The aim of this study was to evaluate the clinical treatment outcomes of recurrent breast cancer with a limited number of isolated lung metastases, and to evaluate the role of pulmonary metastasectomy.
The aim of this study was to evaluate the clinical treatment outcomes of recurrent breast cancer with a limited number of isolated lung metastases, and to evaluate the role of pulmonary metastasectomy.
The authors consecutively enrolled 140 recurrent breast cancer patients with isolated lung metastasis from 1997 to 2007 in Seoul National University Hospital and retrospectively analyzed 45 patients who had <4 metastatic lesions.
Fifteen patients had pulmonary metastasectomy followed by systemic treatment (pulmonary metastasectomy group), and 30 received systemic treatment alone (nonpulmonary metastasectomy group). The 3-year progression-free survival (PFS) and 4-year overall survival (OS) was significantly longer in the pulmonary metastasectomy group than in the nonpulmonary metastasectomy group (3-year PFS, 55.0% vs 4.5%, P < .001; 4-year OS, 82.1% vs 31.6%, P = .001). In multivariate analysis, a disease-free interval (DFI) of <24 months (hazard ratio [HR], 4.53; 95% CI, 1.72-11.90), no pulmonary metastasectomy (HR, 9.52; 95% CI, 3.34-27.18) and biologic subtypes such as human epithelial growth factor receptor-2 positive (HR, 3.00; 95% CI, 1.04-8.64) and triple negative (HR, 3.92; 95% CI, 1.32-11.59) were independent prognostic factors for shorter PFS.
The authors' results demonstrated that DFI and biologic subtypes of tumor are firm, independent, prognostic factors for survival, and pulmonary metastasectomy can be a reasonable treatment option in this population. Further prospective studies are warranted to evaluate the role of pulmonary metastasectomy. Cancer 2010. © 2010 American Cancer Society.
The mainstay of management for patients who have metastatic cancer is a systemic treatment, and the role of local treatment is limited. It is generally accepted that local treatment of primary or metastatic lesions, that is primary tumor resection or metastasectomy, provides no survival advantage once systemic metastasis has occurred. But, the benefit of local treatment in systemic metastasis is emerging for some kinds of malignancies. For colorectal cancer patients with liver metastasis, complete liver resection is the most important determinant for prognosis and remains the only chance of cure; liver resection has a 5-year survival rate of 35% to 40%.1, 2 In addition, the resection of isolated, metastatic, renal-cell cancer,3, 4 sarcoma,5, 6 and melanoma7, 8 has also been associated with longer survival or even cure. However, there have been only limited reports on the role of local treatment in patients with metastatic breast cancer. Although there is a report that primary breast-tumor resection for the purpose of tumor debulking improves survival in patients with metastatic breast cancer,9 few clinical studies have reported the efficacy of curative resection of metastasis.
The lung is the common metastatic site for recurrent breast cancer. In studies of a large autopsy series, the median incidence of lung metastases was 71%.10 And, in the retrospective analysis of 1581 patients with metastatic breast cancer, about 23% of patients had lung metastases, and 5.6% had isolated lung metastases as the first site of recurrence.11 The clinical outcomes of systemic treatment in patients who had metastatic breast cancer with isolated lung metastasis were disappointing. In a retrospective analysis at M. D. Anderson Cancer Center,11, 12 the median progression-free survival (PFS) and overall survival (OS) were 11 months and 22.5 months, respectively, for metastatic breast cancer patients with isolated lung metastasis who were treated with systemic chemotherapy. Recently, some investigators reported that surgical resection of metastases had a significant therapeutic role in metastases confined to lung.13-17 In these pulmonary metastasectomy studies, the median OS was 35-75.6 months, and the 5-year OS was 38% to 54%.13-19 These studies have consistently reported that a long disease-free interval (DFI) was an independent prognostic factor for survival.13-18, 20
However, studies of pulmonary metastasectomy for breast cancer have been criticized for their selection of patients with a low tumor burden and a relatively good prognosis compared with those receiving systemic treatment.11, 19 Patients undergoing surgical resection for metastatic lesions must be medically fit and have a small volume of metastatic disease. Although we could not clearly define small-volume metastatic disease, about 71% to 98% of patients included in the pulmonary metastasectomy studies for breast cancer had <4 metastases.13, 14, 17, 19 Also, in the studies where patients underwent complete resection of a pulmonary metastasis with intent to cure, better OS was observed for patients with <4 metastatic lesions compared with those who had >4.13 These observations could indicate that the patients with >4 lung metastases may not be suitable candidates for actual complete resection, even though they have technically resectable disease, and this may be associated with the high rate of lymphangitic and hematogenous spread around metastatic nodules. In addition, some studies were partly inadequate for determining the therapeutic role of pulmonary metastasectomy because they included the patients who were not completely resected with intent to cure but who were incompletely resected for a diagnostic aim.17, 19 In the largest pulmonary metastasectomy analysis of patients with metastatic breast cancer, complete resection showed a significant survival benefit compared with incomplete resection.14
Thus, we investigated clinical outcomes for recurrent breast-cancer patients with <4 metastases confined to the lung and treated with systemic treatment alone or pulmonary metastasectomy followed by systemic treatment. We also evaluated prognostic factors in this selected population.
At Seoul National University Hospital from January 1997 to December 2007, we consecutively enrolled recurrent breast-cancer patients who had isolated lung metastases after curative breast cancer surgery. The eligibility criteria included 1) primary breast tumors had been resected completely at the time of initial breast-cancer diagnosis, 2) recurrent metastatic disease was limited only to the lungs, and 3) the number of metastatic lesions on contrast-enhanced computed tomography (CT) scan was <4. However, the patients who received palliative systemic chemotherapy before pulmonary metastasectomy were excluded. All eligible patients were evaluated for the status of estrogen receptor (ER), progesterone receptor (PR), and human epidermal growth factor receptor 2 (HER2) from pathologic specimens. According to the status of ER, PR, and HER2 by immunohistochemical staining (IHC), the biologic subtypes were divided into hormone-receptor positive (HR+), human epidermal growth factor receptor 2 positive (HER2+), and triple negative (TN). IHC was performed by using antibodies to the ER (Dako, Carpentaria, Calif) and the PR (Dako, Carpentaria, Calif). ER and PR positivity were defined as the presence of 10% or more positively stained nuclei in 10 high-power fields. HER2 status was evaluated by IHC using an antibody (Novocastra, New Castle-Upon-Tyne, UK) and fluorescence in situ hybridization (FISH). The intensity of HER2 staining was scored as 0, 1+, 2+, or 3+. Tumors scored as 0 or 1+were considered to be negative for HER2 overexpression, whereas tumors scored as 3+ were considered to be positive. And, if HER2 was scored 2+ by IHC, FISH was used for the confirmation of the status of HER2 amplification. HER2 positivity was defined as 3+ on IHC or positive results on FISH. HR+ subtype was defined as ER and/or PR positive tumors with lack of expression of the HER2, the HER2+ subtype defined as HER2-positive tumors regardless of hormone receptor status, and the TN subtype defined as a lack of expression of ER, PR, and HER2. We retrospectively evaluated patient characteristics, treatment of patients, and clinical treatment outcomes. This study protocol was approved by the Seoul National University Hospital Institutional Review Board (IRB), Seoul, Korea (IRB protocol number H-0904-062-280). Because this study was a retrospective analysis that involved no more than minimal risk for the subjects, the IRB granted a waiver of informed consent.
The primary endpoints of this study were the PFS and OS. The PFS was defined as the first day of treatment after relapse to the date on which progressive disease was first documented or to the date of the last follow up. The tumor responses were assessed by contrast-enhanced computed tomography (CT) every 3 cycles or earlier when signs of progression were evident. Objective tumor response was evaluated by using the Response Evaluation Criteria in Solid Tumors (RECIST) criteria, but the World Health Organization (WHO) criteria was also used before RECIST was published in 2000. The OS was calculated from the first day of treatment after relapse to the date of death or the date of the last follow up. The DFI was defined as the duration from initial breast cancer surgery to the first detection of recurrence. The median duration of PFS and OS were estimated by using the Kaplan-Meier method, and 95% confidence intervals (CI) were calculated by the Greenwood formula.21 Comparisons between groups were made by using log-rank tests. Multivariate analysis was carried out by using the Cox regression model, and a significance level of .10 was used for covariate entry. The comparisons of clinical variables between the pulmonary metastasectomy group and nonpulmonary metastasectomy group were made using a chi-square test or the Fisher exact test, for categorical variables, and the Mann-Whitney test for continuous variables. A probability value of <.05 was considered significant. SPSS for Windows, version 12.0 (SPSS; Chicago, Ill), was used for all statistical analyses.
We screened 344 patients who had lung masses and had undergone curative breast-cancer surgery at Seoul National University Hospital from 1997 to 2007. Among them, 278 patients were diagnosed with recurrent breast cancer with lung metastasis and, of the 278 patients, 140 patients had isolated lung metastases. Forty-five of 140 patients had <4 metastatic lesions, and these 45 patients were included in this study (Fig. 1).
The demographic and clinical characteristics of the patients are summarized in Table 1. The median age at recurrence was 49 years (range, 32-67 years). The median DFI was 26.1 months (range, 4.2-168.0 months). Sixteen (36%) patients were ER-positive, 12 (27%) patients were PR-positive, and 14 (31%) patients were HER2-positive. According to hormone receptor and HER2 status, 16 (36%), 14 (31%), and 15 (33%) patients had biologic subtypes of HR+, HER2+, and TN, respectively. Eleven (79%) patients with HER2+ subtype were treated with trastuzumab during the course of disease. Thirty (67%) patients received systemic treatment without pulmonary metastasectomy (nonpulmonary metastasectomy group) as initial treatment after recurrence, and 15 (33%) patients had pulmonary metastasectomy (pulmonary metastasectomy group). The reason that patients did not undergo pulmonary metastasectomy in the nonpulmonary metastasectomy group were surgery was not offered to the patient (n = 26) or surgery was refused by patient (n = 4). In all patients with pulmonary metastasectomy, a complete surgical resection was performed.
|Total N=45||PM Group n=15||Non-PM Group n=30||P|
|Age at recurrence, y||.419|
|F/U duration, mo||.211|
|Stage at diagnosis||.315|
|Stage1||6 (13%)||1 (7%)||5 (17%)|
|Stage2||26 (58%)||11 (73%)||15 (50%)|
|Stage3||13 (29%)||3 (20%)||10 (33%)|
|No. of lung nodules||.036|
|1||21 (47%)||11 (73%)||10 (33%)|
|2||15 (33%)||3 (20%)||12 (40%)|
|3||9 (20%)||1 (7%)||8 (27%)|
|Maximum size of mass||.108|
|Mass <1 cm||7 (16%)||1 (7%)||6 (20%)|
|1 ≤ Mass <2 cm||18 (40%)||5 (33%)||13 (43%)|
|2 ≤ Mass <3 cm||11 (24%)||3 (20%)||8 (27%)|
|3 ≤ Mass||9 (20%)||6 (40%)||3 (10%)|
|Positive||20 (44%)||5 (33%)||15 (50%)|
|Negative||25 (56%)||10 (67%)||15 (50%)|
|Positive||14 (31%)||5 (33%)||9 (30%)|
|Negative||31 (69%)||15 (67%)||21 (70%)|
|Trastuzumab use, n=14||14 (100%)||5 (100%)||9 (100%)||.258|
|Used||11 (79%)||5 (100%)||6 (67%)|
|Not used||3 (21%)||0 (0%)||3 (33%)|
|HR+||16 (36%)||4 (27%)||12 (40%)|
|HER2+||14 (31%)||5 (33%)||9 (30%)|
|TN||15 (33%)||6 (40%)||9 (30%)|
|1997-1999||9 (20%)||2 (13%)||7 (23%)|
|2000-2002||9 (20%)||3 (20%)||6 (20%)|
|2003-2005||16 (36%)||4 (27%)||12 (40%)|
|2006-2007||11 (24%)||6 (40%)||5 (17%)|
|Chemotherapy||40 (88.9%)||12 (80.0%)||28 (93.4%)|
|T alone||14 (31.1%)||6 (40.0%)||8 (26.7%)|
|TH||7 (15.6%)||3 (20.0%)||4 (13.3%)|
|AC → T||5 (11.1%)||2 (13.3%)||3 (10.0%)|
|AT||5 (11.1%)||0 (0%)||5 (16.7%)|
|CAF||3 (6.7%)||1 (6.7%)||2 (6.7%)|
|GT||2 (4.4%)||0 (0%)||2 (6.7%)|
|GN||2 (4.4%)||0 (0%)||2 (6.7%)|
|AC||1 (2.2%)||0 (0%)||1 (3.3%)|
|CMF||1 (2.2%)||0 (0%)||1 (3.3%)|
|Hormonal therapy||5 (11.1%)||3 (20.0%)||2 (6.6%)|
|Tamoxifen||3 (6.7%)||2 (13.3%)||1 (3.3%)|
|AI||2 (4.4%)||1 (6.7%)||1 (3.3%)|
|Systemic agents used|
|Cyclophosphamide||13 (38%)||4 (27%)||13 (43%)||.680|
|Anthracyclines||28 (62%)||4 (27%)||24 (80%)||.001|
|Taxanes||40 (89%)||12 (80%)||28 (93%)||.315|
|Gemcitabine||25 (56%)||4 (27%)||21 (70%)||.010|
|Fluoropyrimidines||21 (47%)||4 (27%)||17 (57%)||.068|
|Vinorelbine||20 (45%)||3 (20%)||17 (57%)||.027|
|Platinums||12 (27%)||2 (13%)||10 (34%)||.283|
|Methotrexate||3 (7%)||0 (0%)||3 (10%)||.540|
|Irinotecan||1 (2%)||1 (7%)||0 (0%)||.333|
|Tamoxifen||7 (16%)||3 (20%)||4 (13%)||.670|
|AI||8 (18%)||2 (13%)||6 (20%)||.699|
|Death at the time of analysis||22 (49%)||2 (13%)||20 (67%)||.001|
Of 45 patients in this cohort, 40 (88.9%) patients were treated with systemic chemotherapy, and 5 (11.1%) patients were treated with hormonal therapy as first-line treatment. The most commonly used chemotherapeutic regimens were taxanes alone (n = 14, 31.1%), which were followed by taxanes/trastuzumab (n = 7), anthracyclines/cyclophosphamide (AC) → taxanes (n = 5), anthracyclines/taxanes (n = 5), cyclophosphamide/anthracyclines/5-fluorouracil(5-FU) (n = 3), gemcitabine/taxanes (n = 2), gemcitabine/vinorelbine (n = 2), AC alone (n = 1), and cyclophosphamide/methotrexate/5-FU (n = 1). Tamoxifen was used in 3 (6.7%) patients, and aromatase inhibitors were used in 2 (4.4%) patients as first-line treatments. Fewer kinds of systemic chemotherapeutic agents in the pulmonary metastasectomy group than nonpulmonary metastasectomy group were used, especially, anthracyclines (27% vs 80%; P = .001), gemcitabine (27% vs 70%; P = .010), and vinorelbine (20% vs 57%; P = .027) (Table 1).
During a median follow-up duration of 50.1 months (range, 5.0-136.2 months), the median PFS was 13.0 months (95% CI, 10.05-15.95 months), median OS was 41.7 months (95% CI, 23.82-59.58 months), and the 4-year OS rate was 45.9%. In the pulmonary metastasectomy group, a median PFS and OS was not reached, and the 3-year PFS and 4-year OS of the pulmonary metastasectomy group were 55.0% and 82.1%, respectively. In nonpulmonary metastasectomy group, median PFS and OS were 10.1 months (95% CI, 6.0-14.2 months) and 34.3 months (95% CI, 22.6-46.0 months), respectively, and the 3-year PFS and 4-year OS of nonpulmonary metastasectomy group were 4.5% and 31.6%, respectively. The median PFS and OS in the pulmonary metastasectomy group were significantly longer than those in the nonpulmonary metastasectomy group (PFS, P < .001; OS, P = .011; Fig. 2A, B)
In univariate analysis for PFS, patients with DFI of >24 months had a significantly prolonged PFS compared with those with a DFI of <24 months (median, 29.2 vs 6.3 months, respectively; P < .001). Patients with pulmonary metastasectomy as initial treatment after recurrence also had a significantly longer median PFS than those without pulmonary metastasectomy (median, not reached vs 10.1 months; 3-year PFS rate, 55.0% vs 4.5%, respectively; P < .001). The biologic subtypes, such as HR+, HER2+, and TN, showed marginally significant trends for PFS (median, 33.3 vs 14.7 vs 10.5 months, respectively; P = .084; Table 2). In multivariate analysis, a DFI of <24 months (hazard ratio, 4.53; 95% CI, 1.72-11.90), no pulmonary metastasectomy (hazard ratio, 9.52; 95% CI, 3.34-27.18), and biologic subtypes such as HER2+ (hazard ratio, 3.00; 95% CI, 1.04-8.64), and TN (hazard ratio, 3.92; 95% CI, 1.32-11.59) were independent prognostic factors for shorter PFS (Table 3).
|Variables||No. of Patients||Median PFS, mo (95% CI)||P||Median OS, mo (95% CI)||P|
|Age at recurrence, y||.332||.815|
|<50||25||13.0 (9.7-16.3)||29.6 (3.0-56.2)|
|≥50||20||11.5 (1.9-21.1)||50.0 (19.4-80.6)|
|<24 mo||20||6.3 (2.8-9.8)||23.6 (18.4-28.8)|
|≥24 mo||25||29.2 (11.1-47.3)||132.9 (24.9-240.9)|
|No. of nodules||.308||.636|
|1||21||22.5 (5.5-39.5)||50.0 (31.1-68.9)|
|2-3||24||12.3 (8.3-16.3)||34.4 (12.2-56.6)|
|Max size of mass||.102||.133|
|<2 cm||25||13.0 (9.3-16.7)||Not reached|
|≥2 cm||20||13.6 (0-36.5)||34.3 (20.7-47.9)|
|PM||15||Not reached||Not reached|
|No PM||30||10.1 (6.0-14.2)||34.3 (22.6-46.0)|
|HER2+||14||14.7 (4.6-24.8)||34.4 (17.4-51.4)|
|TN||15||10.5 (1.9-19.1)||23.6 (13.0-34.2)|
|1997-2002||18||13.1 (10.4-15.8)||45.0 (30.1-59.9)|
|2003-2007||27||13.0 (7.9-18.1)||29.6 (5.6-38.2)|
|Variables||Hazard Ratio||95% CI||P|
In univariate analysis for OS, patients with DFI of >24 months, pulmonary metastasectomy, and HR+ subtype had a superior median OS compared with those with DFI of <24 months (median, 132.9 vs 23.6 months, respectively; P < .001), no pulmonary metastasectomy (median, not reached vs 34.3 months; 4-year OS rate, 82.1% vs 31.6%, respectively; P = .011), and biologic subtypes such as HER2+ and TN (median, 134.2 vs 34.4 vs 23.6 months, respectively; P = .009; Table 2; Fig. 2B, Fig. 3A, B). In multivariate analysis, a DFI of <24 months (hazard ratio, 7.67; 95% CI, 2.00-29.43), and biologic subtypes such as HER2+ (hazard ration, 3.70; 95% CI, 1.03-13.24), and TN (hazard ratio, 3.27; 95% CI, 1.02-10.44) were independent prognostic factors for a shorter OS (Table 3).
According to the DFI and biologic subtypes of a tumor, the patients could be divided into 2 prognostic subgroups, good prognosis and poor prognosis. The good-prognosis subgroup comprised patients who were DFI ≥24 months and HR+ subtype, and the poor-prognosis subgroup comprised patients who were DFI <24months or HER2+/TN subtypes. In an analysis of the poor-prognosis subgroup, the OS was significantly longer in the pulmonary metastasectomy group than in the nonpulmonary metastasectomy group (median, not reached vs 24.5 months; P = .021), whereas in the good-prognosis subgroup, the OS was not significantly different between the pulmonary metastasectomy group and the nonpulmonary metastasectomy group (median, not reached vs 134.2 months; P = .317; Table 4).
|Good Prognosis Subgroup: DFI≥24 mo and HR+ Subtype|
|No. of Patients||Median PFS, mo (95%CI)||P||Median OS, mo (95%CI)||P|
|PM||4||Not reached||.010||Not reached||.317|
|No PM||8||33.3 (8.3-58.3)||134.2 (50.4-166.6)|
|Poor Prognosis Subgroup: DFI<24 mo or HER2+/TN Subtypes|
|PM||11||22.5 (5.4-39.6)||<.001||Not reached||.021|
|No PM||22||6.3 (3.4-9.2)||24.5 (21.2-27.8)|
Seventeen (37.8%) of 45 patients had brain metastasis in their disease course, 3 (33.3%) in the pulmonary metastasectomy group and 14 (46.7%) in the nonpulmonary metastasectomy group, respectively. There was no significant difference between the 2 groups in terms of the incidence of brain metastasis (P = .110)
As mentioned in Figure 1, 140 patients were identified as having metastatic breast cancer with isolated lung metastasis. According to the pulmonary metastasectomy and number of metastasis, we divided the cohort (n = 140) into 4 groups as follows: nonpulmonary metastasectomy and <4 metastases (n = 30), nonpulmonary metastasectomy and ≥4 metastases (n = 86), pulmonary metastasectomy and <4 metastases (n = 15), and pulmonary metastasectomy and ≥4 metastases (n = 9). The median PFS was not reached for the group of pulmonary metastasectomy and <4 of metastases, 10.1 months (95% CI, 6.0-14.2) for the group of nonpulmonary metastasectomy and <4 of metastases, 8.2 months (95% CI, 6.4-10.0) for the group of nonpulmonary metastasectomy and ≥4 of metastases, and 11.6 months (95% CI, 3.7-19.5) for the group of pulmonary metastasectomy and ≥4 of metastases; median PFS was significantly longer in the group of pulmonary metastasectomy and <4 of metastases compared with than that of the other 3 groups (P < .001), but the PFS was similar among the other 3 groups (Fig. 4A). The median OS was also not reached for the group of pulmonary metastasectomy and <4 metastases. Median OS was 34.3 months (95% CI, 22.6-46.0) for the group of nonpulmonary metastasectomy and <4 metastases, 35.8 months (95% CI, 22.6-49.0) for the group of nonpulmonary metastasectomy and ≥4 metastases, and 36.2 months (95% CI, 20.0-73.7) for the group of pulmonary metastasectomy and ≥4 metastases. Median OS was significantly longer in the group of pulmonary metastasectomy and <4 of metastases compared with than that of the other 3 groups (P = .011), but the OS was similar among the other 3 groups (Fig. 4B).
Recent advances in imaging protocols, such as multidetector-row computed tomography (MDCT) and positron emission tomography (PET) allow early detection of small lung metastasis in patients who have undergone breast-cancer surgery. Although pulmonary metastasectomy on several types of malignant tumors has been a safe and effective treatment method, lung resection in breast cancer is a controversial issue.19, 20 Lung metastasis from breast cancer is a systemic disease, and we generally choose systemic treatment. Although it is not clear which populations have prolonged survival due to a surgical approach when compared with systemic treatment alone, pulmonary metastasectomy studies in metastatic breast-cancer patients have emphasized that some patients had a great benefit from surgical resection.13-19 Therefore, if we consider >70% of inclusions in these surgical series had <4 metastatic nodules,13, 14, 17, 19 and a survival benefit also was observed for this set of patients compared with patients with >4,13 it is possible that some proportion of patients with <4 metastatic nodules may benefit from surgical resection. However, until recently, there has been a lack of research investigating the clinical treatment outcomes and the influence of surgical and systemic treatments for patients with <4 isolated lung metastases.
In the current study, 45 patients with a limited number (<4) of isolated lung metastasis who had undergone curative breast cancer surgery were primarily analyzed. In our analysis, all patients treated with pulmonary metastasectomy or systemic treatment had <4 metastases, and all patients who had pulmonary metastasectomy underwent complete resection. The median PFS and OS were 13.0 months (95% CI, 10.05-15.95 months) and 41.7 months (95% CI, 23.82-59.58 months), respectively, and the 4-year OS was 45.9%. The median PFS and OS for patients in the nonpulmonary metastasectomy group were 10.1 months and 34.3 months, respectively. It is possible that the relatively good survival of our population, compared with other studies of systemic chemotherapy alone, is a reflection of a low tumor burden and modern systemic treatment.11, 12 In our analysis, 79% of HER2-positive, recurrent, breast-cancer patients were treated with trastuzumab during the course of their disease. The pulmonary metastasectomy group, compared with the nonpulmonary metastasectomy group, had excellent survival. During a median follow-up duration of 50.1 months (range, 5.0-136.2 months), the median PFS and OS were not reached in the pulmonary metastasectomy group and showed statistically significant differences between the pulmonary metastasectomy group and the nonpulmonary metastasectomy group (PFS, P < .001; OS, P = .011). The 3-year PFS and 4-year OS in the pulmonary metastasectomy group were 55.0% and 82.1%, respectively. The 3-year PFS and 4-year OS in our population were not inferior to other pulmonary metastasectomy studies, even though the direct comparison is difficult.13, 14, 16, 17
Previous studies of pulmonary metastasectomy have shown contrary results in terms of its influence on survival. Some reports support the role of pulmonary metastasectomy, even though pulmonary metastases are a systemic disease,13-17 whereas others denied the therapeutic role of pulmonary metastasectomy.18, 19 We have thought that the contrary pulmonary metastasectomy results may be related to patients' heterogenous characteristics between the pulmonary metastasectomy studies. In an analysis of 47 recurrent breast-cancer patients, Welter et al19 reported that pulmonary metastasectomy was only needed to adjust the adequate medical treatment according to hormone receptor and HER2-expression status and to rule out primary lung cancer. However, in their analysis, 13 (27.7%) of 47 patients had ≥4 metastases, and only 28 (60%) patients had a curative resection. On the other hand, Chen F et al13 reported that OS rate after pulmonary metastasectomy was 51% at 5 years in an analysis of 41 completely resected recurrent breast-cancer patients and that pulmonary metastasectomy had a therapeutic role in some patients. This report also found that the therapeutic role was confined to patients with <4 metastases (2-year OS, 85.6% vs 33.3%; P = .0003). This idea was supported by our extended survival analysis of metastatic breast-cancer patients with isolated lung metastases. In our extended analysis according to the pulmonary metastasectomy and number of metastasis, PFS and OS were significantly prolonged in patients who had <4 metastases that were completely resected. Besides, Kaplan-Meier survival curve of patients with <4 metastases who did not have surgery was incredibly similar to those with ≥4 metastasis who did not have surgery. Thus, our results suggest that development of a treatment strategy according to the number of metastasis should be possible.
The research for prognostic factors in the surgical series has shown that DFI was the important, independent, prognostic factor.14-17, 20 In addition, the number of metastases,13, 19 achievement of complete resection,14 estrogen-receptor status,19 and the initial breast-cancer stage17 have been studied as possible prognostic factors for survival after pulmonary metastasectomy. However, their role has not been confirmed except for DFI. In our present study, DFI was a consistently independent prognostic factor for PFS and OS, as with previous studies.14-17, 19 Interestingly, biologic subtypes of the tumors were also an independent prognostic factor for PFS and OS. To our knowledge, no other pulmonary metastasectomy studies found a prognostic influence of biologic subtypes of the tumors. Only Welter's study19 reported a 5-year OS rate for estrogen receptor-positive patients that was significantly higher than estrogen receptor-negative ones (76% vs 12.1%, P = .002). Our observation that TN breast cancer has a shorter PFS and OS than HR+ or HER2+ subtypes in this population is consistent with other TN breast cancer studies.22-24 This finding suggests that a long DFI and HR positivity may represent slow-growing tumor biology. On the other hand, TN breast-cancer present with more aggressive clinical features and has worse survival.22, 24 Pulmonary metastasectomy, in our analysis, was also an independent prognostic factor for PFS, but not OS. In fact, the pulmonary metastasectomy group enrolled 6 (40%) of 15 patients between 2006 and 2007. Therefore, it is possible that this is responsible for the differences in the number of deaths at the time of analysis (pulmonary metastasectomy group vs nonpulmonary metastasectomy group; 2 [13%] vs 22 [67%]; P = .001; Table 1). Because of the difference between these 2 groups, we think further follow up on the pulmonary metastasectomy group is needed to evaluate the influence of pulmonary metastasectomy on OS.
Although patients with HER2+ had better OS than patients with the TN subtype as mentioned above, the wide gap between HR+ and HER2+ subtype is a very interesting phenomenon. In trials of first-line trastuzumab combined with chemotherapy in patients with metastatic breast cancer, the median OS was between 25.1 and 31.2 months.25, 26 Although direct comparison was inappropriate, median OS (34.4 months) of HER2+ subtype in our analysis was not inferior to that of other studies. On the other hand, median OS in HR+ subtype was 134.2 months, and this is a remarkably long survival for patients with metastatic cancers. Not surprisingly, the better survival in the HR+ subtype may be related to biologic characteristics of this cohort. As mentioned above, HR positivity represented slow growing tumor biology, and patients included in the study had a small-volume metastatic disease. Therefore, the HR+ subtype included in this study may represent the utmost in quiescent tumor biology, and this could explain why the gap between HER2+ and HR+ was wide in the present study.
In the analysis of subgroups according to DFI and biologic subtypes of tumor, the poor prognosis subgroup, comprising patients who were DFI<24months or HER2+/TN subtypes, showed statistically significant OS benefit in the pulmonary metastasectomy group. However, the good prognosis subgroup, comprising patients who were DFI≥24 months and HR+ subtype, did not show a significant OS difference between the pulmonary metastasectomy group and the nonpulmonary metastasectomy group. This observation may be quite different from previous surgical series that have reported that pulmonary metastasectomy was only beneficial in patients with a long DFI. However, caution should be taken in interpreting the results because of the extremely small number of patients included in this good-prognosis subgroup analysis (pulmonary metastasectomy group vs nonpulmonary metastasectomy group; 4 vs 8 patients). The small number of patients may affect this result in the good prognosis subgroup analysis. Hence, considering the consistently favorable pulmonary metastasectomy reports in patients with long DFI and the current low morbidity and mortality rates with lung resection, the patients in the good-prognosis subgroup might have a benefit from the surgical approach. On the other hand, survival benefit of pulmonary metastasectomy for patients with poor-prognosis characteristics, which is DFI<24 months or HER2+/TN subtypes, is an interesting finding. According to our results, if the metastasis is a small-volume disease, the patients with poor-prognosis characteristics also have a survival gain from pulmonary metastasectomy followed by systemic treatment compared with systemic treatment alone. We think this observation may suggest that early surgical resection is a reasonable option for overcoming the drug resistance or biologic aggressiveness of the tumors. This concept is supported by a study that reported remnant tumor cells after cytotoxic chemotherapy may differ from the bulk of the cells in the initial untreated tumor.27 Remnant-tumor cells can undergo epigenetic alteration, resulting in resistant tumor cells. Because drug resistance is proportional to cancer size, tumor with a lower burden may have less clonal heterogeneity. However, tumors may achieve more clonal heterogeneity and drug resistance to cytotoxic chemotherapy when tumors grow larger.28 Thus, early surgical resection in patients who have poor-prognosis characteristics will allow the removal of tumor cells that have the potential to become larger and, consequently, more clonally heterogenous and drug resistant. As a result, our study supports the finding that pulmonary metastasectomy can have a therapeutic role in patients with small-volume metastases, even though the patients have clinically poor prognosis characteristics such as short DFI or HER2+/TN subtypes.
The current study has limitations. First, the small number of patients, heterogeneity of the population, and short follow-up duration are important limitations of this study. Second, the tissues in the metastatic site were not available for 14 (47%) patients who were included in the nonpulmonary metastasectomy group. The biologic subtypes had been changed in 3 (10%) of 31 patients, for whom tissue was available in primary breast tumor and lung metastasis sites. The biologic subtypes at recurrence were changed in 2 patients from TN to HR+ subtype and 1 patient from TN to HER2+ subtype. We could not account for the possibility of changing the biologic subtype of the tumors in patients whose metastatic lung tissue was not available, and that is the major limitation of this study. Nevertheless, our study is the first to provide evidence for the intrinsic biologic characteristics of tumors as a relevant prognostic factor in patients with a limited number of isolated lung metastases from breast cancer.
In conclusion, in patients with a limited number of isolated lung metastases from breast cancer, DFI and biologic subtypes of the tumors were firm, independent, prognostic factors for PFS and OS. Pulmonary metastasectomy can be a reasonable treatment option in patients with a small volume of metastasis despite of poor-prognosis characteristics. Further studies that use a prospective design are warranted to evaluate the role of pulmonary metastasectomy.
This study was partly supported by a grant (A080316) of the Korea Healthcare Technology R&D Project, Ministry for Health, Welfare & Family Affairs, Republic of Korea.