Jpn J Clin Oncol 2009;39(2)97 104 doi:10.

1093/jjco/hyn133

Risk Factors for Initial Lung Metastasis from Breast Invasive Ductal Carcinoma in Stages I III of Operable Patients
Dongwei Gao1,, Jiajun Du2,, Lei Cong1 and Qi Liu3
Department of Oncology, Shandong Provincial Hospital Afliated to Shandong University, Medical School of Shandong University, 2Department of Thoracic Surgery, Shandong Provincial Hospital Afliated to Shandong University and 3Department of Oncology, Shandong Provincial Hospital Afliated to Shandong University, Jinan, Peoples Republic of China
Received August 17, 2008; accepted October 22, 2008; published online December 3, 2008
1

Objective: The aim of this study was to evaluate high-risk factors for initial lung metastases from breast invasive ductal carcinomas in operable patients with Stages I III invasive ductal carcinoma. Methods: Data of all patients who underwent radical mastectomy were reviewed retrospectively, and they were conrmed with invasive ductal breast cancer between January 2003 and December 2007. Routine clinical examination data of patients included in the study at primary diagnosis, adjuvant modes and rst metastasis sites were recorded. Possible risk factors were easily identied from patients. Twenty-eight potential risk factors were investigated. Finally, 78 patients with complete data in the potential factors were found eligible, and univariate and multivariate analyses were conducted. Results: Univariate analyses showed that the status of estrogen receptor (ER) and progesterone receptor (PR) and the status of the epidermal growth factor receptor-2 (Her2) were high-risk factors for invasive ductal breast cancer metastasis to the lung as the rst organ. P values were, respectively, 0.045, 0.049 and 0.026. Multivariate analyses showed that the pN3 stage needs to be combined with vascular invasion to predict initial lung metastasis. The status of ER and PR was also viewed in combination with p53 negative to predict lung metastasis. Further analyses demonstrated that a subtype of four negative in breast cancer was signicantly associated with initial lung metastasis. Conclusions: Patients with pN3 stage and vascular invasion were more likely to develop lung metastasis. A new subtype with Her2 negative, both ER-negative and PR negative combination with p53 negative, had a great tendency to develop initial lung metastasis in breast invasive ductal cancer patients. Key words: breast invasive ductal carcinoma metastasis lung predictor

Downloaded from jjco.oxfordjournals.org by guest on October 10, 2010

INTRODUCTION
In breast cancer, metastasis is generally an organ-selective process whereby cancer cells metastasize to regional lymph nodes, the bone marrow, the lungs and the liver (1,2). The lung is an important distant organ to be affected by metastasis from breast cancer. It is estimated that 60 70% of patients who eventually die of breast cancer have it in their
For reprints and all correspondence: Qi Liu, Department of Oncology, Shandong Provincial Hospital Afliated to Shandong University, 324 Jingwu Road, Jinan 250021, Peoples Republic of China. E-mail: liuqileader@hotmail.com These authors contributed equally to this work.

lungs. In 21% of cases, the lung is the only site of metastasis (3,4). Therefore, it is important to study the selective biology mechanism underlying breast cancer metastasis to lung as the rst site. Several studies have made great advance in the mechanisms involved in organ-specic metastatic processes to lung from breast cancer at different levels. At the gene level, Minn et al. (5) identied a set of genes that marked and mediated breast cancer metastasis to the lungs. As a result, nine genes served dual functions, providing growth advantages both in the primary tumor and in the lung microenvironment. Three genes contributed to aggressive growth selectively in the lung. Gupta et al. (6) revealed inhibitor of

# The Author (2008). Published by Oxford University Press. All rights reserved.

98

Initial lung metastasis from breast cancer

cell differentiation (ID gene family)-mediated tumor re-initiation during breast cancer lung metastasis in the triple negative [TN, estrogen receptor (ER) negative, progesterone receptor (PR) negative and human epidermal growth factor receptor-2, also named Her2 negative] subgroup of human breast cancer based on an animal model and clinical studies. Calvo et al. (7) reported that ve genes were identied from VEGF-regulated genes associated with increased lung metastatic potential in the animal model and in patients with breast cancer. The result suggested that tenascin-C may be a relevant target for therapy against lung metastatic breast cancer (7). In the molecular aspect, Padua et al. (8) not only discovered the TGFb-primed breast cancer for lung metastasis but also made a clear explanation that the selection process of lung metastasis from breast cancer was the induction of angiopoietin-like 4 (ANGPTL4) by TGFb through the Smad-signaling pathway. These researches offer the promise that these genes and molecules as potential targets block breast cancer cells metastasizing to lung in the near future, also costly and specic. However, it will take a long time to put these targets into clinical practice, and its technology is also costly and complicated. It is unknown that which subtype in breast cancer needs this gene detection or target treatment. Therefore, further investigation is needed to identify high-risk subtypes in breast cancer patients who develop initial lung metastasis from the current routine examinations and treatment modes. Some factors from routine examinations have been reported to predict initial lung metastasis. Hasebe et al. (9) found in 1044 patients with invasive ductal carcinoma of the breast that histopathologic factors were associated with initial organ-specic metastasis. They also found that tumor necrosis, ER negative and progesterone receptor (PR) negative, skin invasion, N3 and lymph vessel invasion were signicantly related to lung metastasis as the initial organ. Histopathologic examination is routinely performed at any hospital, and it is a very useful tool in clinical practice in IDC patients. The results in the literature were of prime importance in the prediction of rst organ metastasis according to pathological characteristics. Nevertheless, they focused only on histological factors. There are still many other important factors such as Her2/neu, which remain unknown in relation to initial lung metastasis. In an early work, Schlappack et al. (4) described only the clinical course of lung metastasis from breast cancer and did not show which factors were related to lung metastasis as the rst organ. Many researchers have devoted great efforts to searching for risk factors that can predict the bone, the liver, the brain and unusual sites as the rst site of breast cancer metastasis (10 15). To our knowledge, few reports considered initial pulmonary metastasis from IDC as the main endpoint in clinical research. Many routine examinations are conducted when primary breast cancer is diagnosed and treated. Moreover, with the mature of immunohistochemistry (IHC) technology, molecular pathology examination is becoming routine in breast

cancer. These examinations afford much useful information, which have been used to prognosticate clinical outcomes or predict benets of a given treatment. But it is unknown from this information whether these inuence the process of metastasis to the lung. Even for treatment methods, they play an important role in the decision-making of breast cancer prognosis, but it remains unknown whether they can inuence the development of lung metastases from IDC. The current report was a detailed evaluation of risk factors for initial lung metastasis, by analysis of a population of 78 patients with Stages I III primary IDC based on clinical, laboratory and histopathologic characteristics and adjuvant treatment modes. The objectives in the current study were to explore (i) the characteristics of initial lung metastasis from IDC and (ii) high-risk factors for lung metastasis as the rst distant organ metastasis.
Downloaded from jjco.oxfordjournals.org by guest on October 10, 2010

PATIENTS AND METHODS

REQUIREMENTS
FOR ELIGIBLE PATIENTS

Patients were selected from a database. All patients had accepted complete treatment and follow-up in our hospital so that detailed information was available. Eligibility requirements included the following. (i) At primary diagnosis, all patients underwent routine examinations. They were diagnosed at Stages I III and could be operable based on these tests. The data must be accurate and available in age, Eastern Cooperative Oncology Group (ECOG) status, body mass index (BMI), menopausal status, menarche age, blood routine tests, hepatic tests and tumor markers tests, which were regarded as possible risk factors in the current study. Young women were considered to be no more than 40 years. ECOG status should be 0 or 1. Overweight was described by BMI . 25 kg/m 2 (16). Qualitative analyses were conducted as category data. Routine blood tests were white cell count (WBC), red cell count (RBC), hemoglobin (HGB) and platelet count (PLT). Hepatic tests included only total protein (TP) and globular protein (GP). The other items in hepatic tests were excluded because they were often not available in the data recorder. Tumor markers routinely measured at our hospital were carcinoembryonic antigen (CEA) and cancer antigen 15.3 (CA15-3). The CEA was said to be elevated above a value of 5 ng/ml, and CA15-3 was elevated above 35 U/ml. Age, menarche age, blood routine items, TP, GP, CEA and CA15-3 were analyzed in the research as quantitative factors. (ii) After surgery, pathological examination was conducted. Patients must have invasive ductal carcinomas of the breast. Primary tumor ( pT) stage, lymph node stage ( pN), vascular invasion, histological grade should be well recorded. Good records of primary tumors and lymph nodes were kept in the pathological reports so

Jpn J Clin Oncol 2009;39(2)

99

that pT and pN could be reviewed based on the UICC/ AJCC system (6th edition) (17). pT stage referred to T1: tumor is 2 cm or less across. T2: tumor is more than 2 cm but not more than 5 cm across. T3: tumor is more than 5 cm across. T4: tumor of any size growing into the chest wall or skin. pN stage referred to N1: 1 3 axillary (underarm) lymph node(s), N2: cancer has spread to 4 9 axillary lymph nodes, N3: 10 or more axillary lymph nodes or spread to the lymph nodes above the clavicle. pTNM stage was at Stages I III. Vascular invasion referred to the clear-cut presence of tumor cells within endothelial-lined spaces; no effort was made to distinguish between lymph and blood vessel invasion. The Nottingham system was used for histological grading (18): 1 presented low grade, 2 presented median grade and 3 presented high grade. Molecular pathological tests were also conducted. They included ER and PR and Her2, also named c-erb-B2, Ki-67 and p53 protein. They were evaluated on parafn sections using the IHC technology. The cases were considered positive to ER, PR, p53 and Ki-67 protein expression if more than 10% of the tumors cells showed nuclear staining. The Her2 staining was interpreted as positive when . 80% of the tumor cells showed distinct cell membrane staining. The methods on IHC and criterions of positive to negative in our hospital were consistent with the International Breast Cancer Study Group. (iii) For treatment, patients must accept modied radical or radical mastectomy. Other modes would be excluded because surgery methods maybe inuence breast cancer metastasis. If indicated, postoperation radiation was conducted. Radiotherapy with 50 60 Gy (median, 50 Gy) was given to the chest wall and regional lymph nodal areas. Radiotherapy was performed with similar techniques to similar elds. The chest wall was given 50 Gy in 25 fractions using two tangential elds by 60 Co or 6 8 MV X-rays. The draining lymph node regions (supraclavicle and axilla) were treated by a separate eld with 50 Gy in 25 fractions. The internal mammary nodal areas were treated with 45 50 Gy with 1.8 2 Gy per fraction for medially located tumors only. If patients received adjuvant chemotherapy, they must receive at least two cycles of anthracycline-containing regimens before metastasis. The endocrine therapy regimen used was tamoxifen, for at least 3 months before rst organ metastasis. Other adjuvant modes or drugs would be excluded so as to maximize the decreasing treatment bias. In the study, radiation, chemotherapy and endocrine therapy were, respectively, analyzed as possible risk factors. (iv) At the rst presentation, good records of metastatic sites should have been kept. In follow-up, the organ metastasis was diagnosed based on clinical symptoms, physical examination and radiographic tests. The ultimate diagnosis of the metastatic site was based on signicant evidence, using computed tomography (CT), ultrasound

and emission computed tomography (ECT). If two or more sites presented at the same time, they would be excluded from the study. If sites were sequential in a patient, the rst site was included and others were not. Data were collected between January 2003 and December 2007 from patient les with IDC in the Department of Medical Records, Shandong Provincial Hospital, China. A total of 432 patients les with IDC were screened and 354 patients were excluded because they did not meet the above requirements. A total of 78 patients who meet the above requirements were considered eligible and analyzed in the present study. All patients were Chinese women. Informed consent was obtained from each patient before primary treatment. The study was approved by the institutional review board as it has no conicts with the Helsinki Declaration.
Downloaded from jjco.oxfordjournals.org by guest on October 10, 2010

STATISTICAL

METHODS

The endpoint was rst organ metastasis. Lung metastasis was compared with other organs metastasis. Distant metastasis-free interval (DFS) was also calculated. This was dened as the interval between the date of primary treatment, which may be operation or new adjuvant chemotherapy on breast cancer, and the date of rst presentation with metastatic disease diagnosed by clinical methods. Univariate analyses between qualitative factors and rst distant organ (the lung and other organs) used x2 test or Fishers exact test as appropriate. The Mantel Haenzel x 2 test was used for quantitative variables and multiordinal factors. If necessary, correlations between factors were done using the Kendalls rank test. Signicant factors produced by the above methods were analyzed using multivariate analyses based on conditional logistic regression, in a stepwise manner. The mutual effects between vascular invasion and pN stage, p53 and ER/PR status were also analyzed in the multivariate analyses. Logistic regression analysis was used to calculate odds ratios (ORs), 95% condence intervals (CIs) and corresponding P values after adjusting for age as a covariate. Estimates were considered statistically signicant for twotailed values of P , 0.05. All analyses were performed using SPSS 13.0 statistical software (SPSS Inc, Chicago, IL, USA).

RESULTS
The distribution of metastatic disease in the patient group was shown in Tables 1 and 2. Lung metastases were the second most frequently observed metastases. A total of 17 (21.8%) patients presented initial lung metastases. The most frequent site was the bone that accounted for 56.4% of all cases. Liver and brain were the following sites. Median ages in patients with initial lung metastases were more than other patients with other rst metastatic sites. DFS in lung metastasis (26 months) patients was shorter than the bone (42 months) but longer than the brain and liver metastases (14 months).

100

Initial lung metastasis from breast cancer

Table 1. Characteristics of lung metastasis versus other organs metastasis in clinical characteristics Lung (n 17) Bone Cases % n DFS, month (median) Age, year (median) M, year (median) WBC (109/l) RBC (1012/l) HGB (g/l) PLT (109/l) TP (g/l) GP (g/l) CEA (ng/ml) CA15-3 (U/ml) 21.8 17 26 53 15 5.51 + 0.84 4.35 + 0.27 124.3 + 9.6 220.4 + 63.9 70.7 + 4.14 30.47 + 3.20 3.46 + 1.67 21.14 + 11.48 56.4 44 42 49 15 5.67 + 0.61 4.45 + 0.39 129.5 + 10.1 234.2 + 51.4 70.5 + 3.31 30.23 + 2.36 4.54 + 2.72 27.90 + 17.81 11.5 9 14 47 14 5.77 + 0.95 4.57 + 0.31 121.4 + 9.6 195.0 + 44.7 72.3 + 4.34 31.32 + 2.49 6.67 + 4.14 45.19 + 22.9 9.0 7 14 51 14 5.86 + 0.94 4.42 + 0.41 130.3 + 9.9 184 + 32.2 73.1 + 5.18 31.14 + 2.10 3.11 + 2.24 42.86 + 9.90 1.3 1 38 61 13 5.46 4.3 124 178 Other organ as rst sites (n 61) Liver Brain Pleural

Downloaded from jjco.oxfordjournals.org by guest on October 10, 2010

67 29 2.3 5.36

All P values in these factors from age to CA15-3 were, respectively, 0.067, 0.637, 0.374, 0.336, 0.143, 0.586, 0.331, 0.710, 0.145 and 0.315. They were more than 0.05. Age was considered as an adjusted covariate in multivariate analyses. Others were not conducted multivariate analyses. M, menarche age; DFS, Distant metastasis-free interval; WBC, white cell count; RBC, red cell count; HGB, hemoglobin; PLT, platelet count; TP, total protein; GP, globular protein; CEA, carcinoembryonic antigen.

The P values of quantitative factors showed in Table 1 between lung and other organs were all more than 0.05. Multivariate analyses were not conducted on these factors. Table 2 showed the results of the univariate analysis for ordinal factors. Two factors were signicant correlation with lung metastases. These were ER and PR status (Mann Whitney U 373.00, P 0.049), Her2 status (Pearsons x2 4.983, P 0.026). In Table 2, it was also observed that cases with ER/PR negative, and Her2 negative are more than others. The percentage of them were respectively 51% and 52%, which implied that it might be a TN subtype that was prone to initial lung metastasis. Further statistical analyses showed the TN subtype with signicant statistics (Pearsons x2 5.751, P 0.016). Clinical factors, routine pathological factors and treatment modes did not show signicant association with initial lung metastasis. Although treatment had an important inuence on patient outcomes, it seemed that modes of adjuvant therapy did not predict lung metastasis in this study. There were signicant correlations between ER/PR status and p53 protein (Kendalls tall b coefcient 2 0.232, P 0.036), vascular invasion and pN stage (coefcient 0.209, P 0.044). The mutual effects of ER/PR combination with p53 and pN stage with vascular invasion could not be avoided and multivariate analyses in logistic regression were conducted. Although P53, pN stage and vascular invasion were not statistically correlated with lung metastasis in univariate analyses, they were considered biological factors in relation to breast cancer, multivariate analyses were conducted. Table 3 showed the results of multivariate analysis.

Her2 still kept signicant statistics (P 0.005, OR 0.065). Mutual effect of ER/PR status and p53, pN stage and vascular invasion was also, respectively, signicant. Stratifying analysis was conducted on distribution of lung metastases cases in ER/PR status and p53, pN stage and vascular stage (Table 4). We could see cases in ER negative, PR negative and p53 negative were more than those of other levels. Similar results can also be seen in pN3 combined with vascular invasion positive. Cases in ER, PR and p53 all negative and pN3 combined with vascular invasion, respectively, accounted for 46% and 35% of all lung metastases patients. When Her2 negative was considered, cases in four negative concerns [FN, (Her2, ER/PR and p53) negative] were 41% of all lung metastases, implying that FN might be a new subgroup that had a great tendency to develop initial lung metastasis. Multivariate analysis could not be conducted because of few cases. As a result, univariate analysis was conducted and showed that FN was signicantly associated with initial lung metastasis (Fishers exact test, P 0.045).

DISCUSSION
Metastasis is a crucial event in the progression of breast cancer (19). It is the most common cause of death for breast cancer patients. As breast cancer is a heterogeneous disease, its metastatic behaviors are also diverse (1,2,9 12,19,20). IDC is one of the major types of breast cancer according to pathological examination, accounting for nearly 70% 2 80%

Jpn J Clin Oncol 2009;39(2)

101

Table 2. Univariate analyses results of lung metastasis versus other organs as rst metastasis sites in clinical characteristics, pathological features and treatment modes Factors Total cases (n 78) Lung (n 17) Other organs (n 61) P value

Table 2. Continued

Factors

Total cases (n 78)

Lung (n 17)

Other organs (n 61)

P value

ER/PR status Age (years)  40 . 40 14 (18%) 64 (82%) 1 (5%) 16 (95%) 13 (21%) 48 (79%) 0.281 Double negative Double positive Others 6 (35%) 11 (65%) 34 (56%) 27 (44%) 0.136 Her2 status Negative Positive 65 (83%) 13 (17%) 13 (76%) 4 (24%) 52 (85%) 9 (15%) 0.136 Ki-67 status 41 (52%) 37 (48%) 13 (76%) 4 (24%) 28 (46%) 33 (54%) 0.026 40 (51%) 31 (39%) 7 (10%) 12 (70%) 5 (30%) 0 (0%) 28 (46%) 26 (43%) 7 (11%) 0.049

Menopausal status Pre Post ECOG status 0 1 BMS  25 . 25 UICC pT stage T1 T2 T3 T4 UICC pN stage N0 N1 N2 N3 23 (29%) 20 (26%) 19 (25%) 16 (20%) 4 (24%) 2 (12%) 4 (24%) 7 (40%) 19 (31%) 18 (30%) 15 (25%) 9 (14%) 0.062 17 (22%) 43 (55%) 14 (18%) 4 (5%) 2 (12%) 13 (76%) 2 (12%) 0 (0%) 15 (25%) 30 (49%) 12 (20%) 4 (6%) 0.878 64 (82%) 14 (18%) 13 (76%) 4 (24%) 52 (85%) 9 (15%) 0.491 40 (51%) 38 (49%)

Downloaded from jjco.oxfordjournals.org by guest on October 10, 2010

Negative Positive p53 status Negative Positive Radiation No Yes Chemotherapy No Yes

51 (65%) 27 (35%)

13 (76%) 4 (24%)

38 (62%) 23 (38%) 0.277

57 (73%) 21 (27%)

10 (59%) 7 (41%)

47 (77%) 14 (23%) 0.214

35 (45%) 43 (55%)

8 (47%) 9 (53%)

27 (44%) 34 (56%) 0.838

5 (29%) 73 (71%)

2 (12%) 15 (88%)

3 (5%) 58 (95%) 0.298

Endocrine therapy No Yes 27 (35%) 51 (65%) 4 (24%) 13 (76%) 23 (38%) 38 (62%) 0.277

Vascular Invasion Absent Present Histologic grade 1 2 3 8 (10%) 44 (56%) 26 (33%) 1 (6%) 13 (76%) 3 (18%) 7 (11%) 41 (76%) 13 (22%) 0.929 61 (78%) 17 (22%) 9 (53%) 8 (47%) 52 (85%) 9 (15%) 0.822

Triple negative (TN) No Yes 55 (71%) 23 (29%) 8 (47%) 9 (53%) 47 (77%) 14 (23%) 0.016

Four negative (FN) No Yes 61 (78%) 17 (22%) 10 (59%) 7 (41%) 51 (84%) 10 (16%) 0.045

UICC pTNM stage I II III ER status Negative Positive PR status Negative Positive 43 (55%) 35 (45%) 11 (65%) 6 (35%) 32 (52%) 29 (48%) 0.369 Continued 43 (55%) 35 (45%) 11 (65%) 6 (35%) 32 (53%) 29 (48%) 0.369 10 (13%) 29 (37%) 39 (50%) 1 (6%) 5 (29%) 11 (65%) 9 (15%) 24 (39%) 32 (53%) 0.149

ECOG, Eastern Cooperative Oncology Group; BMS, body mass index; UICC, International Union Against Cancer; pTNM, Pathological TumorNode-Metastasis Staging; ER, estrogen receptor; PR, progesterone receptor; TN, triple negative subgroup (ER/PR double negative and Her2 negative); FN, four negative subgroup (all negative of ER/PR and p53, Her2 negative); Her2, epidermal growth factor receptor-2. Bold indicates P , 0.05. Table 3. Multivariate analyses results Factors Vascular invasion by pN stage Her2 P53 by ER/PR statusa
a a

P value 0.007 0.005 0.019

OR 2.600 0.065 3.285

95% CI 1.3175.136 0.0090.443 1.2268.803

These factors were taken account into mutual effect in logistic regression.

102

Initial lung metastasis from breast cancer

Table 4. Case distribution of patients with lung metastasis in vascular invasion combination with pN stage, p53 negative combination with double negative of ER/PR, and p53 negative combination with triple negative breast cancer pN stage Vascular invasion Positive (n 8) No. of cases N0 N1 N2 N3 ER/PR status 2 0 0 6 p53 status Positive (n 7) No. of cases ER/PR double negative (dn) ER/PR double positive (dp) ER/PR other status ER/PR and Her2 status 4 3 0 p53 status Positive (n 7) No. of cases TN statusa dp and Her2 negative dn and Her2 positive dp and Her2 positive Other status
a

Negative (n 9) % 12 0 0 35 No. of cases 2 2 4 1 % 12 12 23 6

Total cases (n 17) No. of cases 4 2 4 7 % 24 12 23 41

Negative (n 10) % 23 18 0 No. of cases 8 2 0 % 47 12 0

Total cases (n 17)

Downloaded from jjco.oxfordjournals.org by guest on October 10, 2010

No. of cases 12 5 0

% 71 29 0

Negative (n 10) % 12 12 11 6 0 No. of cases 7 2 1 0 0 % 41 12 6 0 0

Total cases (n 17) No. of cases 9 4 3 1 0 % 53 24 18 6 0

2 2 2 1 0

TN triple negative subgroup (ER/PR double negative and Her2 negative).

of total breast cancer. It is also the leading cause of the death due to metastasis. Lung metastasis is commonly seen in IDC (1 3). But it is unknown which patients will develop lung metastasis and how the risk could be identied in the progress to lung even if all data are available. In our results, univariate analyses showed that ER and PR status (P 0.049) and Her2 (P 0.026) were high-risk factors for lung metastasis as the rst metastasis site. Case distribution given in Table 2 was more seen in negative of ER/PR and Her2 than others, suggesting ER/PR double negative status and Her2 negative status being high risk for IDC patients to develop lung metastasis. ER/PR status did not show signicant statistics while ER/PR combined status of p53 was signicantly associated with lung metastasis in multivariate analyses. Case distribution showed patients with initial lung metastasis are at a status of p53 negative, ER negative and PR negative, which meant ER/PR status with

p53 status was a high-risk factor for lung metastasis. pN stage in prediction of lung metastasis needs to be combined with vascular invasion. Her2 was still a strong factor in relation to initial lung metastasis. To further observe case distribution in these factors (Table 4), Her2 negative was an independent predictor, ER negative and PR negative with p53 negative as a whole can be good to predict lung metastasis, pN3 combined with vascular stage is also a powerful predictor. The similar results had been reported in previous studies. Dent et al. (21) showed that patients with TN breast cancers were at higher risk to develop visceral metastases. Another study reported that TN breast cancer was a better predictor for distant metastasis in African women than in white women (22). But they did not explain which visceral organs are prone to be affected. Hasebe et al. (9) demonstrated that ER/PR double negative, pN3 stage and lymph vessel invasion were high-risk factors for initial lung

Jpn J Clin Oncol 2009;39(2)

103

metastasis. Our results agree with ER/PR double negative, vascular invasion, pN3 stage. The differences between Hesebe et al. (9) and ours were tumor necrosis and skin invasion, which were absent in our data and analysis could not be conducted in our study. Although there were four T4-stage patients in our study, only one patient presented skin invasion, who developed to brain metastasis, and others presented chest wall progressed to the bone. In other reports, pN3 and lymph vessel invasion are independent factors for prediction of metastasis (23,24). In our results, vascular invasion as an independent factor showed no relation to lung metastasis. But pN stage was signicantly associated with vascular invasion, suggesting vascular invasion be mainly lymph vessel invasion. In the Hasebe et al. (9) report, pN3 was not a good prediction factor for the lung but for the bone. In our results, lymph vessel invasion is necessary to predict initial lung metastasis. ER negative and PR negative also need to combine with p53 negative for the prediction. The results of the current study show that more attention should be paid to TN breast cancer. TN breast cancer with a poor prognosis still presents a great challenge in the treatment and development of novel therapies for oncologists (21,22,25,26). Its metastatic patterns were different from other subgroups (21,22). Dent et al.s (21) report showed a high risk factor for visceral metastases in TN patients shortly after diagnosis was responsible in large part for the adverse prognosis. The main sites of metastases for patients with this subgroup were the lung and brain. We further found that IDC patients with TN and p53 negative were prone to develop lung metastasis. In our results, lymph node positive was common in TN IDC patients (69% versus 31%), which was similar to western countries (27). Compared with previous studies, we further found a subtype of TN breast cancer with p53 negative was more sensitive to metastasis to the lung. The new characteristics of this phenotype should be paid more attention if the result could be repeated in future study and other institutions. However, the most important prognostic or predictive factors for recurrence or metastasis and survival in primary breast cancer were the histopathologic grades, the primary tumor and the regional nodes, pTNM stage in other studies (28,23,29). In our results, they had no inuence on lung metastasis as the rst organ of recurrence except pN3 in combination with vascular invasion, suggesting routine pathology tests based on morphological features are limited in this study to characterize organ-specic metastasis. Although there were controversies about the molecular labels of p53, Her2/neu etc. in their prognosis of the survival or prediction of organ-specic metastasis roles (9,29), the methods were established for detection of lung metastasis and had been put into clinical practice for more than decades. Moreover, the results in the present study testied previous studies and showed that the labels were powerful in the prediction of lung metastasis, meaning molecular pathology examination was useful and could explain different biological mechanisms in organ-selective metastasis for IDC.

Young age was considered to be a high-risk factor for development to brain metastasis (10). In our results, young women did not show a signicant relation to the lung and brain metastases while older women show initial lung metastasis more than those with other sites. But as older women with breast cancer were also more common, it is necessary to conrm age as a high-risk factor for initial lung metastasis in the future study. In the current study it was considered an adjusted factor. Tumor markers are now widely used in clinical practice to reect treatment response or recurrence of breast cancer. CA 153 was reported to be a good predictor for distant metastases (14). In the current report, CEA and CA 153 at primary diagnosis were taken into account and did not predict lung metastasis. Other clinical factors were also not high risk factors. It was most important that adjuvant therapies had no inuence on progression in metastasis to the lung for IDC, implying new advances to be explored in the treatment of IDC patients who had great tendency to develop lung metastasis, especially in TN breast cancer (26,27). Nowadays, we are entering into a molecular era in cancer detection and treatment. The ultimate goal in cancer management is to give right treatment to match right patients (30). It is necessary to discover new biomarkers for different cancers (31). For IDC with lung metastasis as the rst organ, it is a special subgroup. Some genes and special molecular areas have been found to mediate breast cancer metastasis to lung (5 8,31). New drugs are expected to be soon explored with these advances. Our results are to discover these patients so that they can receive target genes detection and new target drugs treatment to specically block metastasis to lung in further clinical trials. In conclusion, this is the rst study to clearly demonstrate important predictive factors for metastasis to lung based on current routine examinations and treatment modes. The pathological test is still the most useful tool that identies special behavior subgroups from different breast cancers among them. Moreover, molecular pathology is becoming routine examination and playing a more important role in discriminating high-risk factors for lung metastasis from IDC patients. They show that IDC patients with pN3 combination with vascular invasion or four negatives of ER, PR, Her2/neu, P53 have great a tendency to develop to initial lung metastasis. The results will decrease the effort to discover high-risk subgroups in IDC patients who develop lung metastasis and save costs in the detection of genes or individual treatment in the future. However, the study is established on retrospective study and the results may be uncertain to some extent. We will carry out a prospective study involving ID genes and TGF beta to further conrm the results on IDC patients.

Downloaded from jjco.oxfordjournals.org by guest on October 10, 2010

Conict of interest statement

None declared.

104

Initial lung metastasis from breast cancer

References
1. Kamby C, Andersen J, Ejlertsen B, Birkler NE, Rytter L, Zedeler K, et al. Pattern of spread and progression in relation to the characteristics of the primary tumour in human breast cancer. Acta Oncol 1991;30:301 8. 2. Jain S, Fisher C, Smith P, Millis RR, Rubens RD. Patterns of metastatic breast cancer in relation to histological type. Eur J Cancer 1993;15:2155 57. 3. Tomin R, Donegan WL. Screening for recurrent breast cancerits effectiveness and prognostic value. J Clin Oncol 1987;5:627. 4. Schlappack OK, Baur M, Steger G, Dittrich C, Moser K. The clinical course of lung metastases from breast cancer. Klin Wochensch 1988;66:7905. 5. Minn AJ, Gupta GP, Siegel PM, BoS PD, Shu W, Giri DD, et al. Genes that mediate breast cancer metastasis to lung. Nature 2005;436:518 24. 6. Gupta GP, Perk J, Acharyya S, de Candia P, Mittal V, Todorova-Manova K, et al. ID genes mediate tumor reinitiation during breast cancer lung metastasis. Proc Natl Acad Sci USA 2007;104: 1950611. 7. Calvo A, Catena R, Noble MS, Carbott D, Gil-Bazo I, Gonozales-Moreno O, et al. Identication of VEGF-regulated genes associated with increased lung metastatic potential: functional involvement of tenascin-C in tumor growth and lung metastasis. Oncogene 2008;27:537384. 8. Padua D, Zhang XH, Wang Q, Nadal C, Gerald WL, Gomis RR, et al. TGFbeta primes breast tumors for lung metastasis seeding through angiopoietin-like 4. Cell 2008;133:6677. 9. Hasebe T, Imoto S, Yokose T, Ishii G, Iwasaki M, Wada N. Histopathologic factors signicantly associated with initial organspecic metastasis by invasive ductal carcinoma of the breast: a prospective study. Hum Pathol 2008;39:68193. 10. Evans AJ, James JJ, Cornford EJ, Chan SY, Burrell HC, Pinder SE, et al. Brain metastases from breast cancer: identication of a high-risk group. Clin Oncol 2004;16:354 9. 11. Fulford LG, Reis-Fiho JS, Ryder K, Jones C, Gillett CE, Hanby A, et al. Basal-like grade III invasive ductal carcinoma of the breast: patterns of metastasis and long-term survival. Breast Cancer Res 2007;9:R4. 12. Maki DD, Grossman RI. Patterns of disease spread in metastatic breast carcinoma: inuence of estrogen and progesterone receptor status. Am J Neuroradiol 2000;21:1064 66. 13. James JJ, Evans AJ, Pinder SE, Gutteridge E, Cheung KL, Chan S, et al. Bone metastases from breast carcinoma: histopathological radiological correlations and prognostic features. Br J Cancer 2003;89:6605. 14. Keshaviah A, Dellapasqua S, Rotmensz N, Lindtner J, Crivellari D, Collins J, et al. CA153 and alkaline phosphatase as predictors for breast cancer recurrence: a combined analysis of seven International Breast Cancer Study Group trials. Ann Oncol 2007;18:7018.

15. Sanuki-Fujimoto N, Takeda A, Amemiya A, Ofuchi T, Ono M, Yamagami R, et al. Pattern of tumor recurrence in initially nonmetastatic breast cancer patients: distribution and frequency of metastases at unusual sites. Cancer 2008;113:677 82. 16. Dawood S, Broglio K, Gonzalez-Angulo AM, Kau SW, Islam R, Hortobagyi GN, et al. Prognostic value of body mass index in locally advanced breast cancer. Clin Cancer Res 2008;14:1718 25. 17. Singletary SE, Connolly LL. Breast cancer staging: working with the sixth edition of the AJCC cancer staging manual. CA Cancer J Clin 2006;56:37 47. 18. Elston CW, Ellis IO. Pathological prognostic factors in breast cancer. I. The value of histological grade in breast cancer: experience from a large study with long-term follow-up. Histopathology 2007;19: 40310. 19. Rabbani SA, Mazar AP. Evaluating distant metastases in breast cancer from biology to outcomes. Cancer Metastasis Rev 2007;26:66374. 20. Li CI, Uribe DJ, Daling JR. Clinical characteristics of different histologic types of breast cancer. Br J Cancer 2005;93:1046 52. 21. Dent R, Hanna WM, Trudeau M, Rawlinson E, Sun P, Narod SA. Pattern of metastatic spread in triple-negative breast cancer. Breast Cancer Res Treat 2008; DOI 10.1007/s10549-008-008-2. 22. Ihemelandu CU, Naab TJ, Mezghebe HM, Makambi KH, Siram SM, Leffall LD, Jr, et al. Basal cell-like (triple-negative) breast cancer, a predictor of distant metastasis in African American women. Am J Surg 2008;195:153 8. 23. Hacene K, Le Doussal V, Rouesse J. Predicting distant metastases in operable breast cancer patients. Cancer 1990;66:2034 43. 24. Hasebe T, Sasaki S, Imoto S, Ochiai A. Prognostic signicance of the intra-vessel tumor characteristics of invasive ductal carcinoma of the breast: a prospective study. Virchows Arch 2004;444:20 7. 25. Haffty BG, Yang Q, Reiss M, Kearney T, Higgins SA, Weidhaas J, et al. Locoregional relapse and distant metastasis in conservatively managed TN early-stage breast cancer. J Clin Oncol 2006;24:5652 57. 26. Nishimura R, Arima N. Is triple negative a prognostic factor in breast cancer? Breast Cancer 2008;15:3038. 27. Dent R, Trudeau M, Pritchard KI, Hanna WM, Kahn HK, Sawka CA, et al. Triple-negative breast cancer: clinical features and patterns of recurrence. Clin Cancer Res 2007;13:4429 34. 28. Kamby C, Sengelv L. Survival and pattern of failure following locoregional re-currence of breast cancer. Clin Oncol 1999;11:15663. 29. Reed W, Hannisdal E, Boehler PJ, Gundersen S, Host H, Marthin J. The prognostic value of p53 and c-erb B-2 immunostaining is overrated for patients with lymph node negative breast carcinoma: a multivariate analysis of prognostic factors in 613 patients with a follow-up of 14 30 years. Cancer 2000;88:804 13. 30. Oshaughnessy. Molecular signatures predict outcomes of breast cancer. N Engl J Med 2006;355:615 7. 31. Van de Vijver MJ, He YD, van t Veer LJ. A gene-expression signature as a predictor of survival in breast cancer. N Engl J Med 2002;347:19992009.

Downloaded from jjco.oxfordjournals.org by guest on October 10, 2010