Contralateral cervical and mediastinal lymph node metastasis in medullary thyroid cancer: Systemic disease?

Contralateral cervical and mediastinal lymph node metastasis in medullary thyroid cancer: Systemic disease? Andreas Machens, MD,a Hans-Ju¨rgen Holzhausen, MD,b and Henning Dralle, MD,a Halle/Saale, Germany

Background. The current American Joint Committee on Cancer/International Union Against Cancer classification designates cervical and mediastinal lymph nodes as regional lymph nodes. In a unilateral thyroid neoplasm, however, involvement of the contralateral cervical lymph node compartment or the mediastinal lymph node compartment, both of which have been designated ‘‘distant’’ lymph nodes, may serve as a surrogate parameter of distant metastases. Methods. This institutional series consisted of 105 consecutive patients with medullary thyroid cancer who underwent systematic dissections of both distant lymph node compartments. Results. Thirty-eight patients had no distant lymph node metastasis, 36 patients had involvement of only 1, and 31 patients of both distant lymph node compartments. Significant associations (P < .001) were seen on univariate analysis between the number of involved ‘‘distant’’ lymph node compartments (none, one, or both) and extrathyroidal extension (3%, 33%, and 58%), the number of positive lymph nodes (means of 3, 13, and 33), and distant metastasis (8%, 36%, and 61%). In a multivariate logistic regression model, only involvement of one or both ‘‘distant’’ lymph node compartments (versus no distant lymph node metastasis) remained significantly related to distant metastasis in a dose-dependent fashion. Conclusions. ‘‘Distant’’ lymph nodes in medullary thyroid cancer should be regarded as nonregional lymph nodes because their involvement is indicative of distant metastasis. (Surgery 2006;139:28-32.) From the Departments of General, Visceral and Vascular Surgery,a and Pathology,b Martin-Luther-University Halle-Wittenberg, Germany

IN A VARIETY OF MALIGNANCIES, eg, bronchial, esophageal, gastric, ampulla of Vater, colorectal, breast, and skin cancers, the American Joint Committee on Cancer (AJCC)1 and International Union Against Cancer (UICC) Tumor-Node-Metastasis (TNM)2 classifications designate ‘‘distant’’ lymph nodes as nonregional lymph nodes, counting their involvement as distant metastasis (M1 lymph). In thyroid cancer, however, the existence of such nonregional or ‘‘distant’’ lymph nodes currently is not recognized and remains unproven. Dissemination of thyroid cancer cells through the lymphatic system of the neck is believed to evolve in a stepwise fashion: nodal metastases begin in the central (UICC groups 1, 2, and 8)

Accepted for publication June 3, 2005. Reprint requests: Andreas Machens, MD, Department of General, Visceral and Vascular Surgery, Martin-Luther-University Halle-Wittenberg, Ernst-Grube-Straße 40, D-06097, Halle/Saale, Germany. E-mail: [email protected]. 0039-6060/$ - see front matter Ó 2006 Mosby, Inc. All rights reserved. doi:10.1016/j.surg.2005.06.018

28 SURGERY

and the lateral cervical (UICC groups 3 through 7) lymph node compartments on the side of the primary neoplasm (regional lymph nodes), and then progress to the contralateral cervical (UICC groups 3 through 7) and the mediastinal lymph node compartment.3 These latter 2 compartments, hence, may be considered as representing nonregional, ‘‘distant’’ lymph node compartments of the thyroid gland. Notorious for its propensity to spread extensively to lymph nodes, medullary thyroid cancer may be particularly suited for investigation of the association between distant lymph node metastasis and distant metastasis. This notion is also supported by the close correlation between the number of positive lymph nodes in this condition and both the rates of postoperative calcitonin normalization and gross distant metastases.4-6 PATIENTS AND METHODS Patient selection. Between November 1994 and September 2004, 105 of 418 consecutive patients (25%) with medullary thyroid cancer seen at this institution had a total thyroidectomy with systematic

Machens, Holzhausen, and Dralle 29

Surgery Volume 139, Number 1

Table I. Univariate analysis of clinicopathologic parameters grouped by number of involved distant compartments Involvement of distant compartments*

Age at first diagnosis, years, mean [95% CI] Gender, male (%) Primary operation (%) Heredity (RET mutation) (%)§ Diameter of primary neoplasm, mm, mean [95% CI]{ Extrathyroidal extension (%)** Positive lymph nodes mean [95% CI] Distant metastases (%)yy Normalization of serum calcitonin postoperatively (%){

None (n = 38y)

One (n = 36)

Both (n = 31)

P valuez

46 [41,51] 39 26 26 25 [20,31]

43 [37,48] 42 28 31 31 [24,37]

43 [35,50] 48 48 42 31 [24,38]

.62 .73 .12 .40 .42

3 3 [1,5] 8§§ 20

33 13 [10,17] 36 10

58 33 [22,44] 61 0

<.001 <.001 <.001 .06

CI, Confidence interval; RET, rearranged during transfection. *Contralateral cervical lymph node compartment and mediastinal lymph node compartment. yOf whom 28 patients (74%) had node-positive and 10 patients (26%) had node-negative tumors. zTwo-tailed Fisher exact test and unifactorial analysis of variance, respectively. §Among the RET carriers, 70%, 70%, and 100% (P = .08) were index patients. {Based on 87 patients (diameter of primary neoplasm) and 85 patients (normalization of postoperative calcitonin), respectively. **Details unavailable for 1 patient. yyAs ascertained at the time of operation. §§All patients with distant metastases had node-positive tumors.

dissections of the central, and both lateral lymph node compartments of the neck and the mediastinal lymph node compartment, through a combined cervical-transsternal approach, as described elsewhere.3 Informed consent was obtained before operation for each of the above procedures that represented standard practice of care.7 Having undergone comparable operations, these 105 patients (35 primary and 70 reoperative patients) form the study population. All surgical specimens were subjected to pathologic examination. Perioperative calcitonin levels were determined using a solid 2site immunoradiometric assay (ELSA-hCT; CIS bio International, Gif-sur-Yvette, France). Normalization of serum calcitonin was assumed after stimulation with pentagastrin (Peptavlon; Laboratoires SERB, Paris, France) when the upper normal limit of the calcitonin assay (10 pg/mL) was not exceeded. Pathologic examination and staging. The thyroid halves were sectioned horizontally from the superior to the inferior pole, as described previously.8 After fixation in formalin, the whole thyroid gland was embedded in paraffin. Soft tissue and lymph nodes were processed separately. Conventional staining (H & E) and, where appropriate, calcitonin immunohistochemistry were performed on every surgical specimen, using the standard avidin-biotin complex peroxidase approach. The diagnosis of medullary thyroid cancer was based

on the World Health Organization histologic classification of tumors.9 The size of the primary neoplasm, which had been ascertained by direct measurements on the surgical thyroid specimens, was taken from the pathology reports. Extrathyroidal extension was defined as any extension of the thyroid neoplasm beyond the thyroid capsule. When more than one primary neoplasm was identified on histopathologic examination, only the largest primary neoplasm was counted, because this method was shown previously to yield comparable patterns of nodal metastases in multifocal and unifocal medullary thyroid cancer.10 In the patients undergoing reoperation, staging included primary and, where applicable, all subsequent histopathologic findings obtained before reoperation at our institution. A diagnosis of nodal metastasis always required histopathologic confirmation. This need of histopathologic confirmation was waived for distant metastases when there was unequivocal evidence on computerized tomography, magnetic resonance imaging, positron emission tomography, scintiscan, or any combination thereof. Statistical analysis. Categorical and continuous data were tested on univariate analysis with the 2-tailed Fisher exact test and unifactorial ANOVA. A conditional logistic regression model on distant metastasis was then fitted to examine those factors identified as significant in the univariate analysis. The level of statistical significance was set at .05.

30 Machens, Holzhausen, and Dralle

Surgery January 2006

Table II. Logistic regression analysis on distant metastases* Parameter

No.

OR

95% CI

P value

Extrathyroidal (vs intrathyroidal) extensiony Number of positive lymph nodesz Involvement of distant lymph node compartments§ None One Both

31 versus 73 104

2.7 1.00

0.9,7.5 0.97,1.02

.06 .72

Reference 1.2,20.3 2.5,61.2

.03 .002

37 36 31

1.0 4.9 12.3

OR, Odds ratio, serving as an estimate of the relative risk where this cannot be measured directly; CI, confidence interval. *As ascertained at the time of operation. yDetails unavailable for 1 patient. zContinuous variable. §Contralateral cervical lymph node compartment and mediastinal lymph node compartment.

RESULTS Univariate analysis grouped by number of involved lymph node compartments. Univariate analysis (Table I) found associations among the number of involved distant lymph node compartments (the contralateral cervical or mediastinal compartment alone or both) and extrathyroidal extension (3%, 33%, and 58%; P < .001), the number of positive lymph nodes (means of 3, 13, and 33; P < .001), and distant metastasis (8%, 36%, and 61%; P < .001). Demographic parameters such as patient age at first diagnosis of the neoplasm; gender; operative status (primary operation); heredity, ie, genetic evidence of a RET (rearranged during transfection) germline mutation; diameter of the primary neoplasm; and rates of postoperative calcitonin normalization did not differ among the groups. It is noteworthy that normalization of postoperative serum calcitonin was never accomplished despite extensive lymph node dissection when both distant lymph node compartments were involved. Multivariate analysis on distant metastasis. Altogether, 35 of the 105 patients (33%) had distant metastases, which involved 1 organ system in 29 patients (lung, 22 patients; bone, 3 patients; liver, 3 patients; brain, 1 patient), 2 organ systems in 5 patients (lung and liver combined, 3 patients; bone and liver combined, 2 patients), and 3 organ systems in 1 patient (lung, bone, and liver combined). In a multivariate logistic regression model controlled for extrathyroidal extension and the number of positive lymph nodes (Table II), only involvement of one or both distant lymph node compartments (versus lack of distant lymph node metastases) remained significantly related to distant metastasis. The magnitude of this effect, expressed as the odds ratio, was the greater, the more distant lymph node compartments were affected: 4.9 for involvement of 1 lymph node

compartment (P = .03), and 12.3 for involvement of both distant lymph node compartments (P = .002), suggesting a dose-dependent effect. DISCUSSION In this study, distant metastases were often encountered (36% to 61%) when distant lymph node compartments were affected (Table I). Involvement of both distant lymph node compartments was incompatible with postoperative normalization of serum calcitonin. Involvement of a single distant lymph node compartment also rarely (10%) resulted in postoperative normalization of serum calcitonin (Table I). Actual surgical cure rates may even be lower because calcitonin levels can surge even after initial normalization.11 The almost linear, dose-effect relationship between distant metastases and incremental involvement of distant compartments relative to noninvolved distant lymph node compartments (Table II) is indicative of a causal relationship, as is the strength of this association, ie, the size of the odds ratios.12 These data suggest that the contralateral cervical lymph nodes and the mediastinal lymph nodes, designated as regional lymph nodes in the current AJCC1 and UICC TNM2 classifications, in fact, may be nonregional lymph nodes in medullary thyroid cancer because their involvement is largely equivalent to distant metastasis. Involvement of distant lymph node compartments and distant metastasis. In agreement with previous data on lymphatic dissemination of medullary thyroid cancer,3,4 it seems that the ipsilateral cervical lymph node compartments constitute the watershed between local and systemic disease. With strictly ipsilateral lymphatic drainage of tumor cells (ie, involvement of central and lateral cervical lymph node compartments on the side of the primary neoplasm only), surgical cure may be attainable in some patients. In stark contrast, those

Surgery Volume 139, Number 1

patients who may already have progressed beyond these ipsilateral nodal groups usually are not curable by the addition of surgical lymphadenectomy of distant lymph node compartments. Indeed, contralateral lymphatic drainage of tumor cells (ie, involvement of distant lymph node compartments) is indicative of systemic disease, as evidenced by the frequent failure of calcitonin levels to normalize despite extensive lymph node dissection. Of note, the overall number of positive lymph nodes was not associated with distant metastasis when the involvement of distant lymph node compartments was accounted for. It has been shown that calcitonin levels rarely normalize when 10 and more lymph nodes are positive.4-6 This condition was met in our high-risk population when 1 distant lymph node compartment (mean of 13 positive nodes, lower 95% confidence interval [CI] bound of 10 positive nodes) or both distant lymph node compartments (mean of 33 positive nodes, lower 95% CI bound of 22 positive nodes) were involved (Table I). In some instances, distant metastasis may also develop independently of lymph node metastasis; despite extensive neck surgery, 38% of patients with node-negative medullary thyroid cancer did not reach normal calcitonin levels.13 These rates are even lower in unselected patients with node-positive medullary thyroid cancer in which calcitonin levels normalize in only 10% to 32% of cases.5,6,13 Clinical implications. The key element of success in curing sporadic medullary thyroid cancer is early detection, which may be facilitated by liberal use of measurements of serum calcitonin in seemingly benign thyroid disease.14 In symptomatic patients, however, medullary thyroid cancers frequently are node-positive and more advanced. Clearance of lymph node metastases by systematic dissection may afford local control in the neck3,4-6,10,15-19 and mediastinum20 and may prevent tracheal and esophageal invasion owing to tumor penetration of the lymph node capsule.21 From a surgical perspective, it is important to distinguish between local disease, which may be surgically curable, and systemic disease, in which palliation is the ultimate goal of therapy. To make this distinction, especially when preoperative imaging fails to visualize distant metastases, intraoperative proof of contralateral lymph node metastases is necessary. To this end, it is recommended to remove suspicious lymph nodes from the contralateral neck and, where feasible without sternotomy, also from the upper mediastinum, until histopathologic confirmation of nodal metastasis is obtained on fresh frozen section. Sternotomy and mediastinal lymph node

Machens, Holzhausen, and Dralle 31

dissection should be considered when there is radiologic evidence of progressive or symptomatic mediastinal lymph node metastases and in patients undergoing reoperation for node-positive, extrathyroidal medullary thyroid cancer.20 Whereas local disease undoubtedly requires systematic lymph node dissection, there is no universal consensus on how to proceed in systemic disease. Surgical options in this setting range from the sampling of gross lymph node metastases only, or ‘‘berry picking,’’10,16,19 to systematic lymph node dissection of all involved lymph node compartments.3,4,6,11,13 Given the comparative longevity of many patients in the face of systemic disease, systematic dissection of the central and both lateral lymph node compartments of the neck may still have a role, not only with regard to local control, but also because of the 10% chance of postoperative calcitonin normalization that exists when only 1 distant lymph node compartment is involved. Further studies with prolonged periods of follow-up are needed to clarify the optimal extent of lymph node dissection in the contralateral neck. REFERENCES 1. Greene FL, Balch CM, Fleming ID, Fritz A, Haller DG, Morrow M, et al, editors. AJCC cancer staging manual. 6th ed. Berlin: Springer; 2002. 2. Sobin LH, Wittekind C, editors. UICC: TNM classification of malignant tumors. 6th ed. New York: Wiley-Liss; 2002. 3. Machens A, Hinze R, Thomusch O, Dralle H. Pattern of nodal metastasis for primary and reoperative thyroid cancer. World J Surg 2002;26:22-8. 4. Machens A, Gimm O, Ukkat J, Hinze R, Schneyer U, Dralle H. Improved prediction of calcitonin normalization in medullary thyroid carcinoma patients by quantitative lymph node analysis. Cancer 2000;88:1909-15. 5. Weber T, Schilling T, Frank-Raue K, Colombo-Benkmann M, Hinz U, Ziegler R, et al. Impact of modified radical neck dissection on biochemical cure in medullary thyroid carcinomas. Surgery 2001;130:1044-9. 6. Scollo C, Baudin E, Travagli JP, Caillou B, Bellon N, Leboulleux A, et al. Rationale for central and bilateral lymph node dissection in sporadic and hereditary medullary thyroid cancer. J Clin Endocrinol Metab 2003;88:2070-5. 7. Ro¨her HD, Simon D, Goretzki PE. Guidelines in oncologic surgery: malignant thyroid tumors. Langenbecks Arch Surg 1997;114(Suppl II):142. 8. Hinze R, Holzhausen HJ, Gimm O, Dralle H, Rath FW. Primary hereditary medullary thyroid carcinoma—C-cell morphology and correlation with preoperative calcitonin levels. Virchows Arch 1998;433:203-8. 9. World Health Organization. International histological classification of tumours. Geneva: World Health Organization; 1969-1981. 2nd ed. Berlin: Springer-Verlag; 1988. 10. Moley JF, DeBenedetti MK. Patterns of nodal metastases in palpable medullary thyroid carcinoma. Recommendations for extent of node dissection. Ann Surg 1999;229:880-8. 11. Franc S, Niccoli-Sire P, Cohen R, Bardet S, Maes B, Murat A, et al. Complete surgical lymph node resection does not

32 Machens, Holzhausen, and Dralle

12. 13.

14.

15.

16.

prevent authentic recurrences of medullary thyroid carcinoma. Clin Endocrinol 2001;55:403-9. Hennekens CH, Buring JE, Mayrent SL, editors. Epidemiology in medicine. Boston: Little, Brown and Company; 1987. Machens A, Schneyer U, Holzhausen HJ, Dralle H. Prospects of remission in medullary thyroid carcinoma according to basal calcitonin level. J Clin Endocrinol Metab 2005;90:2029-34. Hodak SP, Burman KD. Editorial: the calcitonin conundrum—is it time for routine measurement of serum calcitonin in patients with thyroid nodules? J Clin Endocrinol Metab 2004;89:511-4. Moley JF, Wells SA, Dilley WG, Tisell LE. Reoperation for recurrent or persistent medullary thyroid cancer. Surgery 1993;114:1090-6. Moley JF, Dilley WG, DeBenedetti MK. Improved results of cervical reoperation for medullary thyroid carcinoma. Ann Surg 1997;225:734-43.

Surgery January 2006

17. Gimm O, Dralle H. Reoperation in metastasizing medullary thyroid carcinoma: is a tumor stage-oriented approach justified? Surgery 1997;122:1124-31. 18. Gimm O, Ukkat J, Dralle H. Determinative factors of biochemical cure after primary and reoperative surgery for sporadic medullary thyroid carcinoma. Surgery 1998;22: 562-8. 19. Chen H, Roberts JR, Ball DW, Eisele DW, Baylin SB, Udelsman R, et al. Effective long-term palliation of symptomatic, incurable metastatic medullary thyroid cancer by operative resection. Ann Surg 1998;227:887-95. 20. Machens A, Holzhausen HJ, Dralle H. Prediction of mediastinal lymph node metastasis in medullary thyroid carcinoma. Br J Surg 2004;91:709-12. 21. Machens A, Hinze R, Lautenschla¨ger C, Thomusch O, Dralle H. Thyroid carcinoma invading the cervicovisceral axis: routes of invasion and clinical implications. Surgery 2001;129:23-8.