AJCP / Original Article
Breast Cancer Subtypes Predispose the Site of Distant Metastases Abha Soni, DO, MPH,1 Zhiyong Ren, MD, PhD,1 Omar Hameed, MD,1,3 Diptiman Chanda, PhD,2 Charity J. Morgan, PhD,2 Gene P. Siegal, MD, PhD,1 and Shi Wei, MD, PhD1 From the 1Department of Pathology, School of Medicine and 2Department of Biostatistics, School of Public Health, University of Alabama at Birmingham, and 3Department of Pathology, Microbiology and Immunology, Vanderbilt University Medical Center, Nashville, TN. Key Words: Breast cancer; Metastasis; Subtype Am J Clin Pathol April 2015;143:471-478 DOI: 10.1309/AJCPYO5FSV3UPEXS
ABSTRACT Objectives: The distant organs to which breast cancer preferentially metastasizes are of significant clinical importance. Methods: We explored the relationship between the clinicopathologic factors and the common sites of distant metastasis in 531 consecutive patients with advanced breast cancer. Results: Breast cancer subtype as a variable was significantly associated with all five common sites of relapse by multivariate analysis. The luminal tumors were remarkable for their significant bone-seeking phenotype and were less frequently observed in lung, brain, and pleural metastases and less likely to be associated with multiorgan relapse. The HER2 subtype demonstrated a significant liverhoming characteristic. African Americans were significantly less likely to have brain-only metastasis in patients with brain relapse. Conclusions: These findings further articulate that breast cancer subtypes differ not only in tumor characteristics but also in their metastatic behavior, thus raising the possibility that this knowledge could potentially be used in determining the appropriate strategy for follow-up of patients with newly diagnosed breast cancer.
© American Society for Clinical Pathology
Breast cancer is the most common cancer among women of all ethnicities in the United States. Approximately 232,000 new cases of invasive breast cancer and 40,000 breast cancer deaths are expected to occur among US women annually.1 About 90% of this mortality is a consequence of metastatic diseases that are resistant to adjuvant therapies. There is a 20% to 30% likelihood of developing metastatic disease when diagnosed with breast cancer at an early age.2 The distant organs to which breast cancer preferentially metastasizes, of which bone, liver, lung, and brain are among the most common sites, are of clinical and biological importance and are closely related to the patients’ survival outcome.3,4 Factors influencing the development of breast cancer metastases largely include tumor size, histologic grade, lymphovascular spread, nodal involvement, and receptor status.5-7 It has long been accepted that, like many other tumor types, breast cancer spread demonstrates a nonrandom, organotropic metastatic pattern, a theory known as the “seed and soil” hypothesis first popularized by Paget in 1889.8 The organ-specific metastatic cells found to be preexistent in the parental cells were able to be enriched in vivo in their target organs, as first documented in experimental metastasis assays in the 1970s.9 The relationship between clinicopathologic factors and the patterns of distant relapse is of significant clinical interest but has not been well established. The skeleton is among the most commonly observed sites for distant metastases and represents the first site of relapse for approximately 50% of patients with breast cancer.3,10 It has been documented that patients with estrogen receptor (ER)–positive and progesterone receptor (PR)–positive tumors are more
Am J Clin Pathol 2015;143:471-478 471 DOI: 10.1309/AJCPYO5FSV3UPEXS
Soni et al / Breast Cancer Subtypes and Sites of Metastases
likely to have bone metastases,11,12 whereas patients with HER2-positive or so-called triple-negative (ER-, PR-, and HER2-negative) breast cancer (TNBC) have a predilection for visceral metastases, including to the brain.13,14 While these observations are important in treatment decision making and developing appropriate survey strategies, they are inefficient in elucidating the mechanisms responsible and have been rationalized away as secondary to the highly heterogeneous nature of the tumor. For example, a significant proportion of ER/PR-positive breast cancers does not undergo bone metastasis but develops visceral metastases, yet the tumors with isolated bone metastases are known to have significantly different receptor expression profiles from those primary tumors that develop bone plus other organ metastases, with the latter similar to that of visceraonly metastases.12 Exploring molecular targets for cancer therapy has become critical in the pursuit of personalized medicine. Research to that end has identified a number of breast cancer molecular subtypes with prognostic significance.15,16 With regard to gene expression profiling and the preferential site(s) of distant metastasis, patients with bone relapse have been most abundant among those whose tumors were of the luminal subtypes.17 Gene expression profiling further revealed distinct gene expression signatures for different organ-specific metastatic breast cancer variants. A “lung metastasis gene expression signature” was identified in a mouse xenograft model with human primary breast cancers and predicted for patients at high risk for selective distant relapse to the lungs.18,19 A subset of genes as potential brain metastasis mediators has also been established recently.20 The hindrances of gene expression analysis are its costliness and inconsistent reproducibility due to the highly heterogeneous nature of breast cancer, thus suggesting it may not be applicable in routine clinical practice. In this study, we sought to explore the possible relationship between the clinicopathologic factors of the primary tumor and the common sites of distant metastases in a large cohort of patients with advanced breast cancer to aid in the development of improved disease surveillance modalities.
Materials and Methods After approval by the institutional review board of the University of Alabama at Birmingham, the tumor registry at the authors’ institution was searched to identify breast cancer cases with associated metastasis between 1997 and 2010. Those with locoregional recurrence (skin or soft tissue/chest wall adjacent to breast) or lymph node metastasis were excluded. A total of 531 cases with associated distant
472 Am J Clin Pathol 2015;143:471-478 DOI: 10.1309/AJCPYO5FSV3UPEXS
(bone, visceral organs, or brain) metastasis were identified. The patients’ demographic data and the pathologic features of the primary tumor were recorded, including the patient’s age at diagnosis, race, tumor type, histologic grade, and ER, PR, and HER2 status. The accuracy of the data was further validated for each patient using the electronic medical record and/or surgical pathology database. The clinical stage of breast cancer was based on imaging studies (conventional radiographs, computed tomography, or magnetic resonance imaging) with or without confirmation by tissue biopsy. The ER and PR expression status was examined by immunohistochemistry, and HER2 protein overexpression and/or gene amplification was evaluated by immunohistochemistry or fluorescence/chromogenic in situ hybridization as previously described.21,22 ER and PR positivity was defined as 1% or more of tumor cell nuclei with immunoreactivity. HER2 positivity was defined as either a 3+ immunohistochemistry score (uniform and intensity membrane staining of >30% of tumor cells) or a positive in situ hybridization result. When a receptor (ER, PR, or HER2) status was reported to be unknown in the tumor registry and in the medical chart, the patient was coded as having unknown receptor status for that particular receptor. Breast cancer subtype classification was based on results of hormonal receptor and HER2 status as previously reported.12,16 In brief, tumors were defined as luminal (ER+ and/ or PR+), HER2 (ER–/PR–/HER2+), or TNBC (ER–/PR–/ HER2–). The luminal cancers were further separated into luminal A (ER+ and/or PR+/HER2−, Nottingham grades I-II) and luminal B (ER+ and/or PR+/HER2−, grade III or ER+ and/or PR+/HER2+). Although a subset of tumors with a triple-negative phenotype represents basal-like cancers, this group was not further subclassified. All patients received systemic therapy (endocrine therapy, targeted therapy, cytotoxic chemotherapy, or a combination). All but 119 patients with positive hormonal receptors received endocrine therapy, based on physician and patient discussions. Forty-six of 176 patients with HER2-positive disease received HER2-targeted therapy with trastuzumab. All but 114 patients with receptor-positive disease received cytotoxic chemotherapy. Statistical Analysis The association of clinicopathologic factors and distant organ relapse was modeled with logistic regression analysis. Both unadjusted (univariate) and adjusted (multivariate) odds ratios (ORs) and 95% confidence intervals (CIs) were calculated for each model. A two-sided P value of .05 or less was considered statistically significant. All analyses were conducted by using SAS version 9.3 (SAS Institute, Cary, NC).
© American Society for Clinical Pathology
AJCP / Original Article
❚Table 1❚ Association of Breast Cancer Subtype and the Sites of Relapse Univariate Analysis
Multivariate Analysis
Site of Relapse
χ2
P Value
χ2
P Value
Bone Liver Lung Central nervous system Pleura
23.2872 8.2959 18.4191 10.6705 8.2993
Breast Cancer Subtypes Predispose the Site of Distant Metastases Abha Soni, DO, MPH,1 Zhiyong Ren, MD, PhD,1 Omar Hameed, MD,1,3 Diptiman Chanda, PhD,2 Charity J. Morgan, PhD,2 Gene P. Siegal, MD, PhD,1 and Shi Wei, MD, PhD1 From the 1Department of Pathology, School of Medicine and 2Department of Biostatistics, School of Public Health, University of Alabama at Birmingham, and 3Department of Pathology, Microbiology and Immunology, Vanderbilt University Medical Center, Nashville, TN. Key Words: Breast cancer; Metastasis; Subtype Am J Clin Pathol April 2015;143:471-478 DOI: 10.1309/AJCPYO5FSV3UPEXS
ABSTRACT Objectives: The distant organs to which breast cancer preferentially metastasizes are of significant clinical importance. Methods: We explored the relationship between the clinicopathologic factors and the common sites of distant metastasis in 531 consecutive patients with advanced breast cancer. Results: Breast cancer subtype as a variable was significantly associated with all five common sites of relapse by multivariate analysis. The luminal tumors were remarkable for their significant bone-seeking phenotype and were less frequently observed in lung, brain, and pleural metastases and less likely to be associated with multiorgan relapse. The HER2 subtype demonstrated a significant liverhoming characteristic. African Americans were significantly less likely to have brain-only metastasis in patients with brain relapse. Conclusions: These findings further articulate that breast cancer subtypes differ not only in tumor characteristics but also in their metastatic behavior, thus raising the possibility that this knowledge could potentially be used in determining the appropriate strategy for follow-up of patients with newly diagnosed breast cancer.
© American Society for Clinical Pathology
Breast cancer is the most common cancer among women of all ethnicities in the United States. Approximately 232,000 new cases of invasive breast cancer and 40,000 breast cancer deaths are expected to occur among US women annually.1 About 90% of this mortality is a consequence of metastatic diseases that are resistant to adjuvant therapies. There is a 20% to 30% likelihood of developing metastatic disease when diagnosed with breast cancer at an early age.2 The distant organs to which breast cancer preferentially metastasizes, of which bone, liver, lung, and brain are among the most common sites, are of clinical and biological importance and are closely related to the patients’ survival outcome.3,4 Factors influencing the development of breast cancer metastases largely include tumor size, histologic grade, lymphovascular spread, nodal involvement, and receptor status.5-7 It has long been accepted that, like many other tumor types, breast cancer spread demonstrates a nonrandom, organotropic metastatic pattern, a theory known as the “seed and soil” hypothesis first popularized by Paget in 1889.8 The organ-specific metastatic cells found to be preexistent in the parental cells were able to be enriched in vivo in their target organs, as first documented in experimental metastasis assays in the 1970s.9 The relationship between clinicopathologic factors and the patterns of distant relapse is of significant clinical interest but has not been well established. The skeleton is among the most commonly observed sites for distant metastases and represents the first site of relapse for approximately 50% of patients with breast cancer.3,10 It has been documented that patients with estrogen receptor (ER)–positive and progesterone receptor (PR)–positive tumors are more
Am J Clin Pathol 2015;143:471-478 471 DOI: 10.1309/AJCPYO5FSV3UPEXS
Soni et al / Breast Cancer Subtypes and Sites of Metastases
likely to have bone metastases,11,12 whereas patients with HER2-positive or so-called triple-negative (ER-, PR-, and HER2-negative) breast cancer (TNBC) have a predilection for visceral metastases, including to the brain.13,14 While these observations are important in treatment decision making and developing appropriate survey strategies, they are inefficient in elucidating the mechanisms responsible and have been rationalized away as secondary to the highly heterogeneous nature of the tumor. For example, a significant proportion of ER/PR-positive breast cancers does not undergo bone metastasis but develops visceral metastases, yet the tumors with isolated bone metastases are known to have significantly different receptor expression profiles from those primary tumors that develop bone plus other organ metastases, with the latter similar to that of visceraonly metastases.12 Exploring molecular targets for cancer therapy has become critical in the pursuit of personalized medicine. Research to that end has identified a number of breast cancer molecular subtypes with prognostic significance.15,16 With regard to gene expression profiling and the preferential site(s) of distant metastasis, patients with bone relapse have been most abundant among those whose tumors were of the luminal subtypes.17 Gene expression profiling further revealed distinct gene expression signatures for different organ-specific metastatic breast cancer variants. A “lung metastasis gene expression signature” was identified in a mouse xenograft model with human primary breast cancers and predicted for patients at high risk for selective distant relapse to the lungs.18,19 A subset of genes as potential brain metastasis mediators has also been established recently.20 The hindrances of gene expression analysis are its costliness and inconsistent reproducibility due to the highly heterogeneous nature of breast cancer, thus suggesting it may not be applicable in routine clinical practice. In this study, we sought to explore the possible relationship between the clinicopathologic factors of the primary tumor and the common sites of distant metastases in a large cohort of patients with advanced breast cancer to aid in the development of improved disease surveillance modalities.
Materials and Methods After approval by the institutional review board of the University of Alabama at Birmingham, the tumor registry at the authors’ institution was searched to identify breast cancer cases with associated metastasis between 1997 and 2010. Those with locoregional recurrence (skin or soft tissue/chest wall adjacent to breast) or lymph node metastasis were excluded. A total of 531 cases with associated distant
472 Am J Clin Pathol 2015;143:471-478 DOI: 10.1309/AJCPYO5FSV3UPEXS
(bone, visceral organs, or brain) metastasis were identified. The patients’ demographic data and the pathologic features of the primary tumor were recorded, including the patient’s age at diagnosis, race, tumor type, histologic grade, and ER, PR, and HER2 status. The accuracy of the data was further validated for each patient using the electronic medical record and/or surgical pathology database. The clinical stage of breast cancer was based on imaging studies (conventional radiographs, computed tomography, or magnetic resonance imaging) with or without confirmation by tissue biopsy. The ER and PR expression status was examined by immunohistochemistry, and HER2 protein overexpression and/or gene amplification was evaluated by immunohistochemistry or fluorescence/chromogenic in situ hybridization as previously described.21,22 ER and PR positivity was defined as 1% or more of tumor cell nuclei with immunoreactivity. HER2 positivity was defined as either a 3+ immunohistochemistry score (uniform and intensity membrane staining of >30% of tumor cells) or a positive in situ hybridization result. When a receptor (ER, PR, or HER2) status was reported to be unknown in the tumor registry and in the medical chart, the patient was coded as having unknown receptor status for that particular receptor. Breast cancer subtype classification was based on results of hormonal receptor and HER2 status as previously reported.12,16 In brief, tumors were defined as luminal (ER+ and/ or PR+), HER2 (ER–/PR–/HER2+), or TNBC (ER–/PR–/ HER2–). The luminal cancers were further separated into luminal A (ER+ and/or PR+/HER2−, Nottingham grades I-II) and luminal B (ER+ and/or PR+/HER2−, grade III or ER+ and/or PR+/HER2+). Although a subset of tumors with a triple-negative phenotype represents basal-like cancers, this group was not further subclassified. All patients received systemic therapy (endocrine therapy, targeted therapy, cytotoxic chemotherapy, or a combination). All but 119 patients with positive hormonal receptors received endocrine therapy, based on physician and patient discussions. Forty-six of 176 patients with HER2-positive disease received HER2-targeted therapy with trastuzumab. All but 114 patients with receptor-positive disease received cytotoxic chemotherapy. Statistical Analysis The association of clinicopathologic factors and distant organ relapse was modeled with logistic regression analysis. Both unadjusted (univariate) and adjusted (multivariate) odds ratios (ORs) and 95% confidence intervals (CIs) were calculated for each model. A two-sided P value of .05 or less was considered statistically significant. All analyses were conducted by using SAS version 9.3 (SAS Institute, Cary, NC).
© American Society for Clinical Pathology
AJCP / Original Article
❚Table 1❚ Association of Breast Cancer Subtype and the Sites of Relapse Univariate Analysis
Multivariate Analysis
Site of Relapse
χ2
P Value
χ2
P Value
Bone Liver Lung Central nervous system Pleura
23.2872 8.2959 18.4191 10.6705 8.2993
