Long-Term Survival After Resection of Pulmonary Metastases From Carcinoma of the Breast Louis A. Lanza, MD, Giri Natarajan, MA, Jack A. Roth, MD, and Joe B. Putnam, Jr, MD Department of Thoracic Surgery, The University of Texas MD Anderson Cancer Center, Houston, Texas
Resection of isolated pulmonary metastases may yield improved survival in select patients. Between 1981 and 1991, 44 women (median age, 55 years) with a history of breast cancer underwent 47 thoracotomies with no operative deaths and only three minor postoperative complications (3/47,6.4%). Confirmation of the metastatic origin of the lung lesion was made by direct histological comparison with the primary. Three patients had benign nodules and were excluded, and 4 patients had less than complete resection at thoracotomy. The median survival after thoracotomy of the remaining 37 patients with completely resected metastases was 47 f 5.5 months, and their actuarial 5-year survival was 49.5%. Patients with a disease-free interval of longer than 12 months had a longer survival (median survival, 82 r?: 6 months; 5-year
B
reast cancer, with approximately 100,000 new cases per year, is the most prevalent cancer among women in the United States. Accounting for approximately 30,000 deaths per year, it is second only to lung cancer as a cause of cancer-related deaths in American women [l].It has been estimated that breast cancer will develop in 1 in 14 women in the United States during her lifetime. As the treatment of the primary breast lesion has evolved to less radical procedures, the treatment of systemic disease has become more complex, with availability of combination chemotherapy and antihormonal agents. The majority of patients with metastatic breast cancer will have either disseminated disease or metastases in unresectable sites. Approximately 21% of patients will die of isolated metastases to the lung that are potentially resectable sometime during the course of the disease [2]. Since 1939, when Barney and Churchill performed the first intentional resection of pulmonary metastases resulting in long-term survival [3], Ehrenhaft and colleagues [4, 51 and others [&8] have defined and expanded the applications. Although the role of pulmonary metastasectomy has been established in the treatment of patients with osteogenic and soft tissue sarcomas, and prognostic factors have been defined [9-121, its role in the treatment of patients with carcinomatous metastases is less clear. Most reports contain either a mixture of cancer histologies Presented at the Thirty-eighth Annual Meeting of the Southern Thoracic Surgical Association, Orlando, FL, Nov 7-9, 1991. Address reprint requests to Dr Putnam, Department of Thoracic Surgery, The University of Texas MD Anderson Cancer Center, 1515 Holcombe Blvd, Box 109, Houston, TX 77030-4009.
0 1992 by The Society of Thoracic Surgeons
survival, 57%) than patients with a disease-free interval of 12 months or less (median survival, 15 2 3.6 months; 5-year survival, 0%) ( p = 0.004). Patients with estrogen receptor-positive status (n = 14) tended to have longer survival after resection than patients with estrogen receptor-negative status (n = 15) (median survival, 81 & 9 months versus 23 f 6 months, respectively; p = 0.098). Other clinical variables analyzed did not predict survival after thoracotomy. We conclude that resection of pulmonary metastases in patients with breast cancer can be done safely and may result in long-term survival for a substantial number of patients. Patients with a diseasefree interval of longer than 12 months have an excellent prognosis after complete resection. (Ann Thorac Surg 1992;54:244-8)
or a small number of patients, making evaluation of outcome difficult. In addition, factors that may affect outcome for individual histologies are difficult to determine from available data. In this report, we have retrospectively examined survival and prognostic factors in patients undergoing resection of isolated pulmonary metastases from breast carcinoma. In this select population, long-term survival was achieved in a substantial number of patients. In addition, a prolonged disease-free interval (DFI) was found to be associated with a prolonged postthoracotomy survival after complete resection of metastases.
Material and Methods Patient Population Forty-four patients were retrospectively identified from the patient records of The University of Texas MD Anderson Cancer Center with the diagnosis of breast cancer who underwent thoracotomy for resection of pulmonary nodules between 1981 and 1990. Excluded were patients who underwent thoracotorny for clinically obvious primary lung cancer. The determination of the metastatic origin of the lung lesion was made only by histologic similarity to the primary breast cancer with light microsCOPY. All 44 patients were women. Their median age was 55 years (range, 32 to 79 years). Twenty-two patients underwent treatment of the primary breast cancer at our institution, and 22 underwent treatment elsewhere. All patients had adenocarcinoma of the breast. No patient was lost to follow-up. 0003-4975/92/$5.00
245
LANZA ET AL PULMONARY METASTASES FROM BREAST CARCINOMA
Ann Thorac Surg 1992;54:24&8
Chemotherapy regimens varied during the period of study. Forty of the 44 patients received chemotherapy, tamoxifen, or both at some time during their clinical course. Eight of the 37 patients with resectable disease received treatment before thoracotomy, 12 patients received treatment after thoracotomy, and 15 patients were treated both before and after thoracotomy; 2 patients did not receive treatment. Three patients with unresectable disease received tamoxifen only.
1.0 tb?
-ap
-b 0
a
e
n
0.9
-
0.8
-
Total Fall 37 18
0
0.7-
'
0.6-
-
0.5
.
0.4 -
0.3 -
Stat is t ical Analyses
0.2
-
Postoperative survival and disease-free survival were calculated using the product-limit method of Kaplan and Meier [13]. Statistical comparisons between groups were made using the log-rank test with survival calculated from the date of thoracotomy. All p values resulted from two-sided tests.
0.1
-
0.0 7
Results The 44 patients underwent a total of 47 thoracotomies. Forty-one patients had a single thoracotomy and 3, bilateral staged thoracotomies for resection of disease. Two patients were explored through median sternotomies, and the remaining 42 underwent exploration through lateral thoracotomies. There were no postoperative deaths and only three minor complications consisting of prolonged air leaks for the 47 thoracotomies (3/47, 6.4%). Forty-one of the 47 resections were wedge resections, four were lobectomies, one was an anatomical segmentectomy, and one was a pneumonectomy. The overall median survival for the entire cohort of 44 5 months. Three (6.8%) of the 44 patients was 82 patients were found at thoracotomy to have benign disease only (Histoplasma, 1, and Coccidioides, 2). These 3 patients were excluded from further survival analysis. All 3 were disease-free 1 month, 17 months, and 51 months after thoracotomy. Four patients underwent thoracotomy without complete resection of disease. Two of them were found to have mediastinal lymph node involvement that precluded complete resection, 1 was found to have hilar disease requiring pneumonectomy and underwent biopsy only, and 1 had clinically evident bilateral disease and underwent unilateral biopsy only. Three of the 4 patients died 2, 23, and 26 months after thoracotomy, and 1 remains alive with disease 40 months after thoracotomy. The remaining 37 patients underwent complete resection of all clinically evident disease at thoracotomy. The median survival of these patients was 47 c 5.5 months, and the actuarial 5-year survival was 49.5% (Fig 1).Eleven (30%) of the 37 patients remain clinically disease-free a mean of 49.2 months after thoracotomy (range, 1 to 88 months). Eight (22%) of the 37 patients remain alive but with disease a mean of 67 months (range, 32 to 102 months) after thoracotomy. Eighteen (49%) of the 37 patients are dead a mean of 21 months (range, 6 to 82 months) after thoracotomy. Of the 26 patients in whom disease recurred after complete resection, only 4 experienced recurrence in the lung, whereas 22 patients expe-
*
rienced either disseminated recurrences or recurrence in locations not amenable to resection. No patient underwent attempted resection of extrathoracic recurrent disease after initial thoracotomy. Clinical variables were analyzed to attempt to identify prognostic factors that may predict prolonged survival after thoracotomy. Thirty-two of the 37 patients undergoing complete resection of histologically proven malignant disease were found to have a DFI (time from treatment of primary to development of clinically evident metastases) of more than 1 year, and 5 patients were found to have a DFI of 1 year or less. The median survival of the 32 6 patients with a DFI of more than 1 year was 82 months, and their actuarial 5-year survival was 57%. In contrast, the median survival of the 5 patients with a DFI of 1 year or less was only 15 3.6 months, and all patients were dead at 26 months ( p = 0.004) (Fig 2). Similarly, the postthoracotomy disease-free survival was significantly prolonged in the patients with an initial DFI of greater than 1 year. The median disease-free survival of the 32
*
*
1.o 0.9
0.8 0.7
E
.-
0.6
bn
0.5
e
0.4
0.3 0.2 0.1 0.0
Total
1
4
0
Fall
A 32 14r12Month 0
P2
10
20
30
40
50
5
=
5rl2Month .004
60
70
80
90
100
110
Months
Fig 2. lnitial disease-free interval. Preoperative disease-free internal is associated with prolonged postthoracotomy survival (p = 0.004).
246
LANZA ET AL PULMONARY METASTASES FROM BREAST CARCINOMA
Ann Thorac Surg 1992;54:2448
O.gB\ 0.8
Total
Fail
A 32
0 Y
22 DFI> 12 Month 5 4 D F l i 12Month
0
z L
n 0
e
n
0.4-
0.3 -
I:] 0.0
, 0
10
1, 20
,
,
,
,
,
,
,
,
,
30
40
50
60
70
80
90
100
110
Months
Fig 3 . Postthoracotomy disease-free survival. Preoperative disease-free interval (DFI) predicts postthoracotomy disease-free survival (p = 0.046). (NED = no evidence of disease.)
patients with a DFI of more than 1 year was 30 k 3.5 months, and their actuarial 5-year disease-free survival was 29%. The 5 patients with a DFI of 1 year or less had a median disease-free survival of only 11 3.6 months, and all patients had recurrence by 17 months after thoracotomy ( p = 0.046) (Fig 3). Twenty-nine of the 37 patients had a known estrogen receptor (ER) status; in 8 patients, the ER status could not be determined. In 14 of the 29 patients, the ER status was positive and in 15 patients, negative. The median survival 9 of the 14 patients who were ER positive was 81 months, and their actuarial 5-year survival was 66%. The median survival of the 15 patients who were ER negative was 23 6 months, and their actuarial 5-year survival was 28%. There was a trend toward improved postthoracotomy survival in the patients with a positive ER status ( p = 0.098) (Fig 4). Twenty-seven of the 37 patients underwent resection of
*
*
*
1.0
1
1 0.891
.-a0 C
0
0.7
-
0.6
-
0.5-
L
g
0
n
0.4
~
Total
0.3 -
O 0.1
I
0.0
4 0
0 14
10
20
30
40
50
60
Fall 6 ERP(+)
70
80
90
100
Months
Fig 4 . Estrogen receptor (ERE') status. Estrogen receptor status does not predict postthoracotomy survival (p = 0.098). (DFI = diseusefree interval; NED = no evidence of diseuse.)
a single malignant pulmonary nodule, whereas 10 patients underwent resection of two or more malignant nodules (mean, 7.2; range, 2 to 49; and median, 2). The median postthoracotomy survival for the 27 patients undergoing resection of a single pulmonary nodule was 82 ? 7.5 months, and their actuarial 5-year survival was 59%. The median postthoracotomy survival for the 10 patients undergoing resection of two or more malignant nodules was 24 5 3 months, and their actuarial 5-year survival was 36%. A slight survival advantage was noted for patients undergoing resection of a single pulmonary nodule, but the survival difference was not significant ( p = 0.229). Combining these two clinical variables did not improve their predictive power. Nine patients were identified who had a single pulmonary nodule and also were ER positive. The actuarial survival after thoracotomy was not significantly different for this group compared with the remaining 28 patients ( p = 0.548). Seventeen of the 37 patients had positive axillary lymph nodes at the time of mastectomy, and 18 had negative axillary lymph nodes. The postthoracotomy survival did not differ between these groups ( p = 0.650).
Comment Although most patients with breast cancer will have disseminated disease, as many as 21% die of metastatic disease limited to the lungs [2]. Few data exist to determine whether these patients should undergo pulmonary metastasectomy, and if so, which patients can be expected to obtain the longest survival benefit from this procedure. Mountain and colleagues [14] reviewed the cases of a previous group of patients treated at MD Anderson Cancer Center from 1957 through 1975. One hundred fiftynine patients underwent resection of carcinomatous pulmonary metastases during that 19-year period. Of the 159 patients, 21 underwent resection of metastases from breast carcinoma. The cumulative 5-year survival of this group was 14%, and the median survival as calculated by the method of Gehan was 27 months. In fact, among the various carcinomatous histologies that were analyzed, breast cancer resulted in the lowest cumulative 5-year survival. Differences between that experience and our results may reflect advances in chemotherapy and hormonal therapy of systemic disease during the past two decades. Prognostic factors were analyzed for all histologies combined [14]. Patients with a DFI of 1 year or less had the highest cumulative 5-year survival (37%), whereas those with longer DFIs had lower 5-year survival. Analysis of prognostic factors for mixed carcinomatous histologies is difficult to interpret. McCormack and Martini [15] reviewed their 17-year experience with pulmonary metastasectomy at Memorial Sloan-Kettering Cancer Center. During that 17-year period, 663 resections for pulmonary metastases were performed in 448 patients, 246 of whom had carcinomatous histologies. Of these 246 patients, 34 had breast carcinoma. The median postthoracotomy survival of the 34 patients was approximately 33 months, and the cumula-
Ann Thorac Surg 1992;54:244-8
tive 5-year survival was 30%. Prognostic factors were not analyzed in that group of patients. Wright and co-workers [5] reviewed the experience with pulmonary metastasectomy at The University of Iowa over a 22-year period. One hundred forty-two patients underwent 153 thoracotomies with an operative mortality of 0.7%. One hundred patients had carcinomatous histologies, and of them, 18 had carcinoma of the breast. Wright and co-workers found no significant difference in postthoracotomy survival between patients with sarcomatous and patients with carcinomatous metastases. The actuarial 5-year survival for the 18 patients with breast carcinoma was 27%. The preoperative DFI did not predict postthoracotomy survival in their analyses. In this series, we report a 49.5% actuarial 5-year survival and a median survival of 47 k 5.5 months for this select group of patients with metastatic breast cancer confined to the lungs who underwent complete resection of the metastatic disease. There were no postoperative deaths, and only 3 patients experienced minor complications consisting of prolonged air leaks, none of which required reoperation. Although it is not possible to conclude from these or other retrospective data that pulmonary metastasectomy in patients with breast cancer prolongs survival, the data do suggest that pulmonary metastasectomy can be done safely in this group of patients and may result in long-term survival for a substantial number. A longer DFI is associated with a longer postthoracotomy survival in this select population. This observation has been made previously for other histologies [lo, 161. Whether this represents a favorable tumor biology, a favorable tumor-host relationship, or other unknown factors is impossible to determine. A trend toward prolonged survival in patients with ER-positive tumor was evident, but the survival benefit was not significant. However, it may be that ER status is less influential in predicting survival in patients with systemic disease that is completely resected. Chemotherapy regimens varied during the period of study. Thirty-five of the 37 patients with completely resected metastases received chemotherapy, tamoxifen, or both at some time during their clinical course. Two patients with resected metastases received no chemotherapy. Eight patients received systemic treatment before thoracotomy, 12 patients were treated after thoracotomy, and 15 patients were treated both before and after thoracotomy. Three of these 35 patients received tamoxifen alone. Three patients with unresectable disease received tamoxifen only. The efficacy of systemic chemotherapy or hormonal therapy in prolonging survival in this group of patients cannot be determined. No other analyzed variable demonstrated predictive power for postthoracotomy survival. For other histologies, similar variables have been found useful in predict-
LANZA ET AL PULMONARY METASTASES FROM BREAST CARCINOMA
247
ing postoperative survival 1171. Thus, it is evident that prognostic factors are best determined for homogeneous groups of patients with closely matched histologies and should not be generalized for divergent groups. In conclusion, resection of pulmonary metastases from carcinoma of the breast may result in long-term survival for select patients. Patients with a DFI of greater than 1 year have an excellent prognosis after resection of pulmonary metastases, and more than 50% may survive beyond 5 years. The procedure can be performed with no mortality and only minimal morbidity. Therefore, it represents a viable option for the treatment of patients with metastatic breast cancer confined to the lung.
References 1. Harris JR, Hellman S, Canellas GP, Fisher B. Cancer of the breast. In: DeVita VT, Hellman S, Rosenberg SA, eds. Cancer: principles and practice of oncology. 2nd ed. Philadelphia: JB Lippincott, 1985:1119-77. 2. Ramming KP. Surgery for pulmonary metastases. Surg Clin North Am 1980;60:81524. 3. Barney JD, Churchill EJ. Adenocarcinoma of the kidney with metastases to the lungs cured by nephrectomy and lobectomy. J Urol 1958;42:269-71. 4. Ehrenhaft JL, Lawrence MS, Sensenig DM. Pulmonary resection for metastatic lesions. Arch Surg 1958;77606-12. 5. Wright JO, Brandt B, Ehrenhaft JL. Results of pulmonary resection for metastatic lesions. J Thorac Cardiovasc Surg 1982;83:94-9. 6. Alexander J, Haight C. Pulmonary resection for solitary metastatic sarcomas and carcinomas. Surg Gynecol Obstet 1947;85:129. 7. Marks P, Ferrag MZ, Ashof H. Rationale for the surgical treatment of pulmonary metastases. Thorax 1981;36:679-82. 8. Morton DL, Joseph WL, Ketcham AS, et al. Surgical resection and adjunctive immunotherapy for selected patients with multiple pulmonary metastases. Ann Surg 1973;178:360-6. 9. Takita H. Management of pulmonary metastases from osteogenic sarcoma. Ann Thorac Surg 1983;36:508. 10. Creagan ET, Flemming TR, Edmondson JH, Pairolero PC. Pulmonary resection for metastatic non-osteogenic sarcoma. Cancer 1979;44:1908-12. 11. Putnam JB Jr, Roth JA, Wesley MN, et al. Analysis of prognostic factors in patients undergoing resection of pulmonary metastases from soft tissue sarcomas. J Thorac Cardiovasc Surg 1984;87:260-8. 12. Putnam JB Jr, Roth JA, Wesley MN, Johnston MR, Rosenberg SA. Survival following aggressive resection of pulmonary metastases from osteogenic sarcoma: analysis of prognostic factors. Ann Thorac Surg 1983;36:51&23. 13. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc 1958;53:457-81. 14. Mountain CF, Khalil KG, Hermes KE, Frazier OH. The contribution of surgery to the management of carcinomatous pulmonary metastases. Cancer 1978;41:83340. 15. McCormack I'M, Martini N. The changing role of surgery for pulmonary metastases. Ann Thorac Surg 1979;28:13945. 16. Huth JF, Holmes EC, Vernon SE, et al. Pulmonary resection for metastatic sarcoma. Am J Surg 1980;140:9-16. 17. Roth JA, Putnam JB Jr, Wesley MN, et al. Differing determinants of prognosis following resection of pulmonary metastases from osteogenic and soft tissue sarcoma patients. Cancer 1985;55:1361-6.
248
LANZA ET AL PULMONARY METASTASES FROM BREAST CARCINOMA
A n n Thorac Surg 1992;54:24&8
DISCUSSION DR THOMAS M. EGAN (Chapel Hill, NC): I enjoyed your provocative paper. Do you have data on 5-year survival of patients with breast cancer and a solitary metastasis other than to the lung? In other words, have you been able to demonstrate that your surgical intervention contributed to that 5-year survival? DR LANZA: No, we do not have those data, and they are difficult to obtain from patient records. DR HARVEY I. PASS (Bethesda, MD): My colleagues and I prefer a median sternotomy for resection of metastases. Are there select populations of patients with metastases for whom you do
not think a sternotomy is necessary, or should they all have median sternotomies as opposed to lateral thoracotomies? DR LANZA: Median sternotomies are most helpful when bilateral metastases are present or when bilateral metastases are susuected. as in sarcomas. Patients with metastases in one hemithorax did not frequently have new metastases occur in the other hemithorax during follow-up. Most patients did well with a single lateral thoracotomy. As to whether median sternotomies are tKe preferred route forexploration, I would leave that to the individual patient and surgeon, as number of metastases, location, familiarity with the approach, and histology all weigh in the final decision.
Notice From The Society of Thoracic Surgeons New Abstract Deadline Date The abstract due date is Monday, August 17, 1992, for all papers submitted for the Twenty-ninth Annual Meeting of The Society of Thoracic Surgeons, which will be held at the San Antonio Convention Center in San Antonio, Texas, January 2.527, 1993. No abstracts received after this date
will be considered for presentation. The deadline for receipt of films and video tapes is Monday, August 3, 1992. Details are available in the Call for Abstracts published in this issue of The Annals or from Society headquarters, 401 N Michigan Avenue, Chicago, IL 60611 (telephone: (312) 644-6610).
Resection of isolated pulmonary metastases may yield improved survival in select patients. Between 1981 and 1991, 44 women (median age, 55 years) with a history of breast cancer underwent 47 thoracotomies with no operative deaths and only three minor postoperative complications (3/47,6.4%). Confirmation of the metastatic origin of the lung lesion was made by direct histological comparison with the primary. Three patients had benign nodules and were excluded, and 4 patients had less than complete resection at thoracotomy. The median survival after thoracotomy of the remaining 37 patients with completely resected metastases was 47 f 5.5 months, and their actuarial 5-year survival was 49.5%. Patients with a disease-free interval of longer than 12 months had a longer survival (median survival, 82 r?: 6 months; 5-year
B
reast cancer, with approximately 100,000 new cases per year, is the most prevalent cancer among women in the United States. Accounting for approximately 30,000 deaths per year, it is second only to lung cancer as a cause of cancer-related deaths in American women [l].It has been estimated that breast cancer will develop in 1 in 14 women in the United States during her lifetime. As the treatment of the primary breast lesion has evolved to less radical procedures, the treatment of systemic disease has become more complex, with availability of combination chemotherapy and antihormonal agents. The majority of patients with metastatic breast cancer will have either disseminated disease or metastases in unresectable sites. Approximately 21% of patients will die of isolated metastases to the lung that are potentially resectable sometime during the course of the disease [2]. Since 1939, when Barney and Churchill performed the first intentional resection of pulmonary metastases resulting in long-term survival [3], Ehrenhaft and colleagues [4, 51 and others [&8] have defined and expanded the applications. Although the role of pulmonary metastasectomy has been established in the treatment of patients with osteogenic and soft tissue sarcomas, and prognostic factors have been defined [9-121, its role in the treatment of patients with carcinomatous metastases is less clear. Most reports contain either a mixture of cancer histologies Presented at the Thirty-eighth Annual Meeting of the Southern Thoracic Surgical Association, Orlando, FL, Nov 7-9, 1991. Address reprint requests to Dr Putnam, Department of Thoracic Surgery, The University of Texas MD Anderson Cancer Center, 1515 Holcombe Blvd, Box 109, Houston, TX 77030-4009.
0 1992 by The Society of Thoracic Surgeons
survival, 57%) than patients with a disease-free interval of 12 months or less (median survival, 15 2 3.6 months; 5-year survival, 0%) ( p = 0.004). Patients with estrogen receptor-positive status (n = 14) tended to have longer survival after resection than patients with estrogen receptor-negative status (n = 15) (median survival, 81 & 9 months versus 23 f 6 months, respectively; p = 0.098). Other clinical variables analyzed did not predict survival after thoracotomy. We conclude that resection of pulmonary metastases in patients with breast cancer can be done safely and may result in long-term survival for a substantial number of patients. Patients with a diseasefree interval of longer than 12 months have an excellent prognosis after complete resection. (Ann Thorac Surg 1992;54:244-8)
or a small number of patients, making evaluation of outcome difficult. In addition, factors that may affect outcome for individual histologies are difficult to determine from available data. In this report, we have retrospectively examined survival and prognostic factors in patients undergoing resection of isolated pulmonary metastases from breast carcinoma. In this select population, long-term survival was achieved in a substantial number of patients. In addition, a prolonged disease-free interval (DFI) was found to be associated with a prolonged postthoracotomy survival after complete resection of metastases.
Material and Methods Patient Population Forty-four patients were retrospectively identified from the patient records of The University of Texas MD Anderson Cancer Center with the diagnosis of breast cancer who underwent thoracotomy for resection of pulmonary nodules between 1981 and 1990. Excluded were patients who underwent thoracotorny for clinically obvious primary lung cancer. The determination of the metastatic origin of the lung lesion was made only by histologic similarity to the primary breast cancer with light microsCOPY. All 44 patients were women. Their median age was 55 years (range, 32 to 79 years). Twenty-two patients underwent treatment of the primary breast cancer at our institution, and 22 underwent treatment elsewhere. All patients had adenocarcinoma of the breast. No patient was lost to follow-up. 0003-4975/92/$5.00
245
LANZA ET AL PULMONARY METASTASES FROM BREAST CARCINOMA
Ann Thorac Surg 1992;54:24&8
Chemotherapy regimens varied during the period of study. Forty of the 44 patients received chemotherapy, tamoxifen, or both at some time during their clinical course. Eight of the 37 patients with resectable disease received treatment before thoracotomy, 12 patients received treatment after thoracotomy, and 15 patients were treated both before and after thoracotomy; 2 patients did not receive treatment. Three patients with unresectable disease received tamoxifen only.
1.0 tb?
-ap
-b 0
a
e
n
0.9
-
0.8
-
Total Fall 37 18
0
0.7-
'
0.6-
-
0.5
.
0.4 -
0.3 -
Stat is t ical Analyses
0.2
-
Postoperative survival and disease-free survival were calculated using the product-limit method of Kaplan and Meier [13]. Statistical comparisons between groups were made using the log-rank test with survival calculated from the date of thoracotomy. All p values resulted from two-sided tests.
0.1
-
0.0 7
Results The 44 patients underwent a total of 47 thoracotomies. Forty-one patients had a single thoracotomy and 3, bilateral staged thoracotomies for resection of disease. Two patients were explored through median sternotomies, and the remaining 42 underwent exploration through lateral thoracotomies. There were no postoperative deaths and only three minor complications consisting of prolonged air leaks for the 47 thoracotomies (3/47, 6.4%). Forty-one of the 47 resections were wedge resections, four were lobectomies, one was an anatomical segmentectomy, and one was a pneumonectomy. The overall median survival for the entire cohort of 44 5 months. Three (6.8%) of the 44 patients was 82 patients were found at thoracotomy to have benign disease only (Histoplasma, 1, and Coccidioides, 2). These 3 patients were excluded from further survival analysis. All 3 were disease-free 1 month, 17 months, and 51 months after thoracotomy. Four patients underwent thoracotomy without complete resection of disease. Two of them were found to have mediastinal lymph node involvement that precluded complete resection, 1 was found to have hilar disease requiring pneumonectomy and underwent biopsy only, and 1 had clinically evident bilateral disease and underwent unilateral biopsy only. Three of the 4 patients died 2, 23, and 26 months after thoracotomy, and 1 remains alive with disease 40 months after thoracotomy. The remaining 37 patients underwent complete resection of all clinically evident disease at thoracotomy. The median survival of these patients was 47 c 5.5 months, and the actuarial 5-year survival was 49.5% (Fig 1).Eleven (30%) of the 37 patients remain clinically disease-free a mean of 49.2 months after thoracotomy (range, 1 to 88 months). Eight (22%) of the 37 patients remain alive but with disease a mean of 67 months (range, 32 to 102 months) after thoracotomy. Eighteen (49%) of the 37 patients are dead a mean of 21 months (range, 6 to 82 months) after thoracotomy. Of the 26 patients in whom disease recurred after complete resection, only 4 experienced recurrence in the lung, whereas 22 patients expe-
*
rienced either disseminated recurrences or recurrence in locations not amenable to resection. No patient underwent attempted resection of extrathoracic recurrent disease after initial thoracotomy. Clinical variables were analyzed to attempt to identify prognostic factors that may predict prolonged survival after thoracotomy. Thirty-two of the 37 patients undergoing complete resection of histologically proven malignant disease were found to have a DFI (time from treatment of primary to development of clinically evident metastases) of more than 1 year, and 5 patients were found to have a DFI of 1 year or less. The median survival of the 32 6 patients with a DFI of more than 1 year was 82 months, and their actuarial 5-year survival was 57%. In contrast, the median survival of the 5 patients with a DFI of 1 year or less was only 15 3.6 months, and all patients were dead at 26 months ( p = 0.004) (Fig 2). Similarly, the postthoracotomy disease-free survival was significantly prolonged in the patients with an initial DFI of greater than 1 year. The median disease-free survival of the 32
*
*
1.o 0.9
0.8 0.7
E
.-
0.6
bn
0.5
e
0.4
0.3 0.2 0.1 0.0
Total
1
4
0
Fall
A 32 14r12Month 0
P2
10
20
30
40
50
5
=
5rl2Month .004
60
70
80
90
100
110
Months
Fig 2. lnitial disease-free interval. Preoperative disease-free internal is associated with prolonged postthoracotomy survival (p = 0.004).
246
LANZA ET AL PULMONARY METASTASES FROM BREAST CARCINOMA
Ann Thorac Surg 1992;54:2448
O.gB\ 0.8
Total
Fail
A 32
0 Y
22 DFI> 12 Month 5 4 D F l i 12Month
0
z L
n 0
e
n
0.4-
0.3 -
I:] 0.0
, 0
10
1, 20
,
,
,
,
,
,
,
,
,
30
40
50
60
70
80
90
100
110
Months
Fig 3 . Postthoracotomy disease-free survival. Preoperative disease-free interval (DFI) predicts postthoracotomy disease-free survival (p = 0.046). (NED = no evidence of disease.)
patients with a DFI of more than 1 year was 30 k 3.5 months, and their actuarial 5-year disease-free survival was 29%. The 5 patients with a DFI of 1 year or less had a median disease-free survival of only 11 3.6 months, and all patients had recurrence by 17 months after thoracotomy ( p = 0.046) (Fig 3). Twenty-nine of the 37 patients had a known estrogen receptor (ER) status; in 8 patients, the ER status could not be determined. In 14 of the 29 patients, the ER status was positive and in 15 patients, negative. The median survival 9 of the 14 patients who were ER positive was 81 months, and their actuarial 5-year survival was 66%. The median survival of the 15 patients who were ER negative was 23 6 months, and their actuarial 5-year survival was 28%. There was a trend toward improved postthoracotomy survival in the patients with a positive ER status ( p = 0.098) (Fig 4). Twenty-seven of the 37 patients underwent resection of
*
*
*
1.0
1
1 0.891
.-a0 C
0
0.7
-
0.6
-
0.5-
L
g
0
n
0.4
~
Total
0.3 -
O 0.1
I
0.0
4 0
0 14
10
20
30
40
50
60
Fall 6 ERP(+)
70
80
90
100
Months
Fig 4 . Estrogen receptor (ERE') status. Estrogen receptor status does not predict postthoracotomy survival (p = 0.098). (DFI = diseusefree interval; NED = no evidence of diseuse.)
a single malignant pulmonary nodule, whereas 10 patients underwent resection of two or more malignant nodules (mean, 7.2; range, 2 to 49; and median, 2). The median postthoracotomy survival for the 27 patients undergoing resection of a single pulmonary nodule was 82 ? 7.5 months, and their actuarial 5-year survival was 59%. The median postthoracotomy survival for the 10 patients undergoing resection of two or more malignant nodules was 24 5 3 months, and their actuarial 5-year survival was 36%. A slight survival advantage was noted for patients undergoing resection of a single pulmonary nodule, but the survival difference was not significant ( p = 0.229). Combining these two clinical variables did not improve their predictive power. Nine patients were identified who had a single pulmonary nodule and also were ER positive. The actuarial survival after thoracotomy was not significantly different for this group compared with the remaining 28 patients ( p = 0.548). Seventeen of the 37 patients had positive axillary lymph nodes at the time of mastectomy, and 18 had negative axillary lymph nodes. The postthoracotomy survival did not differ between these groups ( p = 0.650).
Comment Although most patients with breast cancer will have disseminated disease, as many as 21% die of metastatic disease limited to the lungs [2]. Few data exist to determine whether these patients should undergo pulmonary metastasectomy, and if so, which patients can be expected to obtain the longest survival benefit from this procedure. Mountain and colleagues [14] reviewed the cases of a previous group of patients treated at MD Anderson Cancer Center from 1957 through 1975. One hundred fiftynine patients underwent resection of carcinomatous pulmonary metastases during that 19-year period. Of the 159 patients, 21 underwent resection of metastases from breast carcinoma. The cumulative 5-year survival of this group was 14%, and the median survival as calculated by the method of Gehan was 27 months. In fact, among the various carcinomatous histologies that were analyzed, breast cancer resulted in the lowest cumulative 5-year survival. Differences between that experience and our results may reflect advances in chemotherapy and hormonal therapy of systemic disease during the past two decades. Prognostic factors were analyzed for all histologies combined [14]. Patients with a DFI of 1 year or less had the highest cumulative 5-year survival (37%), whereas those with longer DFIs had lower 5-year survival. Analysis of prognostic factors for mixed carcinomatous histologies is difficult to interpret. McCormack and Martini [15] reviewed their 17-year experience with pulmonary metastasectomy at Memorial Sloan-Kettering Cancer Center. During that 17-year period, 663 resections for pulmonary metastases were performed in 448 patients, 246 of whom had carcinomatous histologies. Of these 246 patients, 34 had breast carcinoma. The median postthoracotomy survival of the 34 patients was approximately 33 months, and the cumula-
Ann Thorac Surg 1992;54:244-8
tive 5-year survival was 30%. Prognostic factors were not analyzed in that group of patients. Wright and co-workers [5] reviewed the experience with pulmonary metastasectomy at The University of Iowa over a 22-year period. One hundred forty-two patients underwent 153 thoracotomies with an operative mortality of 0.7%. One hundred patients had carcinomatous histologies, and of them, 18 had carcinoma of the breast. Wright and co-workers found no significant difference in postthoracotomy survival between patients with sarcomatous and patients with carcinomatous metastases. The actuarial 5-year survival for the 18 patients with breast carcinoma was 27%. The preoperative DFI did not predict postthoracotomy survival in their analyses. In this series, we report a 49.5% actuarial 5-year survival and a median survival of 47 k 5.5 months for this select group of patients with metastatic breast cancer confined to the lungs who underwent complete resection of the metastatic disease. There were no postoperative deaths, and only 3 patients experienced minor complications consisting of prolonged air leaks, none of which required reoperation. Although it is not possible to conclude from these or other retrospective data that pulmonary metastasectomy in patients with breast cancer prolongs survival, the data do suggest that pulmonary metastasectomy can be done safely in this group of patients and may result in long-term survival for a substantial number. A longer DFI is associated with a longer postthoracotomy survival in this select population. This observation has been made previously for other histologies [lo, 161. Whether this represents a favorable tumor biology, a favorable tumor-host relationship, or other unknown factors is impossible to determine. A trend toward prolonged survival in patients with ER-positive tumor was evident, but the survival benefit was not significant. However, it may be that ER status is less influential in predicting survival in patients with systemic disease that is completely resected. Chemotherapy regimens varied during the period of study. Thirty-five of the 37 patients with completely resected metastases received chemotherapy, tamoxifen, or both at some time during their clinical course. Two patients with resected metastases received no chemotherapy. Eight patients received systemic treatment before thoracotomy, 12 patients were treated after thoracotomy, and 15 patients were treated both before and after thoracotomy. Three of these 35 patients received tamoxifen alone. Three patients with unresectable disease received tamoxifen only. The efficacy of systemic chemotherapy or hormonal therapy in prolonging survival in this group of patients cannot be determined. No other analyzed variable demonstrated predictive power for postthoracotomy survival. For other histologies, similar variables have been found useful in predict-
LANZA ET AL PULMONARY METASTASES FROM BREAST CARCINOMA
247
ing postoperative survival 1171. Thus, it is evident that prognostic factors are best determined for homogeneous groups of patients with closely matched histologies and should not be generalized for divergent groups. In conclusion, resection of pulmonary metastases from carcinoma of the breast may result in long-term survival for select patients. Patients with a DFI of greater than 1 year have an excellent prognosis after resection of pulmonary metastases, and more than 50% may survive beyond 5 years. The procedure can be performed with no mortality and only minimal morbidity. Therefore, it represents a viable option for the treatment of patients with metastatic breast cancer confined to the lung.
References 1. Harris JR, Hellman S, Canellas GP, Fisher B. Cancer of the breast. In: DeVita VT, Hellman S, Rosenberg SA, eds. Cancer: principles and practice of oncology. 2nd ed. Philadelphia: JB Lippincott, 1985:1119-77. 2. Ramming KP. Surgery for pulmonary metastases. Surg Clin North Am 1980;60:81524. 3. Barney JD, Churchill EJ. Adenocarcinoma of the kidney with metastases to the lungs cured by nephrectomy and lobectomy. J Urol 1958;42:269-71. 4. Ehrenhaft JL, Lawrence MS, Sensenig DM. Pulmonary resection for metastatic lesions. Arch Surg 1958;77606-12. 5. Wright JO, Brandt B, Ehrenhaft JL. Results of pulmonary resection for metastatic lesions. J Thorac Cardiovasc Surg 1982;83:94-9. 6. Alexander J, Haight C. Pulmonary resection for solitary metastatic sarcomas and carcinomas. Surg Gynecol Obstet 1947;85:129. 7. Marks P, Ferrag MZ, Ashof H. Rationale for the surgical treatment of pulmonary metastases. Thorax 1981;36:679-82. 8. Morton DL, Joseph WL, Ketcham AS, et al. Surgical resection and adjunctive immunotherapy for selected patients with multiple pulmonary metastases. Ann Surg 1973;178:360-6. 9. Takita H. Management of pulmonary metastases from osteogenic sarcoma. Ann Thorac Surg 1983;36:508. 10. Creagan ET, Flemming TR, Edmondson JH, Pairolero PC. Pulmonary resection for metastatic non-osteogenic sarcoma. Cancer 1979;44:1908-12. 11. Putnam JB Jr, Roth JA, Wesley MN, et al. Analysis of prognostic factors in patients undergoing resection of pulmonary metastases from soft tissue sarcomas. J Thorac Cardiovasc Surg 1984;87:260-8. 12. Putnam JB Jr, Roth JA, Wesley MN, Johnston MR, Rosenberg SA. Survival following aggressive resection of pulmonary metastases from osteogenic sarcoma: analysis of prognostic factors. Ann Thorac Surg 1983;36:51&23. 13. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc 1958;53:457-81. 14. Mountain CF, Khalil KG, Hermes KE, Frazier OH. The contribution of surgery to the management of carcinomatous pulmonary metastases. Cancer 1978;41:83340. 15. McCormack I'M, Martini N. The changing role of surgery for pulmonary metastases. Ann Thorac Surg 1979;28:13945. 16. Huth JF, Holmes EC, Vernon SE, et al. Pulmonary resection for metastatic sarcoma. Am J Surg 1980;140:9-16. 17. Roth JA, Putnam JB Jr, Wesley MN, et al. Differing determinants of prognosis following resection of pulmonary metastases from osteogenic and soft tissue sarcoma patients. Cancer 1985;55:1361-6.
248
LANZA ET AL PULMONARY METASTASES FROM BREAST CARCINOMA
A n n Thorac Surg 1992;54:24&8
DISCUSSION DR THOMAS M. EGAN (Chapel Hill, NC): I enjoyed your provocative paper. Do you have data on 5-year survival of patients with breast cancer and a solitary metastasis other than to the lung? In other words, have you been able to demonstrate that your surgical intervention contributed to that 5-year survival? DR LANZA: No, we do not have those data, and they are difficult to obtain from patient records. DR HARVEY I. PASS (Bethesda, MD): My colleagues and I prefer a median sternotomy for resection of metastases. Are there select populations of patients with metastases for whom you do
not think a sternotomy is necessary, or should they all have median sternotomies as opposed to lateral thoracotomies? DR LANZA: Median sternotomies are most helpful when bilateral metastases are present or when bilateral metastases are susuected. as in sarcomas. Patients with metastases in one hemithorax did not frequently have new metastases occur in the other hemithorax during follow-up. Most patients did well with a single lateral thoracotomy. As to whether median sternotomies are tKe preferred route forexploration, I would leave that to the individual patient and surgeon, as number of metastases, location, familiarity with the approach, and histology all weigh in the final decision.
Notice From The Society of Thoracic Surgeons New Abstract Deadline Date The abstract due date is Monday, August 17, 1992, for all papers submitted for the Twenty-ninth Annual Meeting of The Society of Thoracic Surgeons, which will be held at the San Antonio Convention Center in San Antonio, Texas, January 2.527, 1993. No abstracts received after this date
will be considered for presentation. The deadline for receipt of films and video tapes is Monday, August 3, 1992. Details are available in the Call for Abstracts published in this issue of The Annals or from Society headquarters, 401 N Michigan Avenue, Chicago, IL 60611 (telephone: (312) 644-6610).
