Improved Survival After Resection of Pulmonarv Metastases From Malignant Melanoma Lyall A. Gorenstein, MD, Joe B. Putnam, Jr, MD, Giri Natarajan, MA, Charles A. Balch, MD, and Jack A. Roth, MD Departments of Thoracic Surgery and General Surgery, The University of Texas M.D. Anderson Cancer Center, Houston, Texas
The value of resecting pulmonary metastases from malignant melanoma was retrospectively examined. Between 1981 and 1989, 56 patients (35 men and 21 women with a mean age of 49 y-ars) had 65 pulmonary resections for histologically proven metastatic melanoma after treatment of the primary tumor. In patients undergoing thoracotomy, 50% (28/56) had pulmonary metastases as the initial site of recurrence. Twenty-eight patients (50%) had local-regional recurrence before the development of lung metastases. Eight lobectomies, two segmentectomies, and 55 wedge excisions were done. Fifty-four patients (54/56, 96%) underwent complete resection, and there were no operative deaths. The postthoracotomy actuarial survival was 25% at 5 years (median interval, 18
M
months). Location of the primary tumor, histology, thickness, Clark level, local-regional lymph node metastases, or type of resection was not associated with improved survival. Patients without regional nodal metastases before thoracotomy had a median survival of 30 months compared with 16 months for all others ( p = 0.04). Patients with lung as the site of first recurrence had a median survival of 30 months compared with 17 months for patients with initial local-regional recurrence ( p = 0.038, log-rank test). Despite systemic spread, patients with isolated pulmonary metastases from melanoma may benefit from metastasectomy. (Ann Thorac Surg 1991;52:20&10)
elanoma varies considerably in its biological behavior despite its tendency for wide dissemination involving many organ systems [l].Pulmonary metastases are the most common initial visceral site [2], and life expectancy, even with maximal adjuvant therapy, is shortened by respiratory failure [ 3 ] . Isolated lung metastases occur in 1.9% to 11%of patients with melanoma [l, 4, 51. Select patients with isolated pulmonary metastases from melanoma can achieve long-term survival after pulmonary resection.
Because of a lack of reliable preoperative prognostic criteria for survival after metastasectomy and because of the otherwise dismal prognosis for patients with melanoma, there is a general reluctance to resect pulmonary metastases from melanoma. We examined various prognostic indicators for their association with improved survival after pulmonary metastasectomy in patients with histologically confirmed melanoma.
For editorial comment, see page 178.
Between 1981 and 1989, 56 patients underwent 65 pulmonary resections for histologically proven metastatic melanoma at The University ofTexas-M.D. Anderson Cancer Center. Retrospective analysis of these patients was performed, and complete follow-up was obtained on all patients. There were 35 men and 21 women. All patients with a known history of malignant melanoma and suspected pulmonary metastases underwent complete metastatic evaluation before resection. This evaluation included plain chest roentgenograms, computed tomography of the chest and abdomen, bone scan, and pulmonary function studies. A physical examination excluded other sites of melanoma. Eight patients had pulmonary metastases from an unknown primary site. The location of the primary tumor was known in the remaining 48 patients. The Clark level of the original melanoma was known in 33 patients' The thickness and Of the primary tumor were known in 33 patients and 28 patients, respectively. Lung was the first site Of recurrence after treatment patients had Of the primary tumor in 28 patients. local-regional nodal recurrence before the development of
Most patients with isolated pulmonary metastases from melanoma are asymptomatic, and metastases are detected on routine surveillance roentgenograms after treatment of the primary tumor. Toxicity and relatively low response rate limit the value of systemic therapy, but resection of pulmonary metastases can be done with minimal morbidity and mortality [6-8].The precise population that benefits from resection of pulmonary metastases from melanoma is not known; however, various prognostic indicators can define a population in whom resection should be performed. These prognostic indicators may reflect the underlying biology of the individual and the tumor. Presented at the Thirty-seventh Annual Meeting of the Southern Thoracic Surgical Association, Dorado, Puerto Rico, Nov %lo, 1990. Address reprint requests to Dr Putnam, Department of Thoracic Surgery, The University of Texas M.D. Anderson Cancer Center, 1515 Holcombe Blvd, Box 109, Houston, TX 77030-4009.
0 1991 by The Society of Thoracic Surgeons
Material and Methods
-
0003-4975/91/$3.50
GORENSTEIN ET AL PULMONARY METASTASES FROM MELANOMA
Ann Thorac Surg 1991;52:20410
pulmonary metastases. Six patients had skin or subcutaneous recurrence; 1 patient had a solitary brain metastasis resected. Two patients had a solitary metastatic subcutaneous nodule removed before resection of the pulmonary metastases. Patients were considered for operation who were free from all extrathoracic metastatic disease and had undergone resection of the primary tumor. All patients had sufficient pulmonary reserve to tolerate the necessary resection. Double-lumen endotracheal tubes were used for deflation of lung parenchyma and to aid identification of all nodules by careful palpation of all pulmonary lobes. Most patients had the initial evaluation and treatment of the primary melanoma before referral to our institution. Twenty-seven patients had local excision alone, and 19 patients underwent local excision of the primary and dissection of the dominant lymph node drainage basin. Systemic therapy was often used before resection of metastases, as 25 patients received chemotherapy and 10 patients received immunotherapy. Factors analyzed for prognostic significance (their association with improved survival) included: site of primary tumor; Clark level, thickness, histology and stage of the primary tumor; site of first recurrence; tumor doubling time (TDT); disease-free interval; number of metastases resected at thoracotomy; number of thoracotomies; completeness of resection; type of excision of the primary tumor; and use of preoperative and postoperative adjuvant therapy. Survival curves were calculated using the method of Kaplan and Meier. Comparisons between groups were made using the generalized Wilcoxon test of Gehan and the log-rank test. All p values resulted from two-sided tests.
Results
Operative Findings Sixty-five operations were performed on 56 patients: 63 thoracotomies and two median sternotomies. Seven patients had bilateral metastases. In 4 of them, nodules developed in the contralateral lung after the initial resection, and the other 3 patients had bilateral staged thoracotomies for resection. Of the 126 nodules removed, 94 contained metastatic melanoma. The remainder consisted of a variety of benign pulmonary nodules. The single exception was 1 patient who was found to have a small bronchogenic carcinoma in addition to the metastasis. Operations performed were lobectomy (8/65, 12%), segmentectomy (2/65, 3%), and wedge resection (55/65, 85%). The majority of patients had disease confined to the pulmonary parenchyma. Invasion of parietal pleura, chest wall, diaphragm, or mediastinum was rare and identified in only seven operations (7/65, 11%).Mediastinal or hilar nodal metastases were found in 11 patients. However, a thorough mediastinal lymph node dissection was not performed in all patients. Complete resection was achieved in 97% (63/65)of explorations. Two patients had incomplete resection because of mediastinal adenopathy and a paraspinal mass.
205
TOTAL FAIL 56 37
0
0.20.1 0.0
f
0
10
20
30
40
50
60
70
80
90
MONTHS
Fig 1. Overall actuarial survival for patients with melanoma undergoing resection of pulmonary metastases. The median survival was 18 months.
No operative deaths occurred in the hospital or within 30 days of operation. Morbidity was minimal. Two patients remained in the hospital for longer than 7 days because of prolonged air leaks, and 1 patient had excessive chest drainage after thoracotomy.
Actuarial Survival The median follow-up is 40 months. The actuarial 5-year survival was 25% (median survival, 18 months) (Fig 1). Thirty-seven patients died of metastatic disease. All patients were followed up to death.
Prognostic lndicators SITE, HISTOLOGY, CLARK LEVEL, AND THICKNESS OF PRIMARY
The characteristics and stage of the primary melanoma are shown in Table 1. The location of the primary tumor was known in 48 patients. Eight patients had pulmonary metastases and no evidence of a primary tumor. The distribution was equal between head and neck, trunk, and extremities. There was no survival advantage after complete resection of pulmonary metastases when the primary tumor arose on an extremity versus the trunk or the head and neck region (Table 2). The histological growth pattern was known in 28 patients. Those with nodular melanomas did not have a different postthoracotomy survival than those with superficial spreading melanomas ( p = 0.74). The Clark level of invasion could be determined in 33 patients; only 1 patient had a level I melanoma, and no patients had level I1 tumors. Thirteen patients had a level I11 melanoma, 15 TUMOR.
206
GORENSTEIN ET AL PULMONARY METASTASES FROM MELANOMA
Ann Thorac Surg 1991;52:20&10
had level IV tumors, and 4 had level V tumors. There was no significant difference in postthoracotomy survival for patients with level I11 melanoma compared with patients with levels IV and V melanoma. Tumor thickness was known for 33 patients. Three had thin tumors (thickness < 0.76 mm), 20 had tumors of intermediate thickness (0.76 to 3.99 mm), and 9 had tumor thickness greater than 4.0 mm. No survival difference was identified among these groups. Staging was performed using the criteria of the American Joint Committee on Cancer. Stage assignment was determined by the highest stage reached before thoracotomy; patients in whom localregional nodal metastases developed were classified as in
STAGE OF PRIMARY TUMOR.
Table 2. Prognostic Factors Examined for Improved Survival After Resection of Pulmonary Metastases From Melanoma No. of
Factor
Patients
Histology of primary
ss N
Thickness of primary (mm)
50.76 >0.76
Characteris tic Location Extremity Trunk Head and neck Mucosal Unknown
15 (27) 16 (29) 14 (25) 3 (5) 8 (14)
Clark level
I I1 111
IV V Unknown Thickness (mm) c1.49 1.5-2.49 2.5-3.99
>4.00 Unknown Histology LM
ss
AL
N Unknown Stage before thoracotomyb
I I1 I11 IV Unknown
1 (2) 0 13 (23) 15 (27)
4 (7) 23 (41) 5 (9) 8 (14) 11 (20) 9 (16) 23 (41)
111
IV and V
0.9 13 19
Location of primary
Head and neck Excision of primary WLE WLE
+ LND
0.9 15 16 14 0.2
26 19 0.04
Stage before thoracotomy
I and I1 I11 Site of initial recurrence Lung Other Tumor doubling time (d)
40 Intrathoracic nodal metastases Absent Present No. of metastases resected tl
>1
26 25 0.07b
28 28 0.1 4 17 0.9 42 12 0.07‘ 36
29
Side of metastases Unilateral Bilateral
0.4 49 7 0.9
No. of thoracotomies 1 (2) 15 (27) 1 (2) 11 (20) 28 (50)
3 (5) 23 (41) 25 (45) 3 (5) 2 (4)
This is Numbers in parentheses are percentages of total population. the highest stage achieved before metastasectomy. LM = lentigo maligna; N = nodular; AL = acryl lentiginous; SS = superficial spreading. a
0.1 3 29
Clark level of primary
Trunk No. of Patients”
0.4 15 11
Extremity
Table I . Characteristics of Primary Tumor
P
Value”
1 >1 Pulmonary resection Lobectomy/segmentectomy Wedge
51 5 0.7
10
55
Significance: p = A p value of
The value of resecting pulmonary metastases from malignant melanoma was retrospectively examined. Between 1981 and 1989, 56 patients (35 men and 21 women with a mean age of 49 y-ars) had 65 pulmonary resections for histologically proven metastatic melanoma after treatment of the primary tumor. In patients undergoing thoracotomy, 50% (28/56) had pulmonary metastases as the initial site of recurrence. Twenty-eight patients (50%) had local-regional recurrence before the development of lung metastases. Eight lobectomies, two segmentectomies, and 55 wedge excisions were done. Fifty-four patients (54/56, 96%) underwent complete resection, and there were no operative deaths. The postthoracotomy actuarial survival was 25% at 5 years (median interval, 18
M
months). Location of the primary tumor, histology, thickness, Clark level, local-regional lymph node metastases, or type of resection was not associated with improved survival. Patients without regional nodal metastases before thoracotomy had a median survival of 30 months compared with 16 months for all others ( p = 0.04). Patients with lung as the site of first recurrence had a median survival of 30 months compared with 17 months for patients with initial local-regional recurrence ( p = 0.038, log-rank test). Despite systemic spread, patients with isolated pulmonary metastases from melanoma may benefit from metastasectomy. (Ann Thorac Surg 1991;52:20&10)
elanoma varies considerably in its biological behavior despite its tendency for wide dissemination involving many organ systems [l].Pulmonary metastases are the most common initial visceral site [2], and life expectancy, even with maximal adjuvant therapy, is shortened by respiratory failure [ 3 ] . Isolated lung metastases occur in 1.9% to 11%of patients with melanoma [l, 4, 51. Select patients with isolated pulmonary metastases from melanoma can achieve long-term survival after pulmonary resection.
Because of a lack of reliable preoperative prognostic criteria for survival after metastasectomy and because of the otherwise dismal prognosis for patients with melanoma, there is a general reluctance to resect pulmonary metastases from melanoma. We examined various prognostic indicators for their association with improved survival after pulmonary metastasectomy in patients with histologically confirmed melanoma.
For editorial comment, see page 178.
Between 1981 and 1989, 56 patients underwent 65 pulmonary resections for histologically proven metastatic melanoma at The University ofTexas-M.D. Anderson Cancer Center. Retrospective analysis of these patients was performed, and complete follow-up was obtained on all patients. There were 35 men and 21 women. All patients with a known history of malignant melanoma and suspected pulmonary metastases underwent complete metastatic evaluation before resection. This evaluation included plain chest roentgenograms, computed tomography of the chest and abdomen, bone scan, and pulmonary function studies. A physical examination excluded other sites of melanoma. Eight patients had pulmonary metastases from an unknown primary site. The location of the primary tumor was known in the remaining 48 patients. The Clark level of the original melanoma was known in 33 patients' The thickness and Of the primary tumor were known in 33 patients and 28 patients, respectively. Lung was the first site Of recurrence after treatment patients had Of the primary tumor in 28 patients. local-regional nodal recurrence before the development of
Most patients with isolated pulmonary metastases from melanoma are asymptomatic, and metastases are detected on routine surveillance roentgenograms after treatment of the primary tumor. Toxicity and relatively low response rate limit the value of systemic therapy, but resection of pulmonary metastases can be done with minimal morbidity and mortality [6-8].The precise population that benefits from resection of pulmonary metastases from melanoma is not known; however, various prognostic indicators can define a population in whom resection should be performed. These prognostic indicators may reflect the underlying biology of the individual and the tumor. Presented at the Thirty-seventh Annual Meeting of the Southern Thoracic Surgical Association, Dorado, Puerto Rico, Nov %lo, 1990. Address reprint requests to Dr Putnam, Department of Thoracic Surgery, The University of Texas M.D. Anderson Cancer Center, 1515 Holcombe Blvd, Box 109, Houston, TX 77030-4009.
0 1991 by The Society of Thoracic Surgeons
Material and Methods
-
0003-4975/91/$3.50
GORENSTEIN ET AL PULMONARY METASTASES FROM MELANOMA
Ann Thorac Surg 1991;52:20410
pulmonary metastases. Six patients had skin or subcutaneous recurrence; 1 patient had a solitary brain metastasis resected. Two patients had a solitary metastatic subcutaneous nodule removed before resection of the pulmonary metastases. Patients were considered for operation who were free from all extrathoracic metastatic disease and had undergone resection of the primary tumor. All patients had sufficient pulmonary reserve to tolerate the necessary resection. Double-lumen endotracheal tubes were used for deflation of lung parenchyma and to aid identification of all nodules by careful palpation of all pulmonary lobes. Most patients had the initial evaluation and treatment of the primary melanoma before referral to our institution. Twenty-seven patients had local excision alone, and 19 patients underwent local excision of the primary and dissection of the dominant lymph node drainage basin. Systemic therapy was often used before resection of metastases, as 25 patients received chemotherapy and 10 patients received immunotherapy. Factors analyzed for prognostic significance (their association with improved survival) included: site of primary tumor; Clark level, thickness, histology and stage of the primary tumor; site of first recurrence; tumor doubling time (TDT); disease-free interval; number of metastases resected at thoracotomy; number of thoracotomies; completeness of resection; type of excision of the primary tumor; and use of preoperative and postoperative adjuvant therapy. Survival curves were calculated using the method of Kaplan and Meier. Comparisons between groups were made using the generalized Wilcoxon test of Gehan and the log-rank test. All p values resulted from two-sided tests.
Results
Operative Findings Sixty-five operations were performed on 56 patients: 63 thoracotomies and two median sternotomies. Seven patients had bilateral metastases. In 4 of them, nodules developed in the contralateral lung after the initial resection, and the other 3 patients had bilateral staged thoracotomies for resection. Of the 126 nodules removed, 94 contained metastatic melanoma. The remainder consisted of a variety of benign pulmonary nodules. The single exception was 1 patient who was found to have a small bronchogenic carcinoma in addition to the metastasis. Operations performed were lobectomy (8/65, 12%), segmentectomy (2/65, 3%), and wedge resection (55/65, 85%). The majority of patients had disease confined to the pulmonary parenchyma. Invasion of parietal pleura, chest wall, diaphragm, or mediastinum was rare and identified in only seven operations (7/65, 11%).Mediastinal or hilar nodal metastases were found in 11 patients. However, a thorough mediastinal lymph node dissection was not performed in all patients. Complete resection was achieved in 97% (63/65)of explorations. Two patients had incomplete resection because of mediastinal adenopathy and a paraspinal mass.
205
TOTAL FAIL 56 37
0
0.20.1 0.0
f
0
10
20
30
40
50
60
70
80
90
MONTHS
Fig 1. Overall actuarial survival for patients with melanoma undergoing resection of pulmonary metastases. The median survival was 18 months.
No operative deaths occurred in the hospital or within 30 days of operation. Morbidity was minimal. Two patients remained in the hospital for longer than 7 days because of prolonged air leaks, and 1 patient had excessive chest drainage after thoracotomy.
Actuarial Survival The median follow-up is 40 months. The actuarial 5-year survival was 25% (median survival, 18 months) (Fig 1). Thirty-seven patients died of metastatic disease. All patients were followed up to death.
Prognostic lndicators SITE, HISTOLOGY, CLARK LEVEL, AND THICKNESS OF PRIMARY
The characteristics and stage of the primary melanoma are shown in Table 1. The location of the primary tumor was known in 48 patients. Eight patients had pulmonary metastases and no evidence of a primary tumor. The distribution was equal between head and neck, trunk, and extremities. There was no survival advantage after complete resection of pulmonary metastases when the primary tumor arose on an extremity versus the trunk or the head and neck region (Table 2). The histological growth pattern was known in 28 patients. Those with nodular melanomas did not have a different postthoracotomy survival than those with superficial spreading melanomas ( p = 0.74). The Clark level of invasion could be determined in 33 patients; only 1 patient had a level I melanoma, and no patients had level I1 tumors. Thirteen patients had a level I11 melanoma, 15 TUMOR.
206
GORENSTEIN ET AL PULMONARY METASTASES FROM MELANOMA
Ann Thorac Surg 1991;52:20&10
had level IV tumors, and 4 had level V tumors. There was no significant difference in postthoracotomy survival for patients with level I11 melanoma compared with patients with levels IV and V melanoma. Tumor thickness was known for 33 patients. Three had thin tumors (thickness < 0.76 mm), 20 had tumors of intermediate thickness (0.76 to 3.99 mm), and 9 had tumor thickness greater than 4.0 mm. No survival difference was identified among these groups. Staging was performed using the criteria of the American Joint Committee on Cancer. Stage assignment was determined by the highest stage reached before thoracotomy; patients in whom localregional nodal metastases developed were classified as in
STAGE OF PRIMARY TUMOR.
Table 2. Prognostic Factors Examined for Improved Survival After Resection of Pulmonary Metastases From Melanoma No. of
Factor
Patients
Histology of primary
ss N
Thickness of primary (mm)
50.76 >0.76
Characteris tic Location Extremity Trunk Head and neck Mucosal Unknown
15 (27) 16 (29) 14 (25) 3 (5) 8 (14)
Clark level
I I1 111
IV V Unknown Thickness (mm) c1.49 1.5-2.49 2.5-3.99
>4.00 Unknown Histology LM
ss
AL
N Unknown Stage before thoracotomyb
I I1 I11 IV Unknown
1 (2) 0 13 (23) 15 (27)
4 (7) 23 (41) 5 (9) 8 (14) 11 (20) 9 (16) 23 (41)
111
IV and V
0.9 13 19
Location of primary
Head and neck Excision of primary WLE WLE
+ LND
0.9 15 16 14 0.2
26 19 0.04
Stage before thoracotomy
I and I1 I11 Site of initial recurrence Lung Other Tumor doubling time (d)
40 Intrathoracic nodal metastases Absent Present No. of metastases resected tl
>1
26 25 0.07b
28 28 0.1 4 17 0.9 42 12 0.07‘ 36
29
Side of metastases Unilateral Bilateral
0.4 49 7 0.9
No. of thoracotomies 1 (2) 15 (27) 1 (2) 11 (20) 28 (50)
3 (5) 23 (41) 25 (45) 3 (5) 2 (4)
This is Numbers in parentheses are percentages of total population. the highest stage achieved before metastasectomy. LM = lentigo maligna; N = nodular; AL = acryl lentiginous; SS = superficial spreading. a
0.1 3 29
Clark level of primary
Trunk No. of Patients”
0.4 15 11
Extremity
Table I . Characteristics of Primary Tumor
P
Value”
1 >1 Pulmonary resection Lobectomy/segmentectomy Wedge
51 5 0.7
10
55
Significance: p = A p value of
