Secretory Carcinoma of the Thyroid: A Case Report and Update of Literature

Abstract

Primary secretory carcinoma (SC) of the thyroid gland is a rare neoplasm, characterized by the presence of oncogenic ETV6::NTRK3 fusions, which are amenable to tropomyosin receptor kinase (TRK) inhibitor therapy. Despite its morphologic, immunophenotypic, and genetic similarities to SC of the salivary and mammary glands, diagnostic pitfalls may arise in differentiating from papillary thyroid carcinoma due to overlapping features such as papillary growth, nuclear irregularity, and variable expression of PAX8. Tumor misclassification may lead to delayed consideration of molecular testing and targeted therapy. A total of 13 cases of thyroid SC have been documented in the literature, indicating a tendency for advanced clinical presentation followed by a protracted clinical course, with most patients surviving until the end of the study period despite some experiencing recurrences. However, tumor-related mortality occurred in around 30% of cases, with the overall survival ranging from days to years, underscoring the variability in tumor behavior and the need for further research efforts. Among documented cases of thyroid SC, prognostic factors established for salivary SC have shown broad distributions, including a mitotic activity ranging from < 1 to 10 per 10 high-power fields and variable presence of necrosis, awaiting additional case experience to better elucidate their relevance in thyroid SC. We hereby present a 61-year-old female patient with widely metastatic thyroid SC treated with larotrectinib and provide an updated review of the literature on the molecular pathogenesis and clinicopathologic characteristics of this rare entity.

Introduction

Primary thyroid secretory carcinoma (SC) is a rare malignancy with histological and immunophenotypic characteristics similar to SC arising in the salivary gland and mammary settings. Since Stevens et al. described the first case in 2015 [1, 2], there have been a total of 13 cases of thyroid SC published to this date (Table 1) [1–11]. Histologically, thyroid SC exhibits various architectural patterns, including cribriform, microcystic, papillary, micropapillary, and solid growth, which often coexist within the same tumor. Additionally, SC produces eosinophilic and occasionally purple-basophilic material resembling colloids. Cytomorphologically, thyroid SC shares several features with papillary thyroid carcinoma (PTC), such as pale chromatin, nuclear irregularity with grooves and rarely pseudoinclusions, and occasional psammomatous calcifications, posing significant diagnostic challenges in fine-needle aspirations and surgical resections with documented instances of initial misclassification [3, 4, 7]. Immunophenotypically, as recently reviewed by Huang et al. [9], thyroid SC frequently expresses PAX8 to variable intensities and cellular percentages using monoclonal [7] or polyclonal [4] antibodies, while being typically negative for TTF1 and thyroglobulin. In the reported cases, the secretory phenotype has been supported by positive S100 and mammaglobin immunoreactivity, further aiding in the differential diagnosis from PTC. All thyroid SC cases molecularly analyzed have shown ETV6 gene rearrangements, with ETV6::NTRK3 fusions universally identified in cases with the translocation partner resolved. Clinically, most thyroid SC presented as T3 to T4 disease (85%) with frequent lymph node involvement (62%) at the initial diagnosis (Table 1) [2–12]. Although most patients survived until the end of the studies, tumor-related mortality has occurred in around 30% of cases due to aggressive initial presentation [8] or resistance to treatment upon recurrence [3, 11]. Early recognition of thyroid SC is critical, especially in aggressive cases, to prompt timely molecular identification of NTRK gene fusions that are highly responsive to tropomyosin receptor kinase (TRK) inhibitors. We herein present the fourteenth case of thyroid SC, notable for gross extrathyroidal extension, widespread distant metastases and response to larotrectinib therapy, and discuss the entity’s clinical behavior and management based on the recent literature.

Table 1.

Reported clinicopathologic characteristics of primary thyroid secretory carcinoma

Study	Age (years)	Sex	Radiation Exposure	Size (cm)	Stage	Mitoses	Necrosis	Distant Metastasis	Background CLT	Molecular Testing	Postoperative Treatment	Followup (months)	Outcome
[3]	36	F	No	4.5	T4aN1	< 1/30 HPF	No	Yes, at recurrence	Present	RT-PCR (ETV6::NTRK3 e4 to e14)	RAI, EBRT, sorafenib, sunitinib	107	Two regional recurrences; DOD
[4]	72	F	No	2.9	T4aN0	1/10 HPF	No	No	Present	FISH (ETV6); NGS (ETV6::NTRK3 e4 to e14)	RAI	238	Two regional recurrences; DOOC
[4]	47	M	No	7.0	T3N1b	1/10 HPF	No	No	Present	FISH (ETV6); NGS (ETV6::NTRK3 e4 to e14)	RAI, EBRT	230	Three regional recurrences; NED
[4]	65	F	No	6.5	T4aN0	1/10 HPF	No	No	Present	FISH (ETV6)	RAI, IMRT	24	NED
[2]	74	M	NR	4.0	T2N1b	10/10 HPF	Present	Yes	NR	FISH (ETV6)	Chemoradiation	12	DOD
[2]	52	F	NR	2.4	T2N1a	3/10 HPF	No	No	NR	FISH (ETV6)	None	26	NED
[1, 2]	55	F	NR	2.6	T4N1a	2/10 HPF	No	No	NR	FISH (ETV6)	Radiation	43	NED
[5, 6]	58	F	No	4.0	T4aN0	7/10 HPF	Present	No	Present	FISH (ETV6)	EBRT, chemotherapy	14	NED
[7]	72	F	NR	2.5	T3N0	NR	NR	No	Present	NR	None	18	NED
[8]	74	F	NR	7.6	T4bN1b	2/10 HPF	Present	Yes, at diagnosis	NR	FISH (ETV6)	None	20 days	DOD
[9]	36	F	No	5.0	T3aN0	NR	NR	Yes, at recurrence	NR	FISH (ETV6); NGS (ETV6::NTRK3 e5 to e15)	Larotrectinib	33	Three regional recurrences; AWD
[10]	71	F	No	3.2	T3N1a	NR	NR	No	NR	NGS (ETV6::NTRK3 e5 to e14)	RAI, radiation, larotrectinib	71	Three regional recurrences; NED
[11]	49	M	No	6.0	T4aN1	3/10 HPF	Present	Yes, at recurrence	No	NGS (ETV6::NTRK3 e4 to e14)	Larotrectinib, radiation, chemotherapy	181	Three regional recurrences; DOD
Current case	61	F	No	5.6	T4bN1a	1/10 HPF	Equivocal	Yes, at diagnosis	Present	NGS (ETV6::NTRK3 e4 to e14)	Chemoradiation, larotrectinib	6	AWD

Abbreviations: AWD, alive with disease; CLT, chronic lymphocytic thyroiditis; DOD, died of disease; DOOC, died of other cause; e, exon; EBRT, external beam radiation therapy; FISH, fluorescence in situ hybridization; HPF, high-power fields; IMRT, intensity modulated radiation therapy; NED, no evidence of disease; NGS, next-generation sequencing; NR, not reported; RAI, radioactive iodine

Case Report

Clinical Presentation

A 61-year-old woman presented with a six-month history of hoarseness and palpable left neck mass. She was an active smoker (27 total pack years) previously diagnosed with hypothyroidism secondary to chronic lymphocytic thyroiditis under levothyroxine supplement. There was neither prior cancer diagnosis nor past radiation exposure. During the initial clinical visit, flexible laryngoscopy was performed and revealed left vocal cord paralysis. Neck sonography showed an irregular hypoechoic left thyroid mass, TI-RADS 5, which underwent fine-needle aspiration and was interpreted as suspicious for papillary carcinoma. Neck soft tissue and chest CT scans demonstrated an ill-defined left thyroid mass that measured 5.6 × 5.2 × 3.7 cm with rightward tracheal deviation. The pharyngeal and laryngeal structures and bilateral parotid and submandibular glands were visualized with unremarkable appearance. The scans also noted lytic changes in the manubrium and a 4.9 cm right adrenal nodule. Overall, a preoperative clinical staging of cT4bN1M1 (stage IVB) was rendered.

Intraoperative and Pathologic Findings

The patient underwent total thyroidectomy with central neck lymph node dissection. Intraoperatively, the left lobe mass was noted to invade across the midline into the right lobe and show gross extrathyroidal extension into the surrounding soft tissue. The left recurrent laryngeal nerve was involved by the mass and appeared nonfunctional. The tumor extended into the mediastinum, where it attached to the sternum and the subclavian vessels. Given the extent of the disease, a negative margin was deemed infeasible, and the tumor was removed with maximal attempt.

Gross pathologic examination showed an ill-defined multinodular mass, at least 4 cm in size, with bilateral lobe involvement. Microscopically, carcinoma cells formed irregular nests with intermingled glandular, papillary, micropapillary, and cribriform patterns, infiltrating an inflamed and fibrotic stroma (Fig. 1a). Tumor papillae and micropapillae within cystically dilated glands were associated with neutrophilic infiltrates and degenerative debris (Fig. 1b). There were scattered psammomatous calcifications (Fig. 1b inset) and rare foci of squamous metaplasia (Fig. 1c). Intracellular and extracellular eosinophilic secretions in globules and patches were seen (Fig. 1d). The tumor cells also demonstrated irregular nuclear contours, pale chromatin, and small nucleoli. Nuclear grooves were readily identified but pseudo-inclusions were not seen. Mitotic activity was low (up to one mitosis per 10 high-power fields). Small collections of amorphous debris suspicious for necrosis were seen in a few scattered compact cribriform/nearly solid tumor nests (Fig. 1e). There was metastatic carcinoma in two perithyroidal lymph nodes without extranodal extension.

Fig. 1.

Histologic features of thyroid secretory carcinoma. At low magnification, the tumor comprised infiltrative glands and nests in an inflamed and fibrotic stroma (a). Tumor papillae and micropapillae protruded into cystically dilated glands associated with inflammatory debris (b). Some micropapillary tumor florets showed psammomatous calcifications (b inset). In addition to background lymphoplasmacytic inflammation in the stroma, most tumor nests had mild to moderate neutrophilic infiltrates (c). Rare foci of squamous metaplasia were seen (c). The cribriform nests showed intracellular and extracellular eosinophilic secretions in globules and patches (d). Cytologically, the tumor cells demonstrated irregular nuclear contours, pale chromatin, and small nucleoli. Small foci of amorphous debris suspicious for necrosis were seen in a few compact cribriform/nearly solid tumor nests (e). Targeted RNA next-generation sequencing revealed an ETV6::NTRK3 gene fusion with the breakpoint localized to exon 4 and exon 14, respectively. Ample read support was visualized using the Integrative Genome Viewer [31] (f)

Immunophenotypically, the tumor showed diffuse expression of CK7 and GATA3 and variable expression of PAX8, S100, and mammaglobin (Fig. 2). The cribriform areas (Fig. 2a through f) showed weak to strong mammaglobin (Fig. 2d) and strong S100 (Fig. 2e) expression and were mostly negative for PAX8 (monoclonal, clone SP348, Fig. 2b). The papillary and micropapillary areas (Fig. 2g through l) showed widespread PAX8 nuclear expression (Fig. 2h) and patchy mammaglobin positivity (Fig. 2j) while exhibiting much attenuated S100 (Fig. 2k). Thyroglobulin was negative other than focal weak, nonspecific reactivity in some cribriform areas (Fig. 2c and i). Negative results were also seen with TTF1, BRAF V600E, SOX10, CDX2, synaptophysin, chromogranin, ER, PR, and WT1. Beta-catenin showed membranous labeling. TRK showed diffuse weak to moderate cytoplasmic immunoreactivity with rare nuclear staining (Fig. 2f and l).

Fig. 2.

Immunoprofile of the tumor in cribriform areas (a through f) and foci of papillary/ micropapillary growth (g through l), stained with PAX8 (b, h), thyroglobulin (TG; c, i), mammaglobin (MG; d, j), S100 (e, k), and TRK (f, l)

Molecular Analysis

Targeted amplicon-based next-generation sequencing (NGS) was performed using RNA extracted from formalin-fixed paraffin-embedded tissue slides to evaluate for rearrangements involving NTRK1, NTRK2, and NTRK3 genes. A list of covered regions and fusion partners with genomic coordinates based on build GRCh37 (hg19) is available at https://www.mayocliniclabs.com/test-catalog/overview/606377#Overview. An in-frame ETV6::NTRK3 rearrangement was detected with kinase domain preservation. The gene fusion breakpoint was localized to the exon 4 of ETV6 (NM_001987.4) and the exon 14 of NTRK3 (NM_001012338.2) (Fig. 1f).

Overall, the tumor was most compatible with primary thyroid secretory carcinoma. Based on the radiologic and intraoperative findings, the staging was most in keeping with T4b as defined by the American Joint Committee on Cancer 8th Edition staging criteria [13]. N stage was at least N1a based on histologically confirmed perithyroidal lymph node involvement, although more advanced nodal involvement was suspected clinically.

Postoperative Treatment

A PET scan performed postoperatively showed persistent infiltrative FDG-avid tumor in the left thyroidectomy bed, along with multifocal metastases, including upper mediastinal lymphadenopathy, bilateral adrenal nodules, and multifocal osseous and soft tissue metastases. A radiation therapy treatment plan was devised that utilized volume-modulated arc therapy with 6 MV photons to deliver 6600 cGy in 33 daily fractions to target gross residual thyroid tumor, involved sternum, and immediately adjacent lymph nodes with concurrent weekly cisplatin. Radiation therapy was initiated 7 days after CT simulation. Image guidance with the first fraction revealed critical tracheal narrowing, requiring urgent stent placement. The patient subsequently completed chemoradiotherapy without further complications but without apparent tumor response. Radioactive iodine was not rendered. She was subsequently started on larotrectinib, 100 mg twice daily. As of the most recent follow-up 6 months after the initial diagnosis and 2 months into the larotrectinib therapy, radiologic response was observed with decreased size of tumors in the thyroid bed (3.6 cm in the largest dimension, reduced from a pre-treatment size of 4.2 cm), bilateral adrenal nodules, and left back muscular metastasis. The patient had persistent left vocal cord paralysis with stridor but remained stable on ongoing larotrectinib treatment without major side effect.

Discussion

The thyroid gland is known to host unusual neoplasms with salivary gland-type differentiation, such as mucoepidermoid carcinoma, sclerosing mucoepidermoid carcinoma with eosinophilia, and secretory carcinoma [14]. Secretory carcinoma (SC) is the newest entity recognized by the World Health Organization classification of thyroid tumors published in 2022. Secretory differentiation is morphologically characterized by vacuolar cytoplasm associated with secretions and immunophenotypically supported by the expression of S100 and mammaglobin. On the intracellular signaling level, overexpression of the signal transducer and activator of transcription 5a (STAT5a) has been found specific to SC as opposed to other carcinoma types in the salivary gland and mammary tissues [15, 16]. STAT5a, also known as mammary growth factor, is physiologically activated by prolactin and expressed in cells undergoing secretory changes during lactation [15]. Recently, Kinnunen et al. showed that artificially expressed ETV6::NTRK3 fusions induced STAT5 phosphorylation in cell line models [17], providing a plausible linkage between the gene rearrangement and the secretory phenotype. Interestingly, the degree of phosphorylated STAT5 (pSTAT5) induction appeared to vary depending on the location of the ETV6::NTRK3 fusion breakpoint [17]. The most common fusion type seen in the thyroid (exon 4 to exon 14) demonstrated weaker pSTAT5 induction compared to the predominant breakpoint observed in salivary and breast SC (exon 5 to exon 15) [17]. This may potentially contribute to the rarity of SC among ETV6::NTRK3 fusion-related thyroid carcinomas, which more commonly exhibit the phenotype of PTC [10]. Cases of hybrid thyroid SC and PTC have been reported [4], in keeping with the previously proposed view that at least a subset of thyroid SC may be follicular cell-derived neoplasms with partial to complete secretory differentiation [14]. In the current case, we noted varying distributions of PAX8 and S100 in different tumor areas. PAX8 showed the strongest expression in foci of papillary growth that resembled PTC morphologically. However, concurrent patchy mammaglobin expression in these areas, along with negative TTF1 and thyroglobulin, argued against a co-existing PTC component. The growth patterns observed were intermingled and consistently demonstrated weak to moderate TRK positivity and negative BRAF V600E throughout the tumor, in keeping with the NGS finding of NTRK3 fusion [18].

The distinction between thyroid SC and PTC is therapeutically important. Recognizing SC (or an SC component) may prompt early consideration of molecular testing for targetable gene fusions, as most thyroid SC patients receiving radioactive iodine therapy have recurred (Table 1). A useful morphologic clue favoring SC is architectural variability with cribriform, papillary, and micropapillary patterns often seen in the same tumor. In cases with prominent papillary growth, it is helpful to recognize micropapillary formation with tumor cells arranged in floret-like clusters and the production of secretory material (Fig. 1b). Immunohistochemical characterization with S100, mammaglobin, and thyroid markers should be performed in tumors showing suggestive morphological features. Additionally, BRAF V600E immunostaining can also be diagnostically informative as the mutation accounts for the majority of PTC and is mutually exclusive with kinase gene fusions in untreated thyroid carcinomas [19]. When systemic treatment is in consideration for advanced thyroid carcinoma, as in the current case, negative BRAF V600E immunostaining has been recommended to trigger kinase fusion testing given the increased likelihood of fusion detection [10].

Due to concerns over the sensitivity of pan-TRK immunohistochemistry [20], the detection of NTRK rearrangements currently relies on molecular analysis, which has evolved from the earlier use of fluorescence in situ hybridization (FISH) and reverse transcription polymerase chain reaction (RT-PCR) to the recently increasing utilization of NGS. In the salivary gland, SC has been found harboring ETV6::NTRK3 fusions in the majority, while minor subsets have shown non-canonical rearrangements such as ETV6::MAML3, ETV6::MET, ETV6::RET, VIM::RET, and CTNNB1::ALK [21–24]. It is noteworthy that some of these non-canonical fusions could be missed by single-gene tests examining solely ETV6 or NTRK3. Although all the thyroid SC sequenced to this date have harbored ETV6::NTRK3, other oncogenic alterations remain possible for future cases. Fortunately, gene fusion testing has, in many laboratories, evolved into a panel approach covering multiple targetable kinase genes. This has laid the foundation for recent American Head and Neck Society expert recommendations that NGS-based panels are preferred to multiple single-gene tests in advanced thyroid cancers defined as bulky, RAI-recalcitrant, recurrent, or metastatic disease [25], which are common features of thyroid SC (Table 1).

Because of its low incidence, thyroid SC’s clinical behavior is poorly understood, and the prognostic significance of commonly considered grading parameters, such as nuclear atypia, mitosis, or necrosis, remains unexplored. Recently, Xu et al. proposed a grading system for SC of the salivary gland, with high-grade tumors defined as those with 5 or more mitoses per 10 high-power fields and/or necrosis, which correlated significantly with disease-free survival after adjusting for T stage and margin status [26]. It is unclear whether the same prognosticators may apply to thyroid SC. Among the published cases (Table 1), two cases of the same tumor size (4.0 cm) showed elevated mitotic activity (7/10 HPF and 10/10 HPF) [2, 6]. Over similar follow-up periods (12 and 14 months), one patient succumbed to tumor progression [2] while the other remained tumor-free [6]. Necrosis was seen in four patients whose clinical outcomes varied from apparent cure after 14 months [6] to tumor-related fatality at 20 days, 12 months, or 15 years after diagnosis [2, 8, 11]. More recently, Baněčková et al. proposed a three-tiered grading system for salivary gland SC based on the prevailing architecture, pleomorphism, tumor necrosis, perineural invasion, lymphovascular invasion, and mitotic activity or Ki-67 index [22]. The predictive power of these parameters in thyroid SC would be a worthy topic for future investigation. As it is difficult to draw conclusions based on the limited number of cases, pathologists’ awareness of this uncommon entity and astute differential diagnosis from PTC is crucial for identifying new cases for future analysis.

TRK inhibitor therapy has been a major achievement for oncologic science and the pharmaceutical industry over the past decade. Currently, larotrectinib and entrectinib are the two first-generation TRK inhibitors approved for treating adult and pediatric solid tumors carrying oncogenic NTRK1, NTRK2, or NTRK3 rearrangements in the United States. In a combined analysis of three clinical trials that included a wide variety of tumor types and primary sites with or without prior systemic therapy, the overall objective response rate (ORR) to larotrectinib was 79%, with a median duration of response of 35.2 months [27]. The efficacy of entrectinib was recently evaluated in 121 patients diagnosed with NTRK-rearranged tumors of 14 different types from three phase I/II clinical trials [12]. The ORR was 61.2%, with a median duration of response of 20.0 months [12]. Both analyses included thyroid tumors (16% [27] and 10.7% [12]) but the histology of non-PTC cases, if included, was not further specified. In patients receiving TRK inhibitors, acquired resistance may arise due to on-target mutations that alter the steric conformation of the NTRK molecule [28]. Two second-generation TRK inhibitors, selitrectinib and repotrectinib, have shown promising preliminary results in overcoming on-target resistance [29, 30] and are currently in clinical trials (NCT03093116, NCT04094610, NCT03215511).

To summarize, primary thyroid SC is a kinase fusion-driven malignancy that is possibly under-recognized due to its rarity and morphologic overlap with PTC, and yet a timely diagnosis is important given the observed propensity for early metastatic spread and molecular actionability. Misclassification as PTC is a common diagnostic pitfall that can be avoided by recognizing the morphologic characteristics (cribriform and micropapillary architecture and eosinophilic secretions) and applying immunohistochemistry, especially in cases presenting with advanced-stage disease at initial diagnosis. Molecular testing is the current mainstay for qualifying patients for TRK inhibitor clinical trials and treatments. While TRK-focused testing currently seems sufficient as all the tested thyroid SC have shown ETV6::NTRK3 to date, a panel-based approach might be advantageous for identifying potential noncanonical fusions in future cases or for characterizing resistance later in the treatment course. Larger studies are needed to provide more evidence for developing treatment and prognostication strategies to best guide patient management.

Author Contributions

YHC and JM conceptualized the case report and initiated the project (JM as the primary pathologist; YHC reviewed the pathology on a consultative basis). YHC, DS, and JM collaboratively wrote the manuscript including revisions. ADB provided clinical insights from a surgical perspective. All authors reviewed the manuscript.

Funding

No research funding was acquired or used.

Data Availability

No datasets were generated or analysed during the current study.

Code availability

Not applicable.

Declarations

Ethical Approval

This case report was performed with ethics approval from the Mayo Clinic Institutional Review Board (IRB application number 23-011312).

Consent for Publication

Consent for publication was obtained from the patient included in the study.

Informed Consent

Informed consent was obtained from the patient included in the study.

Competing Interests

The authors declare no competing interests.