Metastasis to the thyroid gland: a single institution 16-year experience

Abstract

Aims:

Metastasis to the thyroid gland is a rare occurrence that may pose a diagnostic challenge. In this study, we aimed to report the clinicopathologic features, immunoprofile, molecular alterations, and outcome of 30 patients treated at our center from 2003 to 2019.

Results:

The most common site of the primary tumor was the kidney, followed by the lung, lower gastrointestinal tract, and breast. In seven (23%) patients, the thyroid metastases were resected prior to the diagnosis of the primary tumors. Six patients (20%) had thyroid as the sole metastatic site. Three (10%) patients harbored tumor-to-tumor metastasis. 71% had unilateral unifocal thyroid mass, which might be mistaken for primary thyroid tumors. Among the 13 cases that were initially diagnosed at an outside hospital, 4 (31%) were misinterpreted as a thyroid primary. An immunohistochemical panel of thyroid follicular cell markers was most useful to differentiate primary thyroid tumors from metastasis. Molecularly, the metastasis showed alterations characteristic of the primary tumor, which may be helpful in establishing diagnosis and primary site. Although the prognosis was poor with a 5-year disease specific survival of 58%, a long-term cure was possible in cases with oligometastasis treated successfully with surgery.

Conclusions:

Metastasis to the thyroid gland is an uncommon phenomenon with an incidence of 0.36% in all thyroid malignancies. It may present as a solitary thyroid mass before the discovery of the primary tumor, posing a diagnostic challenge. Although the overall prognosis is poor, a subset of patients with oligometastasis can be managed surgically.

Introduction

Metastasis to the thyroid gland is an uncommon occurrence with a reported frequency of up to 2.1% among all thyroid malignancies diagnosed in surgical specimens ^1-3. The most common primary site is the lung in autopsy series and kidney in surgical specimens ⁴. Given its rarity, metastasis to the thyroid may pose a diagnostic challenge especially when the primary tumor is undiagnosed at the time of thyroid sampling, resulting in misinterpretation of metastasis as a primary thyroid carcinoma ^5-7.

In recent years, immunohistochemistry (IHC) of various thyroid-specific markers (e.g. PAX8, TTF-1 and thyroglobulin) ^8-12 as well as molecular diagnostic tools ¹³ became readily available in routine clinical practice, assisting pathologists in reaching a correct diagnosis in challenging cases. However, publications on thyroid metastasis, especially those including comprehensive immunophenotypic and molecular profiles, are relatively sparse.

In this study, we performed a detailed clinical, pathologic, immunophenotypic, and molecular analysis of 30 cases of metastasis to the thyroid gland in patients seen at the Memorial Sloan Kettering Cancer Center (MSKCC, New York, NY, US) spanning a period of 16 years. Additionally, a literature review on the contemporary studies of metastasis to the thyroid gland in surgical pathology within the past 4 decades was conducted, aiming to further our understanding of the diagnosis and management of this rare entity.

Materials and Methods

The study was approved by the Institutional Review Board of MSKCC. The pathology database was searched and reviewed for all surgical thyroid specimens from 2003 to 2019 with a diagnosis of malignancy (including carcinoma, lymphoma, and sarcoma) or noninvasive follicular thyroid neoplasm with papillary-like nuclear features (NIFTP). Thyroid specimens with a pathologic diagnosis of non-neoplastic conditions or adenoma were excluded. A total of 8410 cases diagnosed with malignancy involving the thyroid gland were identified. Among them, 30 patients harbored hematogenous metastasis to the thyroid gland. Cases with secondary involvement of the thyroid gland by direct extension of carcinoma or sarcoma from adjacent organs (e.g. larynx, trachea, and parathyroid gland) were excluded.

All available slides and clinical charts of the 30 cases of thyroid metastases were reviewed by an endocrine pathologist (BX). Pathologic features gathered included specimen type (biopsy vs. resection), diagnosis of the metastasis, size, focality, and laterality of the metastasis, co-existing thyroid neoplasms, and their relationship with the metastasis, vascular invasion, immunohistochemical profile, and molecular alterations. Clinical characteristics included were age, sex, site and pathologic diagnosis of the primary tumor and its temporal relationship to the thyroid metastasis, other sites of distant metastasis, and clinical follow up. The 3-year, 5-year, and 10-year disease specific survival was calculated from the diagnosis of thyroid metastasis using SPSS software 24.0 (IBM corporation, Westchester, NY, USA).

IHC was performed at the discretion of the pathologist at the time of diagnosis. Antibodies utilized in this study included PAX8 (Proteintech, polyclonal, dilution 1:100), TTF-1 (Ventana, 8G7G3/1, monoclonal, ready to use RTU), thyroglobulin (Cell Marque, 2H11+6E1, RTU), Napsin-A (Leica, IP64, monoclonal, dilution 1:100), CD10 (Leica, 56C6, monoclonal, RTU), carbonic anhydrase IX (CAIX, Novus, polyclonal, dilution: 1:2000), synaptophysin (Biogenex, SNP88, monoclonal, dilution 1: 2000), chromogranin (Ventana, LK2H10, monoclonal, RTU), CDX2 (Biogenex, CDX2.88, monoclonal, dilution 1:100), GATA3 (Biocare, L50.823, monoclonal, dilution 1:200), estrogen receptor (ER, Leica, 6F11, monoclonal, RTU), progesterone receptor (PR, Leica, 16, monoclonal, RTU) , Histone H3-G34W (HH3-G34W, RevMab Bioscience, RM263, monoclonal, RTU), smooth muscle actin (SMA, Cell Marque, 1A4, Monoclonal, dilution 1:200), smooth muscle myosin (SMM, Cell Marque, SMMS-1, monoclonal, RTU), CKIT (DAKO, polyclonal, dilution 1:150), CK7 (DAKO, OV-TL-12/30, monoclonal, dilution 1: 800), and anaplastic lymphoma (ALK, Cell Signaling, D5F3, Monoclonal, dilution 1:200).

Molecular profile, when available, was collected from the patient’s chart. Five pulmonary carcinomas were subjected to Sequenom MassARRAY system (Sequenom, San Diego, CA) targeting 8 oncogenes: EGFR, KRAS, BRAF, PIK3CA, NRAS, AKT1, ERBB2, and MAP2K1/MEK as previously described ^{14, 15}. Additionally, 6 tumors, including a pulmonary adenocarcinoma, a pulmonary non-small cell carcinoma, a clear cell renal cell carcinoma, a malignant giant cell tumor, a polymorphous adenocarcinoma of minor salivary gland origin, and an undifferentiated spindle sarcoma, were tested using MSK-IMPACT™ (Memorial Sloan Kettering-Integrated Mutation Profiling of Actionable Cancer Targets) platform, which is a Food and Drug Administration (FDA)-approved deep-coverage, targeted next-generation sequencing assay detecting single nucleotide variants (SNVs), small insertions/deletion (indels), copy number variants (CNVs) and fusion/structural variants in 468 oncogenes, using custom DNA probes designed for targeted sequencing of all exons and selected introns, including canonical and selected non-canonical transcripts ^{16, 17}.

Results

Among the 8410 cases of thyroid biopsies and resections with a diagnosis of malignancy or NIFTP, 30 patients (0.36%) were diagnosed with metastasis to the thyroid gland. The detailed clinical, pathologic, immunophenotypic, and molecular features of the study cohort are shown in Table 1.

Table 1.

Clinicopathologic features and outcomes of metastases to the thyroid gland.

N	Age/ gender	Primary site	Diagnosis	Time interval from primary tumor diagnosis to thyroid metastasis (months ) (a)	Other sites of metastasis	Thyroid as the only site of metastasis	Thyroid as first site of metastasis	Status at last follow up	Follow up period after diagnosis of thyroid metastasis (months )	Specimen type	Laterality and focality of metastasis	Tumor greatest dimension (cm)	Thyroid neoplasms and geographic relationship with the metastasis	Discrepant initial diagnosis
Kidney primary (n=9)
1	65/F	Kidney	Clear cell RCC	−2	Lung	No	Yes	AWD	9	Resection	Unilateral, unifocal	1.1	NIFTP (tumor-to-tumor metastasis)	No
2	59/F	Kidney	Clear cell RCC	−1.5	None	Yes	Yes	NED	108.9	Resection	Bilateral, multifocal	1.8	PMC (not related)	NA
3	62/F	Kidney	Clear cell RCC	92.8	None	Yes	Yes	AWD	45.9	Resection	Unilateral, unifocal	2.5	PMC (not related)	No
4	83/M	Kidney	Clear cell RCC	6.3	Liver	No	Yes	DOD	35.4	Resection	Unilateral, unifocal	1.3	PMC (not related)	NA
5	63/M	Kidney	Clear cell RCC	97.7	skin, bone, brain, lung, spine	No	No	DOD	33.4	Resection	Bilateral, multifocal	5.5	Absent	NA
6	69/M	Kidney	Clear cell RCC	129.5	None	Yes	Yes	NED	61.3	Resection	Bilateral, multifocal	2.9	Absent	No
7	72/M	Kidney	Clear cell RCC	71.7	None	Yes	Yes	NED	35.4	Resection	Unilateral, unifocal	2.4	Absent	NA
8	60/M	Kidney	Clear cell RCC	61.1	Pancreas, lung	No	No	DOD	124.5	Resection	Unilateral, unifocal	2.5	Absent	NA
9	71/F	Kidney	Chromophobe RCC	55.7	Liver	No	Yes	AWD	3.4	Resection	Unilateral, unifocal	1.8	PTC, FV (adjacent)	PTC, FV
Lung primary (n=7)
10	83/F	Lung	Adenocarcinoma	36.8	Bone, spine	No	No	AWD	21.2	Resection	Unilateral, unifocal	1.4	Absent	NA
11	44/F	Lung	Adenocarcinoma	1.1	Bone, liver	No	Yes	DOD	16.4	Resection	Bilateral, multifocal	2.5	Absent	No
12	63/M	Lung	Adenocarcinoma	NA	Brain, intrapulmonary, cervical lymph nodes	No	Yes	AWD	14	Resection	Unilateral, unifocal	3	PMC (adjacent)	ATC
13	28/M	Lung	Adenocarcinoma	NA	Bone, gastrointestinal tract, peritoneum	No	Yes	DOD	6.3	Resection	Unilateral, unifocal	6.2	Absent	NA
14	77/M	Lung	Non-small cell carcinoma	NA	Brain, bone, intrapulmonary, peritoneal	No	Yes	AWD	18.2	Resection	Unilateral, unifocal	6	Absent	ATC
15	65/M	Lung	LCNEC	8.7	Intrapulmonary	No	No	AWD	63.9	Resection	Unilateral, unifocal	7.2	Absent	NA
16	54/M	Lung	Carcinoid tumor	34.7	Liver, intrapulmonary	No	No	DOD	88.2	Resection	Unilateral, unifocal	0.5	PMC (not related)	NA
Colorectal primary (n=4)
17	66/F	Colorectal	Adenocarcinoma	145	Ovary, lung, duodenum	No	No	AWD	52.1	Resection	Unilateral, unifocal	1.2	Classic PTC (tumor-to-tumor metastasis)	NA
18	49/F	Colorectal	Adenocarcinoma	67.8	Liver	No	No	NED	49.4	Resection	Bilateral, multifocal	0.8	PMC (not related)	No
19	63/M	Colorectal	Adenocarcinoma	23.1	Liver, omentum, abdominal wall	No	No	DOD	55.4	Resection	Unilateral, unifocal	2.2	Absent	NA
20	56/M	Colorectal	Adenocarcinoma	124.3	Brain, lung	No	No	DOD	106.2	Resection	Unilateral, unifocal	4	Absent	NA
21	56/F	Colorectal	Adenocarcinoma	71	Lung, liver	No	No	DOD	1	Biopsy	NA	NA	NA	NA
Other primary sites (n=9)
22	59/F	Breast	Mammary carcinoma	161.5	Bone	No	No	AWD	11.2	Resection	Bilateral, multifocal	1.5	PTC, TCV (adjacent)	No
23	58/F	Breast	Mammary carcinoma	76.1	Cervical lymph nodes	No	Yes	NED	9.6	Resection	Bilateral, multifocal	1	Absent	No
24	51/F	Salivary gland (minor, buccal)	PAC	257.4	Mediastinum, brain, spine	No	Yes	AWD	43.4	Resection	Unilateral, unifocal	2	Absent	No
25	78/F	Salivary gland (submandibular)	Adenocarcinoma, NOS	24.6	Lung, bone	No	Yes	DOD	29.1	Resection	Unilateral, unifocal	2	Absent	No
26	72/F	Esophagus	Adenocarcinoma	2.2	None	Yes	Yes	DOUC/NED	123.9	Resection	Unilateral, unifocal	3.4	FA (tumor-to-tumor metastasis)	NA
27	51/F	Bone (femur)	Malignant giant cell tumor	NA	Bone, spine, lung	No	No	DOD	3	Biopsy	NA	NA	NA	NA
28	22/M	Bone (femur)	Osteosarcoma	78.3	Lung, liver	No	No	DOD	12.4	Resection	Unilateral, unifocal	3.9	Absent	NA
29	47/M	Soft tissue (peralaryngeal)	Undifferentiated spindle sarcoma	46.2	Skin, neck soft tissue, spine, lung, pleura	No	No	DOD	45.8	Resection	Unilateral, unifocal	5	Absent	NA
30	81/M	Soft tissue (retroperitonum)	Leiomyosarcoma	NA	None	Yes	Yes	AWD	13.4	Resection	Bilateral, multifocal	6.5	Absent	ATC

F: female, M: male, PAC: polymorphous adenocarcinoma, RCC: renal cell carcinoma, LCNEC: large cell neuroendocrine carcinoma, NOS: not otherwise specified, AWD: alive with disease, NED: no evidence of disease, DOD: dead of disease, DOUC: dead of unknown cause, NED: no evidence of disease, PTC: papillary thyroid carcinoma, TCV: tall cell variant, FV: follicular variant, PMC: papillary microcarcinoma, NIFTP: noninvasive follicular thyroid neoplasm with papillary-like nuclear features, FA: follicular adenoma.

^(a)A negative value indicates that the thyroid metastasis is diagnosed prior to primary tumor. NA: not applicable in four cases in which the primary tumor was not sampled for pathologic diagnosis.

Clinical characteristics

The median age at the diagnosis of thyroid metastasis was 63 years (range: 23 to 83). The male to female ratio was 1:1.

In 23 (77%) cases, the primary tumors were diagnosed prior to the presentation of thyroid metastasis. The median time interval from the pathologic diagnosis of the primary tumor to thyroid metastasis was 64.5 months (range: 1 to 257 months). Late metastasis (10 years or more) to the thyroid gland was observed in 5 patients, including 2 with colorectal adenocarcinoma, 1 with clear cell renal cell carcinoma, 1 with mammary carcinoma, and 1 with salivary gland carcinoma. In 2 cases, the primary tumor was resected to provide a definite pathologic diagnosis approximately 2 months after the diagnosis of thyroid metastasis, whereas in the remaining 5 patients the primary tumors were never sampled.

Among these 5 patients, the site and diagnosis of primary tumors were determined using a combination of clinical, radiologic, immunophenotypic, and molecular findings. Three patients (cases 12, 13, and 14) presented with large pulmonary mass and widely metastatic disease with the thyroid metastasis demonstrated typical immunophenotypic (TTF-1 positive/napsin-A positive/PAX8 negative) and/or molecular features (ALK fusion or KRAS mutation) of lung carcinoma; one patient (case 27) had large femoral mass with multifocal metastasis to lung and other bones showing diagnostic H3F3A mutation of giant cell tumor, and one (case 30) showed 8-cm unresectable stable retroperitoneal mass and multiple bilateral thyroid nodules that were diffusely and strongly positive for myoid marker smooth muscle actin and myosin.

Six patients (20%) had thyroid as the sole metastatic site, including 4 patients with clear cell renal cell carcinoma, 1 with retroperitoneal leiomyosarcoma, and 1 with esophageal adenocarcinoma. The remaining 24 patients had widely metastatic disease involving other organs, most common being lung (n=13), bone (n=8), liver (n=8), and brain (n=5). Among these 24 patients, the thyroid gland was present as the first site of metastasis in 10 patients.

The histologic features

Twenty-five cases had a diagnosis of carcinoma; one case was diagnosed as metastatic carcinoid tumor; whereas the remaining four were mesenchymal neoplasms. Two cases were biopsies, while the remaining 28 cases were resection specimens. The most common type and site of the primary tumor that metastasized to the thyroid gland were clear cell renal cell carcinoma (n=8, 27%) and the kidney (n=9, 30%) respectively. The sites of primary tumors in a descending order were kidney (n=9), lung (n=7), colon/rectum (n=5), breast (n=2), femur (n=2), buccal minor salivary gland (n=1), submandibular gland (n=1), esophagus (n=1), neck soft tissue (n=1), and retroperitoneum (n=1).

The histotypes of the metastasis were clear cell renal cell carcinoma (n=8), chromophobe renal cell carcinoma (n=1), pulmonary adenocarcinoma (n=4), pulmonary non-small cell carcinoma (n=1), pulmonary large cell neuroendocrine carcinoma (n=1), pulmonary carcinoid tumor (n=1), colorectal adenocarcinoma (n=5), mammary carcinoma (n=2), polymorphous adenocarcinoma of salivary gland (n=1), adenocarcinoma not otherwise specified of salivary gland (n=1), esophageal adenocarcinoma (n=1), malignant giant cell tumor (n=1), osteosarcoma (n=1), leiomyosarcoma (n=1), and undifferentiated spindle sarcoma (n=1).

Among the 28 resected metastases, 8 (29%) contained bilateral and multifocal disease in the thyroid gland, while the remaining 20 (71%) were present as a unilateral and unifocal thyroid mass. Vascular invasion was identified in 5 (18%), including 2 with extensive vascular invasion. The median size of the metastatic focus was 2.5 cm (range: 0.5 to 7.2 cm).

Eleven (39%) thyroid resection specimens also contained neoplasms of thyroid follicular cell origin, including papillary microcarcinoma (n=6), NIFTP (n=1), papillary carcinoma follicular variant (n=1), noninvasive encapsulated classic papillary carcinoma with predominant follicular pattern (n=1), papillary carcinoma tall cell variant (n=1), and follicular adenoma (n=1). Among them, 3 showed tumor-to-tumor metastasis (Figure 1), including an esophageal adenocarcinoma to follicular adenoma, a colorectal adenocarcinoma to classic papillary carcinoma, and a clear cell renal cell carcinoma to NIFTP. Two cases had coincidental papillary carcinoma immediately adjacent to the metastasis, whereas in the remaining cases, the thyroid neoplasms and the metastases showed no spatial relationship.

Figure 1. Tumor-to-tumor metastasis.

A colorectal adenocarcinoma (C) metastasizes to an encapsulated classic papillary thyroid carcinoma with predominant follicular growth pattern (P, patient #17). The metastatic colorectal adenocarcinoma contains pseudostratified nuclei, luminal necrosis, and frequent mitotic activity. Note the presence of scattered well-formed papillae within the background thyroid lesion (arrows and insert). Magnification: top panel 20X; bottom panel 200X.

IHC and molecular profile

Among the 22 cases tested for TTF-1 IHC, 5 (23%) were positive for TTF-1, all of which were of pulmonary origin. PAX8 IHC using polyclonal antibody was positive in 4 of 13 (31%) cases, including 2/2 (100%) clear cell renal cell carcinoma, and 2/6 (33%) pulmonary tumors. Thyroglobulin IHC was universally negative in all metastases tested.

Molecular testing was performed in 11 cases, which demonstrated alterations characteristic of the primary tumors in 5 cases, including VHL mutation in clear cell renal cell carcinoma (n=1), KRAS mutations (n=2) or EML4-ALK fusion (n=1) in pulmonary adenocarcinoma, and H3F3A G34W mutation in malignant giant cell tumor.

Diagnostic discrepancy and challenges

Thirteen cases were consultation cases that were initially diagnosed at an outside institute. Among them, 4 cases were misinterpreted as primary thyroid carcinomas. The discrepant diagnoses were as follows (initial – revised): anaplastic thyroid carcinoma (ATC) – metastatic leiomyosarcoma; ATC – metastatic non-small cell lung carcinoma; ATC – metastatic EML4-ALK translocated pulmonary adenocarcinoma; and papillary thyroid carcinoma follicular variant – metastatic chromophobe renal cell carcinoma. The primary tumors were not diagnosed at the time of thyroid metastasis in the first three cases, while the resection of chromophobe renal cell carcinoma was performed 5 years prior to the diagnosis of thyroid metastasis. The two cases of pulmonary metastasis showed an immunoprofile of TTF-1 positivity, napsin-A positivity, and PAX8 negativity, which was incompatible with a diagnosis of ATC (Figure 2). The metastatic chromophobe renal cell carcinoma was positive for PAX8, CK7, and CKIT, whereas negative for TTF-1 and thyroglobulin, which did not support the diagnosis of a well-differentiated thyroid carcinoma (Figure 3).

Figure 2. Metastatic pulmonary adenocarcinoma (P) to the thyroid gland (T),

hematoxylin and eosin (H&E, A) and immunohistochemistry for thyroglobulin (TG, B), napsin-A (C), TTF-1 (D), and PAX8 (E, Case #12). The metastatic pulmonary adenocarcinoma is positive for napsin-A and TTF-1, whereas negative for TG and PAX8. The background thyroid gland is positive for PAX8, TTF-1, and TG, and is negative for napsin-A.

Figure 3. Metastatic chromophobe renal cell carcinoma to the thyroid gland.

(A) At low power (100X), the tumor is separated from the background thyroid (blue arrow) by a fibrous capsule. (B) At high power (400X), the tumor cells have wrinkled nuclei, binucleation, perinuclear halo, and abundant eosinophilic cytoplasm. (C-F) The tumor shows a PAX8-positive, TTF-1-negative, CK7-positive, and CKIT-negative immunoprofile. Note that the background thyroid is TTF-1 positive (blue arrow in panel D).

Clinical outcome

The median follow up of the study cohort after the diagnosis of thyroid metastasis was 35 months (range: 1 to 200 months). The median disease-specific survival from the diagnosis of thyroid metastasis was 94 months. The 3-year, 5-year, and 10-year disease specific survival after metastasis was 70%, 58%, and 39% respectively.

Six cases (20%) were considered as no evidence of disease (NED) given that both the primary tumors and metastases were completely resected surgically; 13 suffered disease specific death; whereas the remaining cases were alive with disease at the time of the last follow up. Among the 6 patients with a status of NED at the last follow up, four patients had metastasis limited to the thyroid gland (3 with clear cell renal cell carcinoma and 1 with esophageal adenocarcinoma), while the remaining two patients had resectable oligometastasis to the liver or cervical lymph node beside the thyroid gland (one with mammary carcinoma, the other with colorectal adenocarcinoma). The 3-year, 5-year, and 10-year disease specific survival was 100%, 100%, and 100% respectively for the subgroup with isolated metastasis limited to the thyroid gland, and 56%, 41%, and 14% for patients with multi-site/widely-spread metastatic disease.

Discussion

A detailed literature review summarizing studies of metastasis to the thyroid gland in surgical pathology from 1985 to present is provided in supplementary table 1. Autopsy series ^18-20, fine needle aspiration reports ^{21, 22}, and surgical studies conducted before 1985 ^{5, 7, 18, 23-25} were excluded to generate a contemporary dataset with supportive diagnostic ancillary studies used in current surgical pathology practice. Two series included lymphoma ²⁶ and leukemia ²⁷ as metastasis: such cases were considered as systemic involvement and were excluded from the literature review. Overall, metastasis to the thyroid gland was exceedingly rare, accounting for 0.15% of all thyroid resections ^{1, 2, 28}, and 0.63% of all thyroid malignancies ^{1, 2}. Of note, the frequency of metastasis among all thyroid malignancies was lower in the present study, being 0.36%, compared with 1.1% to 2.1% as reported in the literature ^1-3. One explanation for the discrepancy could be the selection bias of patients treated at our center with high frequency of malignancy diagnosis in all thyroid pathology specimens and a high rate of surgical resection in patients with metastasis to the thyroid gland. Thyroidectomy is often reserved for those with oligometastasis as reflected in the higher number of patients with metastatic tumors limited to the thyroid (20%) in our study compared to other surgical series ^{29, 30}.

In our cohort, renal cell carcinoma was the most common histotype of thyroid metastasis, followed by pulmonary tumors, colorectal adenocarcinoma, and mesenchymal tumors. Similarly, the literature review showed that the kidney was the most common primary site, accounting for 42% of all cases ^{1, 2, 28-39}. Other common sites of primary tumors in descending order of frequency were lung (16%), head and neck (10%), breast (9%), lower gastrointestinal tract (7%), upper gastrointestinal tract (5%), melanoma (2%), and sarcoma (2%). However, the rate of head and neck carcinoma metastasis to the thyroid gland needs to be interpreted with caution, as several studies included thyroid involvement from carcinomas of the adjacent head and neck organs, such as trachea, larynx, and parathyroid gland ^{2, 31, 33, 39}. It is questionable whether these cases represent a direct extension of the primary tumor into the thyroid gland given the anatomic proximity.

To date, all renal cell carcinomas metastasized to the thyroid gland reported in the literature were described as of clear cell type ^{1, 12, 35, 40}. In this study, we reported the first case of hematogenous spread of chromophobe renal cell carcinoma to the thyroid gland, widening the differential diagnoses in this specific setting.

Contrary to the common belief that thyroid metastases frequently present as multifocal deposits ⁴¹, we found that the majority (71%) manifested as a unifocal and unilateral distinct nodule within the thyroid gland. Additionally, 39% of resection specimens contained primary thyroid neoplasms as well, including scenarios of tumor-to-tumor metastasis or adjacent primary and metastatic tumors. Furthermore, in two cases, the primary tumor was discovered after the thyroid resection. Together, these factors resulted in the possibility of a metastasis masquerading as a primary thyroid tumor, which presented diagnostic challenges to the practicing pathologists. Indeed, in our study, 36% (4/11) of the consult cases were initially misinterpreted as primary thyroid carcinomas, in particular ATC and less frequently papillary thyroid carcinoma. In this setting, a high clinical suspicion and a combined IHC panel of thyroid-specific markers (in particular TTF-1, thyroglobulin, and PAX8), as well as primary site/tumor-specific markers, might lead to the correct diagnosis ^8-11. Among the thyroid-specific markers, PAX8 is particularly useful in the setting of ATC as a large proportion of it is positive for PAX8 and entirely negative for TTF-1 and thyroglobulin ^{8, 11}, while pulmonary carcinomas are in their vast majority typically PAX8 negative. However, as shown in this study and others, lung carcinomas can express PAX8 if a polyclonal antibody is used ⁴². Some authors found that 55 (8%) of 687 lung carcinomas are PAX8 positive using a polyclonal antibody while no immunopositivity was seen in these lung carcinomas stained with monoclonal PAX8 ⁴². We therefore recommend the use of monoclonal PAX8 when metastasis is suspected especially a lung metastasis. At the other end of the spectrum, thyroglobulin IHC is the most specific marker and is universally expressed in differentiated thyroid carcinoma ^{10, 43}. We however encountered in our practice non-thyroid carcinomas focally positive for thyroglobulin because of cross-reactivity (especially with Müllerian primaries) or passive diffusion from normal thyroid follicles, a well-known phenomenon ⁴⁴. The confusion of a metastatic tumor to the thyroid with a primary thyroid carcinoma can have significant clinical impacts on the patient leading to morbidities. In our opinion, the best approach to avoid such a mistake is to use a combination of immunohistochemical markers along with a careful clinical-pathologic correlation.

Among all pathologic features studied, tumor-to-tumor metastasis deserves a special mention. It is a rare phenomenon defined as metastasis into another tumor ^45-48. Thyroid tumors are one of the most common “recipient” tumors ⁴⁷, and the most common histotypes reported in the literature are follicular adenoma and follicular variant of papillary thyroid carcinoma ^45-48. In this study, we reported three cases of tumor-to-tumor metastasis with the “recipient” tumors being a follicular adenoma, a noninvasive encapsulated classic papillary carcinoma with predominant follicular growth pattern, and a NIFTP. Coincidentally, all three tumors are indolent neoplasms with a predominant or exclusive follicular growth pattern.

Molecular testing, when performed, might show signature alterations of the primary tumors, which might be useful in identifying the primary site in an appropriate clinical setting, e.g. widely metastatic disease without a clear-cut primary. The characteristic molecular alterations reported in our cohort included VHL somatic mutation in clear cell renal cell carcinoma ⁴⁹, KRAS mutations, or EML4-ALK translocation in pulmonary adenocarcinoma ⁵⁰, and H3F3A G34W mutation and immunopositivity of the mutation product HH3-G34W in giant cell tumor ^{51, 52}.

Literature varied regarding the outcome of thyroid metastasis. The reported median overall survival ranged from 12 to 34 months ^{2, 29, 37}, 5-year overall survival ranged from 31% to 51% ^{28, 35, 36}, and the rate of disease-specific mortality ranged from 33% to 73% ^{1, 28, 35, 36, 38, 39}. We herein reported a median disease specific survival of 94 months, 5-year overall survival of 58%, and disease-specific mortality of 43%. The prolonged survival in our cohort might in part be attributed to the finding of a small but significant amount of cases (20%) with thyroid being the only metastatic site. Indeed as mentioned above, the proportion of patients with metastasis restricted to the thyroid was much lower in other series ^{29, 30}. A subgroup of patients (6/30, 20%) was considered clinically as disease-free following a curative surgical attempt to remove oligometastasis. Similarly, a recent meta-analysis has found surgical removal of the metastasis to the thyroid gland was associated with improved survival compared to a conservative approach in all metastasis as well as in subsite analysis focusing specifically on metastases from kidney, colon, or pulmonary origin ³⁰. Together, these data suggest that surgical approach with curative intent may be considered in the setting of oligometastasis to the thyroid gland, which may result in improved survival and even long-term cure.

In conclusion, we herein reported the comprehensive clinical, pathologic, immunophenotypic, and molecular features of 30 cases of metastasis to the thyroid gland, a rare entity accounting for less than 1% of all thyroid malignancies. Thyroid metastasis typically is presented as a unifocal thyroid mass and might occur prior to the discovery of the primary tumor leading to its misinterpretation as a thyroid primary in a significant number of cases. This error has drastic consequences for the patient. Immunostains for TTF-1, thyroglobulin, and PAX8, as well as molecular testing are useful diagnostic tools in the appropriate clinical setting that can prevent this serious misdiagnosis. Although the outcome was poor overall with a 5-year disease-specific survival of 58%, long-term remission might be achieved with surgery for patients with oligometastasis.