completely examined. Three cases (60%) had an associated differentiated thyroid carcinoma component including diffuse sclerosing subtype and invasive encapsulated follicular variants in one and two cases, respectively. In two out three cases with associated differentiated thyroid carcinoma, the PDTC component accounted for at least 30% of the tumor volume whereas the PDTC component accounted for 75% of the tumor volume in the third case. The differentiated thyroid carcinoma component lacked high-grade features (e.g., necrosis or ≥ 5 mitoses per 2 mm2). No abnormal p53 expression (overexpression or global loss) was recorded in all tested tumors.

Fig. 1
figure 1

DICER1-mutant pediatric poorly differentiated thyroid carcinoma (PDTC). PDTCs are characterized by invasive follicular cell-derived thyroid carcinomas with exclusive solid, trabecular and insular growth. The tumor lacked conventional nuclear alterations of papillary thyroid carcinoma. There was tumor necrosis along with convoluted nuclei (A, Hematoxylin and Eosin X20). The tumor showed 3 mitoses per 2 mm2 (B, Hematoxylin and Eosin X40)

Full size image

Four patients had follow-up data. The mean follow-up time was 60.25 months, (range: 18–86 months). One patient with no evidence of disease recurrence died of an unrelated cause 18 months after the diagnosis. All remaining patients were alive at their last visit in 2022. Of the 4 patients with lymph node (LN) dissection, none but one with an associated component of diffuse sclerosing subtype papillary thyroid carcinoma component showed central and lateral LN metastasis at the time of the initial diagnosis. Metastatic foci revealed both papillary carcinoma and PDTC components. Twenty-five months later, the same patient also developed cervical lymph node metastasis of the differentiated thyroid carcinoma component. The FDG PET/CT also showed a suspicious activity in the cervical lymph node metastasis on his last follow-up at 86 months after the initial diagnosis. None of the patients developed distant metastasis during the follow-up period.

DICER1 and TERT Promoter Mutation Analyses

DICER1 hotspot mutation was detected only in one (20%) of five PDTC. The mutation was identified in codon E1813 of exon 27 (Fig. 2). The mutation was confirmed to be somatic by using matched-non-tumor tissue DNA. None of the patients with PDTCs showed pathogenic TERT promoter mutation.

Fig. 2
figure 2

D1CER1 mutation. This electropherogram illustrates p.E1813K mutation identified in case 4

Full size image

Clinicopathologic Variables with Respect to DICER1 Status

The DICER1-mutantPDTC had neither associated differentiated thyroid carcinoma component nor other pathological findings in the non-tumorous thyroid parenchyma. The DICER1-related pediatric PDTC showed widely invasive growth confined to the thyroid parenchyma. Despite the widely invasive growth, the tumor lacked vascular and lymphatic invasion. Two DICER1 wild-type PDTCs had lymphocytic thyroiditis and another one had underlying follicular nodular disease and/or follicular adenomas. Since the original surgical procedure of the patient with DICER1-mutant PDTC was a partial thyroidectomy, a completion thyroidectomy with lymph node dissection was performed. During the follow-up period (72 months) no local recurrence or distant metastases was detected.

Discussion

The current series of 5 PDTCs expands the clinicopathologic spectrum of pediatric-onset PDTCs with respect to clinicopathological variables as well as TERT promoter and DICER1 mutations. While no tumor harbored TERT promoter mutation, this finding is in line with the widely reported absence of TERT promoter mutations in pediatric-onset follicular cell-derived thyroid carcinoma [18, 20, 24, 29, 30]. However, exceptionally TERT promoter mutations may still occur in follicular cell-derived pediatric thyroid carcinomas [31]. Similarly, we did not identify any abnormal p53 expression pattern to suggest a mutant TP53 reactivity in 5 pediatric PDTCs examined in this cohort.

In the current series, DICER1 hot spot mutation was identified in one (20%) of five pediatric PDTCs diagnosed using Turin criteria. This finding still makes DICER1 mutations to be much more prevalent in pediatric PDTCs compared to adult-onset PDTCs with an incidence of 2.5% in the MSK-IMPACT Next Generation Sequencing platform [32]. However, this figure is much lower than the reported frequency of DICER1 mutations in two recent series consisting of 2 (67%) of 3 pediatric PDTCs by Gallant et al. [20] and 5 (83%) of 6 pediatric PDTCs by Chernock et al. [18]. While the variation in the incidence of DICER1 harboring pediatric PDTCs may be attributable to selection bias and/or very low number of overall DICER1-mutant PDTCs in all series, the combined incidence is 57% (8 of 14) of pediatric PDTCs (range: 20–83%) with a female predominance (Female to Male ration: 6 to 2).

Similar to the current series, DICER1 mutations were ‘hotspot’ mutations encoding the metal-ion binding sites of the RNaseIIIb domain of DICER1-related pediatric PDTCs [18, 20]. Chernock et al. reported one patient harboring germline pathogenic DICER1 variant, and this finding underscored the need of germline testing in patients with DICER1-harboring pediatric PDTCs [18]. Although the non-tumorous parenchyma was assessed with respect to hot spot mutations in the RNAseIIIb domain, the lack of the entire DICER1 gene sequencing and unavailability of germline DNA testing data represent methodological limitations of the current cohort to ensure the absence of other rare somatic DICER1 mutations and germline/mosaic DICER1 variants in this series.

Interestingly, two (50%) out of four DICER1-mutant pediatric PDTCs with available follow-up data died of disease with recurrence/distant metastases [18]. Both tumors harbored additional somatic mutations including ATM and CDC73 mutations as well as additional germline heterozygous missense ERCC2 variant in one patient with PDTC (tumor had 40 mitoses per 10 high power field as per the original data), and MAP2K2, RBM10, ARID1A and FLT3 in the other patient (tumor had 37 mitoses per 10 high power field as per the original data) [18]. Both tumors had exceptionally high mitotic activity for a PDTC. In that series, a similar outcome was also identified in an additional pediatric PDTC with no DICER1 mutation and a mitotic activity of 9 mitoses per 10 high power field (as per the original data). In the series of Gallant et al. [20], one of the two DICER1-related pediatric PDTCs with recurrent disease was found to have copy number alterations and gene expression alterations (data taken from supplementary data, e.Fig. 1 from Gallant et al.). The other remaining DICER1-related pediatric PDTCs with progressive disease had no gene expression alterations and no copy number alterations [20]. Unlike the 4 former DICER1-mutant pediatric PDTCs, the patient with DICER1-mutant PDTC that was identified in the current cohort was alive with no recurrent/metastatic disease (follow-up: 72 months). Despite the widely invasive growth, which was confined to the thyroid gland, there was no angioinvasion or lymphatic invasion in this tumor that was entirely submitted for microscopic examination. Of note, this tumor and another PDTC with no dismal outcome also showed tumor necrosis, a morphologic finding that Saliba et al. reported as an independent prognostic factor in pediatric high-grade follicular cell-derived non-anaplastic thyroid carcinomas [19]. In addition, two DICER1-mutant pediatric PDTCs in the series of Chernock et al. were also alive with no evidence of disease (follow-up 3–4 years) [18].

Conclusion

This study shows that indolent clinical behavior can be seen in DICER1-mutant pediatric PDTCs. The occurrence of additional genomic alterations (likely late events) as discussed in some earlier reports [18, 20] may be contributing more to the tumor progression and/or biologic aggressivity of pediatric PDTCs rather than DICER1 mutations alone. Furthermore, some DICER1-related pediatric data did not distinguish lymphatic invasion from angioinvasion, and did not specify the extent of vascular invasion. The lack of vascular invasion (angioinvasion) in the current pediatric PDTC may also explain an indolent biologic outcome. With this observation in mind, the authors would like to underscore prognostic limitations of genomic findings. Therefore, careful interpretation of DICER1 mutation-related risk in pediatric PDTC should be done in conjunction with the presence of additional genetic events as well as well-established clinical and pathologic variables in order to ensure predictive dynamic risk stratification. Additional studies from larger series are needed to corroborate the findings of this study and advance our understanding of pediatric thyroid neoplasia.