Pulmonary Metastasectomy for Head and Neck Cancers

Pulmonary Metastasectomy for Head and Neck Cancers

David Liu, MD, Daniel M. Labow, MD, Nick Dang, AB, Nael Martini, MD, Manjit Bains, MD,Michael Burt, MD, PhD, Robert Downey, Jr., MD, Valerie Rusch, MD,

Jatin Shah, MD, and Robert J. Ginsberg, MD

Background: Distant metastases from carcinomas that arise from the head and neck region areinfrequent. The most common site is the lung. To evaluate the results of resection of pulmonarymetastases for head and neck cancers, we reviewed our own cases of these metastases.

Methods: Between November 1966 and March 1995, 83 patients with pulmonary metastasesfrom head and neck cancers underwent 94 thoracic operations. All patients had obtained or hadobtainable locoregional control of their primary head and neck cancers. Kaplan-Meier and Coxregression models were used to analyze the prognostic factors for survival after metastasectomy.

Results: Median age was 53 years (range, 17–77). Fifty-nine were male and 24 were female.Forty-one patients had squamous cell cancers, and 36 had glandular tumors that consisted mostly ofthyroid and adenoid cystic carcinomas. The median disease-free interval from the time of treatmentof the head and neck primary cancers to the development of pulmonary metastases was 27 months.Sixty-eight (82%) patients had complete resection. Overall operative mortality rate was 2%. Overallactuarial survival rate after metastasectomy was 50% at 5 years. Patients with glandular tumors hada 5-year survival rate of 64% compared with 34% for patients with squamous cell cancers. Whenthe patients with glandular tumors were analyzed according to their histology, patients with adenoidcystic carcinomas had an 84% 5-year survival, but none remained disease-free. Patients with thyroidcancers fared similarly whether they were treated medically or surgically. On multivariate analysis,the adverse prognostic factors for patients with squamous cell cancers were incomplete resection,age greater than 50 years, and disease-free interval less than or equal to 2 years.

Conclusions: Approximately 30% of patients with pulmonary metastases from squamous cellcancers of the head and neck who underwent complete resection of all their metastases can expectto achieve long-term survival. The role of pulmonary resection for patients with glandular tumorsis unclear.

Key Words: Head and neck cancer—Lung metastases—Prognostic factors.

In 1998, there were an estimated 60,000 new cases ofhead and neck cancers in the United States,1 and theanticipated incidence of distant metastases from thesetumors ranged from 11% to 40%.4,5 The most commonsite of metastasis is the lung. Although pulmonary resec-tion has become the standard therapy for a variety ofmetastatic malignancies to the lungs, including sarco-

mas, germ cell tumors, colorectal cancers, and renal cellcancers, there have been only a few studies examiningthe role of pulmonary metastasectomy for patients withhead and neck cancers. Pulmonary resection is acceptedfor the treatment of a solitary lesion metastatic fromsquamous cell cancers of the head and neck. However,its role in the management of head and neck cancerpatients with multiple lung metastases and glandulartumors is unclear. We report our experience with pulmo-nary metastasectomy in patients with head and neckcancers and assess its role in their treatment.

MATERIALS AND METHODS

Between November 1966 and March 1995, 10,478patients with a diagnosis of head and neck cancer were

Received May 28, 1998; accepted April 23, 1999.From Thoracic Service (DL, DML, ND, NM, MB, MB, RD Jr, VR,

RJG) and Head and Neck Service (JS), Department of Surgery, Me-morial Sloan-Kettering Cancer Center, New York, New York.

Presented in part at the 51st Annual Symposium Meeting of theSociety of Surgical Oncology, San Diego, March 26–29, 1998.

Address correspondence to: Robert J. Ginsberg, MD, Thoracic Ser-vice, Department of Surgery, 1275 York Avenue, New York, NY10021; Fax: 212-639-2807; E-mail: [email protected]

Annals of Surgical Oncology, 6(6):572–578Published by Lippincott Williams & Wilkins © 1999 The Society of Surgical Oncology, Inc.

572

seen at Memorial Sloan-Kettering Cancer Center. Twohundred of these patients (1.9%) had developed pulmo-nary lesions in the course of their follow-up. Pulmonarylesions were determined to be either a primary cancer ormetastasis according to the criteria previously de-scribed.2 Of these, 87 patients were treated surgically andfound to have pulmonary metastases from their head andneck cancer. All patients who underwent resection oftheir pulmonary metastases met the following criteria:(1) pulmonary lesions were deemed resectable by radio-logical examinations, (2) metastatic disease was limitedto the lungs, and (3) locoregional control of their headand neck primary cancer was obtained or obtainable.Eighty-three patients met these criteria. Four patientswith thyroid cancers who had an open biopsy for adiagnosis were excluded. Complete clinical and patho-logical data were reviewed. The clinical outcome wasmeasured from the initial thoracotomy to last follow-upexamination or death. Last follow-up categories weredefined as no evidence of disease (NED), alive withdisease (AWD), dead of other causes (DOC), and dead ofdisease (DOD).

All statistical analyses were performed using SPSS 6.1for Windows. The Kaplan-Meier and Cox regressionmethods were used to analyze prognostic factors forsurvival after resection of metastatic disease. A P valueof less than .05 was considered statistically significant.

RESULTS

DemographicsEighty-three patients with pulmonary metastases from

head and neck primary cancers underwent a thoracotomyand resection. Fifty-nine patients were male and 24 werefemale. The median age was 53 years, with a range of17–77 years. Forty-one patients had squamous cell can-cers; 36 patients had glandular tumors; and 6 patients hadmiscellaneous tumors, which included a melanoma, an

osteogenic sarcoma, an ameloblastoma, a neuroblastoma,and two spindle cell carcinomas. The squamous cellcancers originated from the pharynx, larynx, and oralcavity. Adenoid cystic16 and thyroid carcinomas8 com-posed the majority of glandular tumors and were mostlyfrom the salivary and thyroid glands (Tables 1 and 2).Other glandular tumors included were mucoepidermoid,5

adenocarcinoma,4 and parathyroid.3 Table 3 lists thepulmonary clinical features of the patients at the time ofpresentation. Five patients presented with pulmonarymetastases within 6 months of the treatment of theirprimary head and neck cancer. Of the remainder, themedian interval time to the development of pulmonarymetastases was 27 months. Thirty-six patients (43%) hada solitary pulmonary metastasis. For the 47 patients withmultiple metastases, which were confirmed by pathol-ogy, the median number of lesions was 3 (range, 2–25).Twenty-four of these patients with multiple lesions hadmetastases that were confined to one lung.

TreatmentSixty-two patients underwent a unilateral thoracotomy

for resection of their metastases, whereas the remainder

TABLE 1. Histologies of head and neck primaries

Primary histologyNumber of patients

(%)

Squamous cell 41 (49)Adenoid cystic 16 (19)Thyroid 8 (10)

Papillary 4Follicular 1Hurthle cell 2Anaplastic 1

Mucoepidermoid 5 (6)Adenocarcinoma 4 (5)Parathyroid 3 (4)Other 6 (7)

TABLE 2. Location of head and neck primaries

Tumor type Location

Number ofpatients

(%) Total

Squamous cell carcinomas Larynx 19 (46)Pharynx 13 (32) 41Oral cavity 9 (22)

Glandular tumors Salivary gland 14 (39)Thyroid 8 (22)Oral cavity 6 (17)Pharynx 4 (11) 36Parathyroid 3 (8)Skin 1 (3)

Other Pharynx 3 (50)Oral cavity 1 (17)Salivary gland 1 (17) 6Skin 1 (17)

TABLE 3. Pulmonary clinical features at presentation

Number of patients(%)

No. of pulmonary metastasesSolitary 36 (43)Multiple 47 (57)

Extent of lung involvementUnilateral 59 (71)Bilateral 24 (29)

Presentation of metastasesSynchronous 5 (6)Metachronous 78 (94)

573PULMONARY METASTASECTOMY FOR HEAD AND NECK CANCERS

Ann Surg Oncol, Vol. 6, No. 6, 1999

had a staged bilateral thoracotomy, a median sternotomy,or a clamshell thoracotomy (bilateral anterior thoracoto-mies with transverse sternotomy) for bilateral lesions(Table 4). An attempt was made to conserve as muchlung tissue as possible. This was reflected by the fact thatwedge resections were the most common proceduresperformed (Table 5). Sixty-eight patients underwent cur-ative resection. The remaining 15 patients were found tohave tumors that were either unresectable at the time ofoperation4 or had an incomplete resection (either bygross or histological examination).11 Ten of these 15patients had intraoperative isotopes implanted for adju-vant local therapy. The patients who had incomplete orunresectable lesions included 10 of 41 (24%) with squa-mous cell carcinomas, 2 of 5 (40%) with mucoepider-moid carcinomas, and 3 of 16 (19%) with adenoid cysticcarcinomas.

Two patients died postoperatively, one from cardiacarrest and the other from pneumonia, for an overalloperative mortality of 2%. Eleven patients (13%) hadpostoperative complications that included prolonged airleaks in three; hypotension in two; and arrhythmia, uri-nary retention, pulmonary embolism, bleeding, atelecta-sis, and postthoracotomy syndrome in one patient each.

Twenty-seven patients remained free of disease aftertheir first thoracic operation, and 56 patients developedrecurrences, the majority of which were located in thelung and locoregional sites (Table 6). The median time torecurrence after pulmonary resection was 16 months. Ofthe 24 patients with unilateral disease who had recur-rences in the lungs after pulmonary resection, the recur-rence occurred in the opposite lung in 17 patients. Thirty-six patients had recurrences at one site, 17 patients at twosites, 2 patients at three sites, and 1 patient at four sites.

SurvivalThe overall survival after metastasectomy was 50% at

5 years and 35% at 10 years. Forty-four percent and 23%of the patients were rendered free of disease 5 and 10years after pulmonary resection, respectively (Fig. 1).Twenty patients are currently alive without evidence ofdisease (NED) and 13 with disease (AWD). Forty-eightpatients died of disease (DOD) (includes 2 patients whodied postoperatively), and 2 patients died of other causes(DOC). The median time at follow-up examination was38 months.

The 41 patients with squamous cell cancers faredworse than patients with glandular tumors. The overall5-year survival rate for the patients with squamous cellcancers was 34% compared with 64% for patients withtumors of glandular origin (P � .014; Fig. 2). When theglandular tumors were separated by histology, patientswith thyroid cancers fared best with a 5-year survivalrate of 75% (Fig. 3). Six of eight patients who hadresection of all their thyroid metastases without furthertherapy were alive with a median follow-up of 50 months(range, 15–353). Patients with adenoid cystic carcinomashad a 5-year survival rate of 84%, but their overallsurvival curve never reached a plateau and continued to

FIG. 1. Overall and disease-free survival of patients following pul-monary resection (n � 83).

TABLE 4. Operative approach

Type of exposureNumber of patients

(%)

Thoracotomy 62 (75)Bilateral staged thoracotomy 11 (13)Sternotomy 6 (7)Clamshell 4 (5)

TABLE 5. Extent of resection

Type of operationNumber of patients

(%)

Wedge resection 62 (75)Lobectomy 16 (19)Pneumonectomy 1 (1)Exploratory (no resection) 4 (5)

TABLE 6. Sites of recurrence after initial thoracotomy(n � 56 patients)

LocationNo. of patients with

site involved

Locoregional 18Distant

Lung 40Bone 10Brain 9Liver 3

574 D. LIU ET AL.


decline until there were no disease-free survivors after 14years. In contradistinction, patients with mucoepider-moid cancers behaved the same as those with squamouscell cancers.

Recognizing that patients with squamous cell cancershad a different biological behavior than those with glan-dular tumors, several prognostic factors were analyzedfor the squamous cell carcinoma group only. On multi-variate analysis, the adverse prognostic factors for worsesurvival were incomplete resection, disease-free intervalless than or equal to 2 years, and age greater than 50years (Table 7). Patients with a solitary metastasis fromsquamous cell cancers had a 33% 10-year survival ratecompared with 22% for patients with multiple metastases(Fig. 4).

Resection of all metastases was possible in 68 patients(82%). The overall 5-year survival in this group was 56%.Fifteen patients who were either found to be unresectable atthe time of thoracotomy or had incomplete resection had anoverall 5-year survival rate of 22% (P � .0004; Fig. 5).

DISCUSSION

With improved local and regional control in head andneck cancers, distant metastases, as the only sites offailure, have become increasingly common and occurwith a frequency of 11% to 40% with the lung being themost common site.3–7 Historically, patients who devel-oped distant metastases had a poor prognosis and were

FIG. 2. Overall survival according to type of tumors.

FIG. 3. Overall survival according to histologies.FIG. 4. Overall survival according to number of metastases (squa-mous cell cancers).

TABLE 7. Results of univariate and multivariate analysesin patients with squamous cell carcinomas (n � 41)

Variables

No. ofpatients

(%)Univariate

P valueMultivariate

P value

Age�50 y 14 (34) 0.0015 0.016�50 y 27 (66)

SexMale 34 (83) 0.34Female 7 (17)

Disease free interval�2 y 24 (58) 0.032 0.030�2 y 17 (42)

Location of primaryLarynx 19 (46) 0.62Pharynx 13 (32)Oral cavity 9 (22)

No. of pulmonary metastasesSolitary 20 (49) 0.29Multiple 21 (51)

Largest size of metastases�1 cm 6 (15) 0.58�1 cm 35 (85)

Location of metastasesUnilateral 32 (78) 0.52Bilateral 9 (22)

Extent of resectionComplete 31 (76) 0.0008 0.0003Incomplete 10 (24)

Presentation of metastasesSynchronous 3 (7) 0.97Metachronous 38 (93)



not considered for resection. Lefor et al.8 reviewed 15patients with pulmonary metastases from head and neckcancers, who did not undergo pulmonary resection, andfound the median survival to be 3 months after diagnosisof the lung lesion with a 1-year survival of 7%. Iftreatment was considered at all, it consisted of chemo-therapy, which was mostly palliative. There were a fewresponses with the majority of them being short-lived.

Pulmonary metastasectomy has been accepted as astandard treatment for metastatic lesions to the lungcaused by a variety of primary tumors, including sar-coma, testicular, renal, and colorectal cancers.9 However,to date, there have been a few, limited studies thatexamine the role of pulmonary resection for metastasesfrom head and neck cancers. The majority of these stud-ies have focused on squamous cell histologies. Mazer etal.10 reported 44 patients who underwent pulmonary re-section for metastatic squamous cell carcinoma andfound the overall 5-year survival to be 43%. Finley andassociates11 compared 54 patients with squamous cellcancers of the head and neck who did or did not have athoracotomy for their pulmonary metastases. The 18patients who underwent pulmonary resection had an es-timated 5-year survival rate of 29% compared with lessthan 5% for the remainder who had no operation. Fur-thermore, the subgroup of patients who had achievedlocoregional control of the head and neck primary wasexamined and those who underwent pulmonary resectionhad a 5-year survival of 43% versus 10% for those whohad no thoracotomy. Wedman et al.12 found the presenceof squamous cell carcinoma to be a significant prognos-tic factor for poor survival rate after the diagnosis ofpulmonary metastases. Clearly, there are different bio-logical characteristics between squamous cell cancersand glandular tumors. In our study, patients who hadmetastatic squamous cell cancers from primary head andneck cancers had the worst outcome with 5-year survival

rates of 32% compared with 64% for patients who hadtumors of glandular origin.

Whereas pulmonary resection is accepted as a stan-dard treatment for patients with solitary metastases fromsquamous cell cancers of the head and neck, the role ofpulmonary resection in patients with multiple metastasesand glandular tumors is unclear. When patients withglandular histologies were separated into various sub-groups, patients with thyroid cancers had the best prog-nosis with a 5-year survival rate of 75%, which remainedunchanged thereafter. Standard first-line therapy for pa-tients with distant metastases from thyroid cancers con-sists of radioactive iodine, I131, for functioning tumors.13

Three of the four patients in our series with a diagnosisof papillary or follicular thyroid carcinoma, who under-went a thoracotomy for diagnostic intent only withoutresection of all their metastases and were treated witheither thyroid suppression or I131, are still alive with amedian follow-up of 52 months (range, 10–188 months).However, when there is limited or no uptake of I131,patients seem to benefit from surgical removal of thepulmonary lesions. Of the eight patients with thyroidcancer who had resection of all their pulmonary lesionsas their only treatment, six patients are still alive with amedian follow-up of 50 months (range, 15–353).

Adenoid cystic carcinomas are rare tumors, character-ized by a slow, protracted course with multiple recur-rences. Unlike our patients with squamous cell carcino-mas or thyroid cancers, the survival rate of these patientscontinued to decline without ever reaching a plateau.Kim et al.14 evaluated the survival of patients with ade-noid cystic carcinoma after distant metastases and foundthe 3-year survival rate to be 41.3%, which declined to15.5% at 5 years. It is unclear from their study whattreatment the patients with distant metastases received.Mazer et al.10 reported a 5-year overall survival of 63%in 13 patients. In our series, patients with pulmonarymetastases who had undergone resection had a 5-yearestimated survival rate of 84%, which continued to de-cline until there were no survivors after 14 years. Therewere no differences in survival rates between those pa-tients with solitary or multiple lesions. Although it isunclear whether pulmonary resection prolonged the sur-vival of these patients, resection did not cure these pa-tients.

All of the mucoepidermoid cancers in our series orig-inated from the salivary glands, and their behavior afterpulmonary resection was worse than that of thyroid oradenoid cystic cancers and akin to the squamous cellcancers.

Some studies have found the number of metastases tobe a significant prognostic factor, whereas others have

FIG. 5. Overall survival according to completeness of resection.

576 D. LIU ET AL.


not.10–12 In our series, patients with a solitary metastasisfrom squamous cell cancers had a 10-year 33% survivalrate compared with 22% for those with multiple metas-tases. These results were confirmed by a large, interna-tional study of long-term results of more than 5000 casesof lung metastasectomy for a variety of tumors.15

Recurrence after pulmonary resection of head andneck metastases was a common problem. In 67% of ourpatients, cancer recurred after pulmonary resection withmore than half of them recurring in the lungs. This ismost likely a result of the presence of micrometastases inthe lungs at the time of operation and not as a result ofinadequate resection because most patients had gross andmicroscopically clear margins. Furthermore, 17 of 24patients with unilateral disease had recurrences in thecontralateral lung. The importance of obtaining locore-gional control of the primary head and neck cancer is inthe fact that the second most common site of recurrencewas the head and neck, which represented 32% of pa-tients who did have a recurrence. One of the criteria inselecting patients for pulmonary resection was that thepatients have locoregional control of the primary cancer.Presumably, these patients who had a recurrence at thehead and neck sites had occult disease at their locore-gional site at the time of thoracotomy.

Patients with head and neck cancers are at an in-creased risk for developing a new and separate primarylung cancer with the reported incidence ranging from4.5% to 6.9%.16–18 When a solitary pulmonary nodule isdetected in a patient with head and neck cancer, primarylung cancers compose 78% of malignant lesions.19 Sol-itary pulmonary nodules are considered primary lungcancers if the histology differs from that of the primaryhead and neck cancer or, if on pathological examination,endobronchial carcinoma in situ is identified. In theabsence of these findings, differentiating a primary can-cer from a metastasis is difficult. We eliminated 113patients who were identified as having primary lungcancers according to these criteria. Twenty patients inour study had a solitary pulmonary nodule of squamouscell carcinoma that could not be excluded as a newprimary lung carcinoma. Ten of these patients underwenta wedge resection, and six underwent a lobectomy. Al-though there were no significant differences in survivalrates between the two different treatment groups, thesmall numbers make it difficult to interpret these results.Presumably, some of these patients who had wedgeresections for presumed metastatic disease may have hada primary lung cancer. The importance of making thisdistinction relates to the selection of appropriate treat-ment for the two different groups. In the future, geneticmarkers may be beneficial to distinguish between meta-

static and primary solitary nodules. Chung and associ-ates,20 using p53 analysis, showed that two-thirds of lungtumors in patients with previous head and neck cancerswere indeed primary cancers.

CONCLUSION

Patients with pulmonary metastases from head andneck cancers undergoing complete resection of all theirlesions can have long-term survival even if there aremultiple metastases. Patients with squamous cell andmucoepidermoid cancers do worse than patients withother glandular tumors. The role of pulmonary resectionis unclear in patients with indolent glandular tumors.Pulmonary resection may prolong the survival in patientswith adenoid cystic carcinomas, but pulmonary resectiondoes not seem to result in a cure in this group. As in othercancers, when selecting candidates for pulmonary resec-tion, the following criteria should be met: (1) the meta-static disease is limited to the lungs, (2) the lesions are allresectable, and (3) locoregional control of their head andneck primary is obtained or obtainable.

REFERENCES

1. Landis SH, Murray T, Bolden S, Wingo PA. Cancer statistics. CACancer J Clin 1998;48:6–29.

2. Martini N, Melamed MR. Multiple primary lung cancers. J ThoracCardiovasc Surg 1975;70:606–12.

3. Vikram BV, Strong EW, Shah JP, Spiro R. Failure at distant sitesfollowing multimodality treatment for advanced head and neckcancer. Head Neck Surg 1984;6:730–3.

4. Dennington ML, Carter DR, Meyers AD. Distant metastases inhead and neck epidermoid carcinoma. Laryngoscope 1980;90:196–201.

5. Merino OR, Lindberg RD, Fletcher GH. An analysis of distantmetastases from squamous cell cancers of the upper respiratory anddigestive tracts. Cancer 1977;40:145–51.

6. Calhoun KH, Fulmer P, Weiss R, Hokanson JA. Distant metastasesfrom head and neck squamous cell carcinomas. Laryngoscope1994;104:1199–205.

7. Papac RJ. Distant metastases from head and neck cancer. Cancer1984;53:342–5

8. Lefor AT, Bredenberg CE, Kellman RM, Aust JC. Multiple ma-lignancies of the lung and head and neck. Arch Surg 1986;121:265–70.

9. McCormack P. Surgical resection of pulmonary metastases. SeminSurg Oncol 1990;6:297–302.

10. Mazer TM, Robbins KT, McMurtrey MJ, Byers RM. Resection ofpulmonary metastases from squamous carcinoma of the head andneck. Am J Surg 1988;156:238–42.

11. Finley RK, Verazin GT, Driscoll DL, et al. Results of surgicalresection of pulmonary metastases of squamous cell carcinoma ofthe head and neck. Am J Surg 1992;164:594–8.

12. Wedman J, Balm AJM, Hart AAM, et al. Value of resection ofpulmonary metastases in head and neck cancer patients. Head Neck1996;18:311–6.

13. Schlumberger M, Challeton C, De Vathaire F, et al. Radioac-tive iodine treatment and external radiotherapy for lung and



bone metastases from thyroid carcinoma. J Nucl Med 1996;37:598–605.

14. Kim KH, Sung MW, Chung PS, et al. Adenoid cystic carcinoma ofthe head and neck. Arch Otolaryngol Head Neck Surg 1994;120:721–6.

15. The International Registry of Lung Metastases. Long-term resultsof lung metastasectomy: prognostic analyses based on 5206 cases.J Thorac Cardiovasc Surg 1997;113:37–49.

16. Shaha AR, Hoover EL, Mitrani M, et al. Synchronicity, multicen-tricity, and metachronicity of head and neck cancer. Head NeckSurg 1988;10:225–8.

17. Deleyiannis FW, Thomas DB. Risk of lung cancer among patients

with head and neck cancer. Otolaryngol Head Neck Surg 1997;116:630–6.

18. Vikram B, Strong EW, Shah JP, Spiro R. Second malignant neo-plasms in patients successfully treated with multimodality treatmentfor advanced head and neck cancer. Head Neck Surg 1984;6:734–7.

19. Sercarz J, Ellison D, Holmes EC, Calcaterra TC. Isolated pulmo-nary nodules in head and neck cancer patients. Ann Otol RhinolLaryngol 1989;98:113–8.

20. Chung KY, Mukhopadhyay T, Kim J, et al. Discordant p53 genemutations in primary head and neck cancers and correspondingsecond primary cancers of the upper aerodigestive tract. CancerRes 1993;53:1676–83.

578 D. LIU ET AL.


Documents

Pulmonary Metastasectomy for Head and Neck Cancers