Management and Prognosis of Metastases to the Thyroid Gland

Eric Mirallié, MD, Jérôme Rigaud, MD, Muriel Mathonnet, MD, PhD, Hélène Gibelin, MD, Nicolas Regenet, MD, Antoine Hamy, MD, Frédéric Bretagnol, MD, Loïc de Calan, MD, Jean-Claude Le Néel, MD, Jean-Louis Kraimps, MD

BACKGROUND: Intrathyroid metastases (ITM) of extrathyroid cancer are rare and have a poor prognosis. The aim of this work was to identify the sites of primary cancers and the treatment and prognosis of these lesions. STUDY DESIGN: This retrospective study was carried out on patients treated between 1982 and 2002 in the surgical departments of the University hospitals of Poitiers, Limoges, Tours, and Nantes, France. All diagnoses were confirmed by cytology or histologic examination. RESULTS: Twenty-nine patients (41 to 78 years) had ITM. Primary cancers were renal cell in 16 patients, lung in 4, digestive in 4, sarcoma in 1, melanoma in 1, neuroendocrine in 1, and of unknown origin in 1 patient. For 10 patients, diagnoses of primary cancer and ITM were synchronous. For 19 patients, delay between diagnosis of the primary cancer and ITM was 6.8 years (2 months to 16 years). Diagnosis was confirmed with fine-needle aspiration 3 times and with histologic examination of the thyroid 26 times. Twenty-seven patients had thyroidectomy; two were not operated on. After treatment of ITM, 13 patients had new metastatic sites. Mean followup for all patients was 2.3 years. Seven patients (24%) (6 with renal cancer) were disease free (followup 4.5 years). Four patients were alive with disease (followup 1.4 years). Eighteen patients (62%) died of their disease at a mean delay of 1.4 years. CONCLUSIONS: ITM are rare but the diagnosis should be borne in mind when patients have a history of cancer (mainly renal cancer). Preoperative diagnosis and complete evaluation could avoid unnecessary thyroidectomy because numerous patients had diffuse metastases. (J Am Coll Surg 2005;200: 203–207. © 2005 by the American College of Surgeons)

Clinical series of intrathyroid metastases (ITM) of extra- incidence of metastatic disease in the thyroid in patients thyroid cancers are rare and samples limited.1-9 ITM who die as a result of primary or metastatic cancer.10-13 represent between 1.4% and 2.5% of thyroid cancers3,9 The most frequently reported primary cancers are clear and between 0.05% and 0.1% of thyroid disease.2,8 cell renal carcinoma, bronchial carcinoma, and breast Conversely, the incidence of ITM is higher in autopsy cancer.1,3,7 series. Autopsy series demonstrate a 1.25% to 24.4% The aim of our study was to record ITM to identify the modes of revelation, location of primary cancers, treatment, and prognosis of the ITM. No competing interests declared. Received June 18, 2004; Revised October 4, 2004; Accepted October 4, METHODS 2004. From the Department of Surgery, Clinique Chirurgicale A, Hôtel Dieu, Uni- Our retrospective study focused on patients presenting versity Hospital Nantes, Nantes, France (Mirallié, Rigaud, Le Néel); Depart- an ITM who were treated between 1982 and 2002 in the ment of Surgery, Limoges Hospital, Limoges, France (Mathonnet); Depart- ment of Surgery, Poitiers Hospital, Poitiers, France (Gibelin, Kraimps); Surgical Departments of the University Hospitals of Department of Surgery, Clinique Chirurgicale II, Hôtel Dieu, University Tours, Limoges, Nantes, and Poitiers, France. All pa- Hospital Nantes, Nantes, France (Regenet); Department of Surgery, Clinique Chirurgicale I, Laënnec Hospital, University Hospital Nantes, Nantes, France tients had confirmed cytologic or histologic diagnosis of (Hamy); Department of Surgery, Tours Hospital, Tours, France (Bretagnol, thyroid metastasis. Patients with lymphoma or with di- de Calan). Correspondence address: Eric Mirallié, MD, Clinique Chirurgicale A, Hôtel rect extension of tumor from primary lesions in the head Dieu, CHU Nantes, 44093 Nantes Cedex 1, France. and neck were excluded. The information collected and

cancers were all clear cell cancers. Digestive cancers included two colonic cancers, one adenocarcinoma of the cardia, and one epidermoid carcinoma of the esophagus. The sarcoma was a leiomyosarcoma of the forearm. This patient had had a lower left pulmonary lobectomy for a metastasis 2 years before the ITM. Diagnoses of primary and thyroid localizations were synchronous in 10 patients (34.5%). For these patients, primary cancers were four lung cancers, two digestive cancers, one renal cancer, one melanoma, one neuroendocrine tumor, and the cancer of unknown origin. In six of these patients, the ITM revealed the cancer (three lung, one renal, one neuroendocrine, and one digestive). Three of the four other patients (with synchronous ITM) had tracheal compression. The patient with unknown primary cancer had been operated on for a brain lesion, which was a secondary tumor. Postoperatively, CT scan revealed a thyroid lesion. One of these four patients, without cervical com- Figure 1. Location and number of primary cancers. pression, had FNAC that led to diagnosis. All four of these patients had clinical examination, cervical ultrasonography, and CT scan. Thyroid locations were meta- analyzed included: patients’ age and gender, clinical pre- chronous in 19 patients (65.5%). They appeared with a sentation of the ITM, location of primary cancer, time mean delay of 6.8 years (2 months to 16 years) after from the initial cancer diagnosis to appearance of metas- diagnosis of the primary cancer. Delayed appearance of tasis, location and number of synchronous or metachro- the ITM in patients with renal cancer (excluding the nous metastases, and treatment and course of the disease patient with synchronous ITM) was 7.6 years (2 months after diagnosis of ITM. Information concerning fol- to 16 years). For nonrenal cancer, the mean delay for lowup was obtained by written questionnaire or direct metachronous ITM was 3.1 years (1 to 5 years). communication with the physicians or the patient. Twenty-one patients had total thyroidectomy and six StatView 5.0 was used for statistical analysis of data. isthmolobectomy (one of whom had a unilateral recur- Survival curves were drawn according to the Kaplan- rent nerve resection because of a massive infiltration). Meier method. Comparison of overall survival was per- Seven of the patients had lymph node resection because formed with the log-rank tests. Statistical significance of the presence of cervical lymphadenopathies (six of was set at a threshold of p Ͻ 0.05. those treated by total thyroidectomy and one treated by isthmolobectomy). Two patients had frozen section dur- RESULTS ing operation. In one patient, diagnosis was clear cell Twenty-nine patients were included in this series: 15 tumor; in the other patient, an intrathyroid parathyroid women and 14 men, with a mean age of 62 years (range adenoma was suspected. In both patients, it was an ITM 41 to 78 years). Revealing symptoms were a multinod- of renal origin on definitive examination. Total thyroid- ular goiter in 17 patients (2 had unilateral recurrent ectomy was performed because of massive infiltration: a nerve palsy), a thyroid nodule in 8, and a cervical lymph presence of nodule in the opposite lobe. In six patients node in 1 (data unknown for 1 patient). Four patients the reason was unknown. Two patients were not oper- had a tracheal compression. All patients but one had ated on because they had other distant metastases and normal thyroid function. One patient had hyperthy- had a preoperative diagnosis by FNAC. One was treated roidism. Five patients had fine-needle aspiration cytol- with chemotherapy (adenocarcinoma of the cardia) and ogy (FNAC). In three patients, FNAC led to diagnosis the other did not receive treatment (lung carcinoma and of ITM. In two patients, it was nonconclusive. synchronous ITM). Five patients had postoperative cer- The primary cancers are shown in Figure 1. Renal vical radiotherapy associated with one session of chemo- Vol. 200, No. 2, February 2005 Mirallié et al Intrathyroid Metastases 205

Table 1. Thirteen Patients with New Metastases During Followup Primary cancer Treatment of ITM Location Delay* Outcomes† Kidney Lobectomy Pleura 2 y Dead (3 y) Kidney TT Parotid, pancreas 1 y Alive with tumor (2.5 y) Kidney TT Kidney 1 y Dead (3 y) Kidney TT Bone 2 y Dead (4 y) Liver, pancreas 3 y Kidney TT Pancreas, kidney 5 y Dead (8 y) Kidney Lobectomy Pancreas 2 y Dead (2.2 y) Melanoma TT, Chemotherapy Bone 3 mo Dead (4 mo) Melanoma TT Skin 1 mo Dead (2 mo) Cardia Chemotherapy Mediastinal nodes 3 mo Dead (4 mo) Colon TT, LND Adrenal gland 2 mo Dead (4 mo) Lung Lobectomy, LND Brain 2 y Alive with tumor (2.2 y) Esophagus TT Bone 2 mo Dead (10 mo) Unknown TT, LND Bone 2 mo Dead (3 mo) ITM, intrathyroid metastases; LND, lymph node dissection; TT, total thyroidectomy. *ITM—new metastases. †Followup since diagnosis of ITM. therapy (patient suffering from a sarcoma). Two patients initial cancer diagnosis to the metastasis, and outcomes had adjuvant chemotherapy after thyroid intervention. for these patients are shown in Table 1. Histologic examination of the thyroids of the 27 pa- Actuarial patient survival as determined by Kaplan- tients who were operated on showed 9 patients with Meier analysis for all patients is shown in Figure 2. Dif- massive thyroid infiltration: 4 unilateral and 5 bilateral. ference in survival between patients with primary renal The primary tumors of the thyroid infiltrations were cancers and the others is shown in Figure 3. This differ- three digestive cancers, two melanomas, one lung, one ence was statistically significant (p ϭ 0.0001). At the neuroendocrine, one unknown origin, and one kidney. end of our study, 7 patients (24%) were alive and disease In 10 patients, there was a solitary metastatic nodule and free (6 renal cancers and the sarcoma) after 4.5 years 6 had multiple metastatic nodules (data unknown for 2 (range 1 to 9 years). Four patients (14%) were alive with patients). All patients who had lymph node resection a progressive lesion after 1.4 years (range 2 to had involved nodes. One patient also had a 5-mm pap- 30 months). And, finally, 18 patients (62%) have died at illary cancer. The 16 histologic examinations of primary a mean delay of 1.4 years (range 1 month to 8 years) after renal cancers were compared with thyroid lesion and diagnosis of ITM. Patients who died of an extrarenal thyroglobulin immunostaining was performed. It was cancer all survived for less than 1 year. Among the 16 negative. These lesions were not clear cell primary thyroid cancer. After diagnosis of ITM (within 1 month), a complete evaluation was performed in all patients. Eight patients had other metastatic sites: five had pulmonary metastases, four hepatic metastases, three bone metastases, and two adrenal metastases. The primary sites of these cancers were two lung, two melanomas, and one each in the colon, kidney, and neuroendocrine, and one of unknown origin. Mean followup was 2.3 years (range 3 months to 9 years) after diagnosis of ITM. During this followup, 13 patients developed new metastases. The primary cancers, treatments of ITM, metastatic sites, time from the Figure 2. Actuarial survival for all patients. 206 Mirallié et al Intrathyroid Metastases J Am Coll Surg

current nerve palsy. Preoperative distinction between a primary versus a secondary thyroid neoplasm is almost impossible.7 These nodules appear as inhomogeneous, hypoechoic mass on ultrasound. The true metastatic nature of the tumor is recognized only after tumor sam- pling. Aspiration cytology was carried out in only five patients, and three of these led to a diagnosis. Its great interest was to avoid unnecessary thyroidectomy in two patients with poor prognosis. If this examination had been performed systematically, we would certainly have avoided operating on other patients with numerous metastases at the time of thyroidectomy. The small number of FNAC carried out is because this series took place over Figure 3. Actuarial survival for patients with renal versus extrarenal cancers. a period of 20 years, and FNAC was less widely practiced in the past. Noncontributory FNAC can result from interpretation difficulties (particularly if the primary patients presenting a renal cancer, 7 died after a mean cancer is unknown) or from the fact that the puncture delay of 3.3 years (range 3 months to 8 years) and 9 concerned a colloidal nodule of multinodular goiter. An- (56%) were alive (recurrence free for 6 patients) after a other potential source of confusion in cytologic interpre- delay of 3.7 years (3 months to 9 years). Among the 8 tation is the difficulty of distinguishing primary anaplas- patients who had ITM associated with other metastatic tic thyroid carcinoma from metastatic high-grade sites, 7 died (mean delay 6 months, range 2 to malignancy. Positive immunostaining for thyroglobulin 16 months) and the only surviving patient had a pro- suggests a primary thyroid malignancy.7 Currently, we gressive tumor and a short followup (3 months). Among think that aspiration cytology should be performed in the 21 patients who had isolated ITM, 11 were deceased cases of thyroid nodules, particularly in patients with a (mean delay 2.1 years, range 2 months to 8 years), 7 were history of cancer. alive without tumor, and 3 were alive with tumor. Not all of our patients had total thyroidectomy despite the recommendations of some authors.3 Histologic DISCUSSION examinations showed that, with the exception of massive In our series, as described in the literature, clear cell renal infiltrations of thyroid tissue, the nodular metastasis was cancer is the most commonly found (55%) primary site solitary in 10 patients out of 16. We think that total with ITM.1,4,5,8 Other primary cancers were mainly lung, thyroidectomy is only advisable where bilateral nodules digestive, and melanoma. Unlike other authors, we did are present on preoperative ultrasonography. A lymph- not observe any ITMs in patients with breast cancer.1,3,8 adenectomy was carried out on seven patients. It was The site of the primary cancer is not always localized, only performed for cases of preoperative macroscopic despite all the radiologic examinations.1,3 Numerous pri- lymph node involvement, which explains the fact that all mary cancers were reported in the literature: colon, pa- these patients had involved nodes. Because of the small rotid gland, gastric leiomyosarcoma or adenocarcinoma, number of lymphadenectomies performed, we cannot esophagus, liver, uterus, uterine cervix, pancreas, blad- draw any conclusions about the mode of dissemination der, liposarcoma, and Merkel cell carcinoma.1,5,14-19 In (hematogenic or lymphatic) to explain the ITMs. In the six of our patients, the ITM revealed the primary cancer. literature, the hematogenic pathway is the more com- This has already been noted for cancers of the kidney, monly cited.3 liver, and lungs.3,7,16,20 In Heffess and colleagues’7 series In cases of renal cancers, we observed very long delays based on 36 ITMs of renal cancer, it revealed the pri- between diagnosis of the primary tumor and that of the mary cancer in one-third of patients. ITM (mean delay 7.6 years), as has already been re- The mode of revelation of the ITMs in our series was ported: 9.4 years for Heffess and colleagues7 and1to no different from a primary thyroid disease: multinod- 26 years for Nakhjavani and colleagues.1 Long disease- ular goiter, thyroid nodule, cervical lymph node, or re- free intervals have also been described for breast and Vol. 200, No. 2, February 2005 Mirallié et al Intrathyroid Metastases 207 uterine adenocarcinomas.1 There is no explanation for REFERENCES this long latent period between the identification of the 1. Nakhjavani M, Gharib H, Goellner J, van Heerden J. Metastasis primary cancer and the development of clinical thyroid to the thyroid gland. Cancer 1997;79:574–578. 7 2. Wychulis AR, Bearhs OH, Woolner LB. Metastasis of carci- metastasis. noma to the thyroid gland. Ann Surg 1964;160:169–177. In our series, 62% of patients died and only 24% were 3. Ménégaux F, Chigot JP.Les métastases thyroïdiennes. Ann Chir alive without evidence of disease.The prognosis depends 2001;126:981–984. 4. De Ridder M, Sermeus A, Urbain D, Storme G. Metastases to essentially on the primary cancer and the isolated nature the thyroid gland. A report of six cases. Eur J Intern Med 2003; of the ITM. Patients with renal cancer had a better prog- 14:377–379. nosis than others (see Fig. 3). Their mortality rate is 5. Chen H, Nicol T, Udelsman R. Clinically significant, isolated lower and the survival time of the deceased patients is metastatic disease to the thyroid gland. World J Surg 1999;23: 177–181. greater than that of patients presenting an extrarenal 6. McCabe D, Farrar W, Petkov T, et al. Clinical and pathologic cancer. Patients presenting a nodular ITM also had a correlations in disease metastatic to the thyroid gland. Am J Surg better prognosis than those who had a thyroid infiltra- 1985;150:519–523. 7. Heffess CS, Wenig BM, Thompson LD. Metastatic renal cell tion, but the primary cancers of the latter were eight carcinoma to the thyroid gland. A clinicopathologic study of 36 times out of nine nonrenal cancers. The patient present- cases. Cancer 2002;95:1869–1878. ing with sarcoma on the forearm was the only 8. Benoit L, Favoulet P, Arnould L, et al. Les métastases intrathy- roïdiennes du cancer du rein à cellules claires: présentation de 7 recurrence-free survivor with an extrarenal tumor. cas et revue de la littérature. Ann Chir 2004;129:218–223. These results indicate that surgical intervention 9. Lin JD, Weng HF, Ho YS. Clinical and pathological character- should mainly be performed on patients with renal can- istics of secondary thyroid cancer. Thyroid 1998;8:149–153. cer after establishing the solitary nature of the ITM, or at 10. Berge T, Lundberg S. Cancer in Malmö 1958–1969: an autopsy study. Acta Pathol Microbiol Scand Suppl 1977;260:1–235. least that the other metastases can be removed. When 11. Abrams HL, Spiro R, Goldstein N. Metastases in carcinoma: the ITM affects patients with extrarenal primary cancer, analysis of 1000 autopsied cases. Cancer 1950;3:74–85. widespread examinations should be carried out. It is very 12. Silverberg SG, Vidone RA. Metastatic tumors in the thyroid. 9 Pacific Med Surg 1966;74:175–180. often a polymetastatic disease. For polymetastatic dis- 13. Shimoaka K, Sokal JE, Pickren JW. Metastatic neoplasms in the ease, thyroid operation should only be performed in thyroid gland: pathological and clinical findings. Cancer 1961; patients presenting signs of tracheal compression. 15:557–565. 14. Ok E, Sözüer E. Thyroid metastasis from gastric carcinoma: FNAC is an essential tool to improve preoperative diag- report of a case. Surg Today 2000;30:1005–1007. nosis of thyroid nodule. 15. Barouk J, Hamy A, Mirallié E, et al. Goitre métastatique d’un cancer colique. Gastroenterol Clin Biol 2000;24:131–132. 16. Masuda T, Fukuya T, Ono M, et al. Thyroid metastasis from Author Contributions hepatocellular carcinoma as an initial presentation: a case report. Radiat Med 2001;19:43–46. Study conception and design: Mirallié, Kraimps 17. Cheung A, Donner L, Capen C. Metastatic adenocarcinoma of the Acquisition of data: Mirallié, Rigaud, Mathonnet, Gibelin, uterine cervix to the thyroid gland. Clin Oncol 2000;12:60–61. 18. Hsiao PJ, Tsai KB, Lai FJ, et al. Thyroid metastasis from intra- Regenet, Hamy, Bretagnol, de Calan, Le Néel ductal papillary-mucinous carcinoma of the pancreas. A case Analysis and interpretation of data: Mirallié, Rigaud, report. Acta Cytol 2000;44:1066–1072. Mathonnet, Kraimps 19. Wood K, Vini L, Harmer C. Metastases to the thyroid gland: the Drafting of manuscript: Mirallié, Rigaud, Regenet Royal Marsden experience. Eur J Surg Oncol 2004;583–588. 20. Giuliani A, Caporale A, Borghese M, et al. Papillary renal cell Critical revision: Kraimps carcinoma presenting as a nodal metastases to the neck. J Exp Supervision: Kraimps Clin Cancer Res 1999;18:579–582.