Pulmonary metastases of breast cancer. When is resection indicated?

Transcription

1 European Journal of Cardio-thoracic Surgery 34 (2008) Pulmonary metastases of breast cancer. When is resection indicated? Stefan Welter a, *, Jan Jacobs a, Thomas Krbek a, Martin Tötsch b, Georgios Stamatis a Abstract a Department of Thoracic Surgery, Ruhrlandklinik, Tüschener Weg 40, Essen, Germany b Department of Pathology, University Hospital Duisburg-Essen, Hufelandstr. 55, Essen, Germany Received 1 May 2008; received in revised form 28 July 2008; accepted 30 July 2008; Available online 27 September 2008 Objective: While resection of pulmonary metastases is a common treatment in other primaries, the role of breast cancer metastasectomy is still unclear. The objective of the present study was to investigate the clinical outcome of our operated patients with pulmonary breast cancer metastases and discuss the different indications for metastasectomy. Methods: From January 1998 to December 2007 we retrospectively analysed 47 patients with histologically proven pulmonary metastases from breast cancer. The mean age of the 47 female patients was 56.2 years, the median disease-free interval (DFI) was 3.66 (0 25.8) years and the median follow-up was 20.6 months ( ). Results: The grading of the metastases was higher than the primary tumour in 12 of 45 (26.7%) and lower in 6 of 45 (13.3%) patients. R0, R1 and R2 resections were achieved in 27, 6 and 14 cases. The oestrogen receptor status of the metastases differed from the primary tumour in 11 out of 39 (28.2%) tested cases. Her2- neu receptor status differed in 4 out of 16 tested patients. The histological reports described a tumour spread around the metastasis in lymph or blood vessels in at least one resection specimen in 25 out of 47 (53.2%) patients. The rate of major complications was 5.8%. The overall survival from the first pulmonary metastasectomy was 32 months with a 5-year survival of 36%. The main prognostic factor was the oestrogen receptor status with a 5-year survival for receptor positive patients of 76% and 12.1% for receptor negative ones ( p = 0.002). A similar survival difference was found for the status of Her2-neu receptor ( p = 0.037). No prognostic influence could be demonstrated for age, number of metastases, initial tumour stage, complete versus incomplete resection, lymphatic spread, lymph node or parietal pleural involvement. Conclusion: The gain in life expectancy in breast cancer patients with pulmonary metastases is based on chemotherapy and antihormone treatment. Tissue of the lung metastasis is needed to adjust medical therapy to oestrogen and Her2-neu expression and to reliably rule out primary lung cancer. In case of proved pulmonary metastases, the level of evidence for a curative approach is low but some patients might benefit. # 2008 European Association for Cardio-Thoracic Surgery. Published by Elsevier B.V. All rights reserved. Keywords: Breast cancer; Metastasectomy; Hormones; Chemotherapy 1. Introduction Breast cancer is well known to be a systemic disease even in early stages. While resection of pulmonary metastases is a common treatment in other primaries, the role of breast cancer metastasectomy is still unclear, despite many other publications. There is no doubt about the necessity to rule out lung cancer in breast cancer patients presenting with pulmonary nodules, because lung cancer prognosis is worse [1,2] and tumour specific aggressive and potentially curative treatment is indicated. Furthermore, histology might be essential in case of disseminated nodules to distinguish Abbreviations: CT, computed tomography; DFI, disease-free interval; MBC, metastatic breast cancer; MST, median survival time; PET, positron emission tomography; S1, survival from initial diagnosis; S2, survival from first pulmonary metastasectomy. * Corresponding author. Tel.: ; fax: addresses: stefan.we@t-online.de, stefan.welter@ruhrlandklinik.de (S. Welter). interstitial fibrosis or rare infectious diseases. In case of histology proven metastasis from breast cancer, the level of evidence for curative resection is low [3]. The median survival for patients with metastatic disease treated with conventional chemotherapy doses and regimens is months [4,5], so that metastatic breast cancer (MBC) is considered an incurable disease. However, a small number of patients remain disease-free for a long time raising the hope that MBC is not always fatal [6]. Recently new treatment possibilities came up, concerning chemotherapy, targeted therapy and antihormone treatment with good response rates [7]. Under these circumstances it becomes more important to make clear what still are the indications for pulmonary resections. The objective of the present study was to investigate the clinical outcome of our operated patients with pulmonary breast cancer metastases taking into account prognostic factors and discuss the different indications for metastasectomy. Incomplete resections were included in the analysis to line out a possible benefit of R0 metastasectomy /$ see front matter # 2008 European Association for Cardio-Thoracic Surgery. Published by Elsevier B.V. All rights reserved. doi: /j.ejcts

2 S. Welter et al. / European Journal of Cardio-thoracic Surgery 34 (2008) Materials and methods We retrospectively analysed all patients that were resected for histologically proven pulmonary metastases from breast cancer and excluded those who had pleural biopsy and talc poudrage for known pleural dissemination. Our clinical review board waived the need for patient consent taking into account the anonymous data interpretation. From January 1998 to December 2007 altogether 47 patients were included. All medical records were reviewed for biographic data, time, stage and hormone receptor status of the primary breast cancer and the metastases, indication for thoracic surgical intervention, mode of intervention, number and diameter of metastases, resection margins, lymph node involvement, and postoperative complications. Follow-up data about further treatment, including the treatment with trastuzumab, tumour progression and survival information were collected from patients, relatives, the oncologists and in two cases from the registry offices. All pathological reports from the primary and secondary tumour were reviewed for subtype and the reports of the metastases were additionally reviewed for invasion of blood and lymph vessels and a continuous spread into the parietal pleura. Table 1 presents the 47 patients pretreatment characteristics, including prognostic factors. Among them, seven had metastases in other organs in complete remission. The admission of patients for lung surgery was dependent on the opinion of the pretreating oncologists. When a complete Table 1 Patient characteristics (n = 47) Age (years) Median 55.0 Range DFI (years) Median 3.7 Range Initial tumour stage (UICC-classification 1997) Stage I n = 12 (26%) Stage II n = 25 (53%) Stage III n = 6 (13%) Stage IV n = 4 (9%) Resected local breast cancer recurrence 12/47 (26%) Indication for metastasectomy Curative 28 (60%) Suspected lung cancer 8 (17%) Proof of histology 11 (23%) Metastases in other organs in remission Before pulmonary resection 7 (15%) Resection at the first metastasectomy (n = 47) Wedge 31 (66%) Segment 11 (23%) Lobe 4 (9%) Sternum + wedge 1 (2%) Resection at the second metastasectomy (n = 9) Wedge 4 (44%) Segment 2 (22%) Lobe 1 (11%) Sternum 2 (22%) Number of metastases (first/second metastasectomy) 1 metastasis 29/6 2 metastases 2 3 metastases 3 >3 metastases 13/3 removal of metastases was strived for, other organ deposits had been ruled out by magnetic resonance imaging or computed tomography scan (CT) of the brain, CT of the abdomen and bone scan or PET-CT scan in two cases. The decision for resection depending on the functional status was made by the thoracic surgeon and a pneumologist. Preoperative investigations consisted of lung function tests, blood gas analysis, fiber optic bronchoscopy, CT scan of the chest and an ultrasound scan of the abdomen. Preoperative selection criteria were as follows: completely resectable lesions in the lung, not more than five seen on CT scan, no evidence of metastases other than in the lung, no suspicion of mediastinal lymph node metastases, good performance score (WHO 0 2), sufficient lung function and no contraindication due to concomitant disease. If only biopsy was planed, no restrictions were made despite the need to tolerate anaesthesia. The indication for surgery was complete removal of a single tumour site, removal of remaining tumour after chemotherapy, the extraction of tumour tissue for immunohistochemical work-up or diagnostic only to rule out lung cancer Statistical analysis Survival was defined as the time elapsed between the date of primary tumour resection (S1) or the first pulmonary metastasectomy (S2) and the date of death or last contact. The data were analysed using the program SPSS 15 for Windows (SPSS Inc., Chicago, IL). Probability of survival was calculated according to the method of Kaplan Meier [8]. The prognostic influence of variables on survival was analysed using the log-rank test [9]. Differences between proportions were calculated using the chi-square test. A multivariate analysis using the Cox proportional hazards model [10] was used to perform a multivariate analysis of prognostic factors for both start points. A probability value of 0.05 or less was considered significant. Stepwise regressions of parameters due to highest p value were achieved with the likelihood quotient method (LR-method). 3. Results The patients characteristics are listed in detail in Tables 1 3. The median follow-up was 20.6 months with a range of months after the first pulmonary metastasectomy. Twenty-nine patients had a disease-free interval (DFI) beginning from the resection of the primary to the first metastasectomy of >3 years and only nine of them died until the end of the study. The type of largest resection was wedge resection in 31, segmental resection in 11, lobectomy in 4 and one lung wedge with major sternal resection. A second lung resection was done in nine patients, including wedge resection in four, segmental resection in two, lobectomy in one and extended resection with sternum in two patients. Concerning the first pulmonary metastasectomy a microscopic complete resection was achieved in all resected specimen in 27 patients, while 6 were described as R1. The tumour grading differed between the primary tumour and the metastases in 18 of 45 patients (40%). The oestrogen receptor status of the metastases as well as of the primary was tested

3 1230 S. Welter et al. / European Journal of Cardio-thoracic Surgery 34 (2008) Table 2 Pathologic macroscopic characteristics Largest diameter of R0 or R1 resected specimen (first metastasectomy) (n = 33) Largest diameter of resected specimen (second metastasectomy) (n = 9) Median 1.7 cm (range ) Median 2.5 cm (range ) Lymph node involvement in R0 or R1 resected patients (n = 33) No involvement 17 (52%) Not investigated 7 (21%) Hilar/mediastinal 9 (27%) Parietal pleural spread around the metastasis n = 7 (47) Pulmonary resection margins (first metastasectomy) Free of tumour (R0) 27 Microscopic tumour (R1) 6 Macroscopic tumour left (R2) 14 Pulmonary resection margins (second metastasectomy) Free of tumour (R0) 7 Microscopic tumour spread 2 (sternum resections) in 39 cases among which 16 were positive and 23 were negative with a significant ( p = 0.006) accordance to the status of the primary tumour. Four out of 20 metastases were oestrogen receptor positive while the primary was negative and 7 out of 19 were negative while the primary tumour was positive for oestrogen receptors. Her2-neu receptor status was tested rarely in the primary. One out of 6 metastasis was positive while the primary was negative and 3 metastases out of 10 were negative while the primary was positive. The statistical accordance (chi-square) was not significant ( p = 0.059). The histological reports described a tumour spread around the metastasis in lymph or blood vessels in at least one resection specimen in 25 out of 47 patients equivalent to 53.2%. This frequent occurrence of lymphangitic spread might be a breast cancer specific manner of growth in lung tissue. Table 3 Pathologic microscopic characteristics Histology of primary breast cancer Ductal invasive 34 (72%) Lobular 2 (4%) Others 8 (17%) Unknown 3 (7%) Hormone receptor status (oestrogen) of the primary/metastasis Negative 21/16 (4 positive, 1 unknown) Positive 21/12 (7 negative, 2 unknown) Unknown 5/0 (3 positive, 2 negative) DAKO-Hercep test (0/1 negative, 2/3 positive) of the primary/metastasis Negative 9/5 (1 positive, 3 unknown) Positive 11/7 (3 negative, 1 unknown) Unknown 27/12 (8 negative, 7 positive) Histologic grading of the primary/metastasis G1 4/0 (2 G2, 2 G3) G2 18/10 (8 G3) G3 23/17 (6 G2) Unknown 2/(2 G3) Local spread in the resected specimen Lymph vessels 15 (32%) Blood vessels 3 (6%) Both 7 (15%) None 22 (47%) Loco regional lymph nodes had been removed if suspicious; nodal involvement was found in 9 and none in 17 patients. In the remaining 21 cases no lymph node dissection was performed at all. The intention of treatment was curative in 33 and diagnostic only in 14. Major complications were found in 3 of 52 (5.8%) operations, including contra lateral resections and redo surgery, consisting of one case each of pneumonia, of acute renal failure and of bronchial stump fistula after a lower bilobectomy, which healed after a thoracostomy and later stump coverage with latissimus dorsi muscle. The 30-day mortality was 0%. Thirty-two out of 47 (68.1%) patients faced tumour recurrence or tumour progression, including other organs. All 15 patients without recurrence had a positive receptor profile and have ongoing antihormone medication or adjuvant trastuzumab treatment. The overall median survival from initial diagnosis (S1) to the end of the study was 118 months with a 5- and 10-year survival of 76% and 48%. The overall median survival from the first pulmonary metastasectomy (S2) was 32 months with a 5-year survival of 36% (Fig. 1). The main survival difference was found concerning the oestrogen receptor status (Table 4). From the time of first pulmonary metastasectomy the 5-year survival with a positive hormone receptor status was 76% (median not reached), whereas the corresponding survival with a negative receptor status was 12.1% with a median survival of 25.6 months. This difference was significant ( p = 0.002) (Fig. 2). A similar pattern could be demonstrated beginning from the primary diagnosis (S1) for the status of Her2-neu receptor: 5-year survival with a positive 92.9%, and 62.5% with a negative receptor status. The difference was significant ( p = 0.038). The corresponding receptor associated S2 difference was significant as well ( p = 0.037) (Fig. 3). Analysing the R0 and R1 resected patients only for the influence of the receptor status, oestrogen reached Fig. 1. Overall (n = 47) probability of survival after the first pulmonary metastasectomy in pulmonary MBC patients. Months Patients at risk Events

4 S. Welter et al. / European Journal of Cardio-thoracic Surgery 34 (2008) Table 4 Kaplan Meier survival analysis (first metastasectomy) and log-rank test Prognostic group Cases, n MST (months)/95% confidence interval 5-year survival Sig., p Overall ( ) 36% Age (years) ( ) 16% > ( ) 56% 0.46 Initial tumour stage I ( ) 44% II ( ) 29% 0.75 III ( ) 33% 0.86 IV % 0.98 Disease-free interval (DFI) 3 years ( ) 22% >3 years ( ) 50% Oestrogen receptor (metastasis) Positive 19 77% Negative ( ) 12% Oestrogen receptor (n = 33, R0/R1 only) Positive ( ) 15% Negative ( ) 71% Her2-neu receptor (metastasis) Negative ( ) (3-y.) 22% Positive 15 74% Number of metastases ( ) 41% > ( ) 22% Resection margins R0/R ( ) 33% R ( ) 39% 0.74 Local spread (lymphatic, vascular) L0/V ( ) 38% L1/V ( ) 31% 0.96 Parietal pleural involvement No ( ) 38% Yes ( ) n.r Lymph node involvement (curative approach) No ( ) 26% Yes 9 Not reached 66% 0.33 Adjuvant therapy after pulmonary metastasectomy Hormone % Chemo ( ) 19% Bold values indicate significance. significance ( p = 0.040), whereas Her2-neu did not reach ( p = 0.52) a significant impact on survival. No prognostic influence could be demonstrated for age, DFI, number of metastases, diameter of metastases, initial tumour stage, complete versus incomplete resection, lymphangitic spread, other organ metastases in remission, lymph node or parietal pleural involvement. Concerning adjuvant therapy after pulmonary metastasectomy, the survival of patients with antihormone treatment was significantly longer than for those with chemotherapy alone ( p = 0.003), reflecting as well the prognostic importance of oestrogen receptor status. The multivariate analysis identified the oestrogen receptor status of the metastasis as the single factor of prognostic relevance ( p = 0.002). This analysis was restricted because Her2-neu could not be included into stepwise regression analysis due to missing values. 4. Discussion In our study we could show an overall median survival of 32 months and a 5-year survival of 36% for complete and incomplete resected patients. That is quite similar to most of the other published surgical series [1 3,11,12]. Some found better results with a 5-year survival of up to 54% [13,14]. The extraordinary good results of Yoshimoto et al. [13] may reflect the highly selected collective with 63% having a DFI >3 years, 70.2% having a single metastasis only and all having a complete resection. The limitations of this study are the short median followup period, the restriction to a single centre and, most importantly, the retrospective study design. However, in our opinion it is almost impossible to generate more reliable data: the inclusion of patients treated before the mid 1990s suggests a less effective standard of care, the number of

5 1232 S. Welter et al. / European Journal of Cardio-thoracic Surgery 34 (2008) Fig. 2. Probability of survival (n = 41) depending on the oestrogen receptor status of the metastases. The MST with a negative receptor status is 25.6 months, the 5-year survival 12% and 77% when hormone receptor is positive. Months Patients at risk Negative Positive resected patients is small everywhere so that prospective trials seem to be impossible or too time consuming and the indication for resection underlies a variety of oncologist opinions. Furthermore this study suffers from an important selection bias because the thoracic surgeon was only included into therapeutic decisions when asked by the oncologist. In Fig. 3. Probability of survival (n = 31) depending on the Her2-neu status of the metastases. The MST with a negative receptor status was 29.2 months, the 4- year survival 22% and the 5-year survival 74% when the Her2-neu receptor is positive. Months Patients at risk Negative Positive this circumstance it is important to underline that the investigation of hormone receptors and Her2-neu status of the metastases, when it was different from the primary tumour, was followed by a change in the systemic antitumour therapy. Because almost every patient with MBC received chemotherapy in all published series, it is difficult to prove that resection of metastases is beneficial. A case series of patients with breast cancer metastases confined to the lungs and treated with chemotherapy only was published from the M.D. Anderson Cancer Center in 1998 [15]. The overall median survival was 22 months with a 5-year survival around 16%. Important to recognise is the fact that the hormone receptor profile was unknown in 77% of the collective. One possible explanation for our better results (median survival 32 months) could be the fact that 39 (83%) patients had an investigation of hormone receptors of the metastases and the adjuvant treatment was adjusted to these results. Patients with minimum disease (one or two lesions up to 2 cm diameter) achieved a median overall survival of 57 months [15], which is almost identical to the surgical data of Friedel et al. [16]. He described women in the best prognostic group with a long DFI of over 3 years and a single completely resected metastasis having a median survival of 59 months. On the other hand, only 33% of the patients with an isolated lung metastasis achieved a complete response under chemotherapy, which would be comparable to a complete surgical resection. Complete responders achieved a median overall survival of 37 months. In a word, minimal tumour burden and complete response to chemo stays for better survival [15]. A further long-term follow-up study of 1581 patients with metastatic breast cancer treated with chemotherapy [17] could demonstrate a response rate of 65%, including 16.6% complete responders. The overall survival was 21.3 months and 41.8 months for complete responders. That indicates that the median survival for resected patients [1,2,3,11] was no better than that documented for patients receiving only modern chemotherapy, especially with minimal tumour burden and with a complete response. Further studies could address the question if resectable patients not reaching a complete remission under chemotherapy would benefit most from surgical intervention. Several prognostic factors were described for MBC patients: complete resection, long DFI, single metastasis and positive hormone receptor status. All factors have to be discussed separately Resection In the presented study we could not show a survival difference concerning complete and incomplete metastasectomy ( p = 0.744). The slightly better survival of R2- versus R0 and R1 resected patients in our series might be due to the distribution of oestrogen receptors: 50% in the R2 and only 36% of the R0 and R1 resected patients were positive. No significant survival difference was found by McDonald et al. [3] and Planchard et al. [12], whereas Friedel et al. [16] found a significant survival benefit ( p = ) for patients with complete resection. This might be due to the study s inclusion criteria; as patients from 1960 and later were

6 S. Welter et al. / European Journal of Cardio-thoracic Surgery 34 (2008) included, long before modern chemotherapy was established, only 25.1% had adjuvant chemotherapy and incomplete resection was an intraoperative not expected finding and mainly due to locally advanced irresectable tumours. McDonald et al. [3] even found a better 5-year survival for patients having incomplete resection (42.1%) versus 35.6% after complete resection, which would underline the superior impact of medical treatment for MBC patients DFI A long disease-free interval seems to be of prognostic relevance concerning resected patients [2,12 14,16,19] and medically treated patients as well [7]. The interval reaching significance is 1 year and longer. A long disease-free interval seems to represent slowly growing tumours, long effective antihormone treatment and minor tumour burden Number of metastases The number of metastases is well known to be an important prognostic factor in other tumours. The results of 5206 cases of the international registry of lung metastases [18], including various histological entities showed a survival of 43% at 5 years for single metastasis and 27% for more than four metastases, but this difference did not reach significance in most publications concerning MBC [1,2,12 14,16,19]. An approach of the significance level ( p = 0.055) was described only by McDonald et al. [3]. Our study showed a significant (0.050) survival difference between one and three metastases and more than three consisting of 1 case with 8 and 12 cases with multiple, incompletely resected lesions. Therefore the interpretation of this significance is problematic. The high rate of lymphangitic and vascular spread around the metastases, 53.2% in our study, underlines the necessity to keep wide safety margins to assure long lasting local control Receptor status Oestrogen receptor status was examined only in a limited number of studies for its prognostic relevance. Our presented study demonstrates a survival benefit (S2) for patients with oestrogen receptor positive metastases on a significant level ( p = 0.002). To our knowledge, no other surgical series [3,14,19,20] described a significant prognostic influence of oestrogen receptor status to date. While citing several studies where the addition of tamoxifen (oestrogen receptor positive patients) significantly improved response rate and survival, Bergh et al. [4] concluded in his overview that the addition of hormonal therapy to chemo adds no survival benefit to chemotherapy alone. This data interpretation is inapprehensible. Rahman et al. [17] found a median survival difference of 28.6 months for positive versus 18.1 months for receptor negative patients ( p < ) in a series with 1581 MBC cases. He could also show the same response rate to doxorubicin based chemotherapy, but a much longer progression free survival and overall survival for those expressing oestrogen receptors. This effect is closely connected with the addition of tamoxifen to chemotherapy. The better survival of patients with a positive status of Her2-neu in this study can be explained by the fact that most of these patients (9/15) had trastuzumab in the following period and only 3 of them had a tumour progression whereas 8 out of 16 patients with a negative status of Her2-neu faced tumour progression and no one had trastuzumab in the course of their disease Chemotherapy In patients with MBC, doxorubicin-containing, standarddose chemotherapy achieved a response rate of 65% with a median overall survival of 21.3 months with a clear difference concerning survival between complete responders (MST 41.8 months), partial responders (MST 24.6 months) and patients with progressive disease (MST 3.8 months) [17] Tumour recurrence The rate of tumour recurrence after pulmonary metastasectomy is well known to be high. It was 56% in the international registry of lung metastases series [16] and was 82.1%, described by McDonald et al. [3], despite complete resection of metastases, and 68.1% in the presented study Indication The most important question for the thoracic surgeon going to operate on a patient with pulmonary nodules and a history of breast cancer will be: is there any benefit for the patient? To answer this question, Tanaka et al. [1] and Rena et al. [2] investigated their results concerning the histology found in those nodules. They found % pulmonary metastases, % primary lung cancers and % benign lesions. That is most important to plan further therapy. Even multiple nodules were not necessarily metastases. 5. Conclusion Every pulmonary nodule in breast cancer patients that are otherwise free of metastases should be aggressively diagnosed because lung cancer has to be ruled out and benign lesions could alter the clinical tumour stage and ongoing therapy. Because patients with MBC confined to the lung have almost the same survival patterns after complete resection and after complete response to modern chemotherapy, a benefit for complete metastasectomy is not proven. Furthermore, there is no significant survival difference between single and multiple metastases in surgical series nor is there a clear difference between complete and incomplete resection in all series. This indicates that the survival depends more on chemo- or antihormone treatment than on R0-resection. As we could show, a difference between primary and secondary tumour concerning the grading, and the receptor profile is not rare and important to know for further adjusted treatment. Unfortunately this series is too small to demonstrate a benefit concerning local control by R0-resection. Further studies might address this point. Altogether we can support two indications for pulmonary resection of nodules in patients with a history of breast cancer:

7 1234 S. Welter et al. / European Journal of Cardio-thoracic Surgery 34 (2008) To find and resect primary lung cancer and rule out benign lesions. 2 To prove the histology and investigate grading, hormone receptor status and Her2-neu status to adapt medical treatment. References [1] Tanaka F, Li M, Hanaoka N, Bando T, Fukuse T, Hasegawa S, Wada H. Surgery for pulmonary nodules in breast cancer patients. Ann Thorac Surg 2005;79(5): [2] Rena O, Papalia E, Ruffini E, Filosso PL, Oliaro A, Maggi G, Casadio C. The role of surgery in the management of solitary pulmonary nodule in breast cancer patients. Eur J Surg Oncol 2007;33(5): [3] McDonald ML, Deschamps C, Ilstrup DM, Allen MS, Trastek VF, Pairolero PC. Pulmonary resection for metastatic breast cancer. Ann Thorac Surg 1994;58(6): [4] Bergh J, Jönsson PE, Glimelius B, Nygren P, SBU-group. Swedish Council of Technology Assessment in Health Care. A systematic overview of chemotherapy effects in breast cancer. Acta Oncol 2001;40: [5] Stockler M, Wilcken NR, Ghersi D, Simes RJ. Systematic reviews of chemotherapy and endocrine therapy in metastatic breast cancer. Cancer Treat Rev 2000;26: [6] Greenberg PA, Hortobagyi GN, Smith TL, Ziegler LD, Frye DK, Buzdar AU. Long-term follow-up of patients with complete remission following combination chemotherapy for metastatic breast cancer. J Clin Oncol 1996;14: [7] Atalya G, Biganzoli L, Renard F, Paridaens R, Cufer T, Coleman R, Calvert AH, Gamucci T, Minisini A, Therasse P, Piccart MJ, on behalf of the EORTC Breast Cancer and Early Clinical Studies Groups. Clinical outcome of breast cancer patients with liver metastases alone in the anthracyclinetaxane era: a retrospective analysis of two prospective, randomised metastatic breast cancer trials. Eur J Cancer 2003;39: [8] Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Soc 1958;53: [9] Peto R, Peto J. Asymptotically efficient rank and invariant procedures. J R Stat Soc 1972;135: [10] Cox DR. Regression models and life-tables. J R Stat Soc 1972;34: [11] Friedel G, Linder A, Toomes H. The significance of prognostic factors for the resection of metastases of breast cancer. Thorac Cardiovasc Surg 1994;42(2):71 5. [12] Planchard D, Soria JC, Michiels S, Grunenwald D, Validire P, Caliandro R, Girard P, Le Chevalier T. Uncertain benefit from surgery in patients with lung metastases from breast carcinoma. Cancer 2004;100: [13] Yoshimoto M, Tada K, Nishimura S, Makita M, Iwase T, Kasumi F, Okumura S, Sato Y, Nakagawa K. Favourable long-term results after surgical removal of lung metastases of breast cancer. Breast Cancer Res Treat 2008;110: [14] Lanza LA, Natarajan G, Roth JA, Putnam Jr JB. Long term survival after resection of pulmonary metastases from carcinoma of the breast. Ann Thorac Surg 1992;54(2): [15] Diaz-Canton EA, Valero V, Rahman Z, Rodriguez-Monge E, Frye D, Smith T, Buzdar AU, Hortobagyi GN. Clinical course of breast cancer patients with metastases confined to the lungs treated with chemotherapy. Ann Oncol 1998;9: [16] Friedel G, Pastorino U, Ginsberg RJ, Goldstraw P, Johnston M, Pass H, Putnam JB, Toomes H. Results of lung metastasectomy from breast cancer: prognostic criteria on the basis of 467 cases of the International Registry of Lung Metastases. Eur J Cardiothorac Surg 2002;22: [17] Rahman ZU, Frye DK, Smith TL, Asmar L, Theriault RL, Buzdar AU, Hortobagyi GN. Results and long term follow-up for 1581 patients with metastatic carcinoma treated with standard dose doxorubicin-containing chemotherapy. Cancer 1999;85: [18] Pastorino U, Buyse M, Friedel G, Ginsberg RJ, Girare P, Goldstraw P, Johnston M, McCormack P, Pass H, Putnam JB. Long-term results of lung metastasectomy: prognostic analysis based on 5206 cases. J Thorac Cardiovasc Surg 1997;113: [19] Ludwig C, Stoelben E, Hasse J. Disease-free survival after resection of lung metastases in patients with breast cancer. Eur J Surg Oncol 2003; 29(6): [20] Kayser K, Biechele U, Kayser G, Dienemann H, Andre S, Bovin NV, Gabius HJ. Pulmonary metastases of breast carcinomas: ligandohistochemical, nuclear, and structural analysis of primary and metastatic tumors with emphasis on period of occurrence of metastases and survival. J Surg Oncol 1998;69: