The significance of circulating tumor cells (CTCs) in blood and of disseminated tumor cells (DTCs) in bone marrow (BM) in patients with early stage breast cancer is unclear. In this study, the authors investigated the occurrence of CTCs and DTCs in women with early stage breast cancer and evaluated the correlation of their presence with other prognostic markers.
Blood and BM aspirations were collected at the time of primary breast surgery. CTCs were detected by using the CellSearch assay, and DTCs were detected by immunostaining BM aspirates for pancytokeratin. The presence of CTCs and DTCs was correlated with tumor classification (T1 vs T2), tumor histologic grade, estrogen receptor (ER) status, progesterone receptor (PR) status, human epidermal growth factor receptor 2 (HER2) status, and lymph node (LN) status.
Of 92 patients who were included in the study, 49 had T1 tumors, and 43 had T2 tumors. CTCs were detected in 31% of patients, and DTCs were detected in 27% of patients. There was no correlation between the occurrence of CTCs and DTCs with the tumor classification (T1 vs T2) or histologic grade. CTCs were detected in 33% of patients with ER-positive disease versus 26% of patients with ER-negative disease, in 32% of patients with PR-positive disease versus 30% of patients with PR-negative disease, and in 25% of patients with HER2-positive disease versus 31% of patients with HER2-negative disease. DTCs were observed in 23% of patients with ER-positive disease versus 37% of patients with ER-negative disease, in 22% of patients with PR-positive disease versus 32% of patients with PR-negative disease, and in 0% of patients with HER2-positive disease versus 29% of patients with HER2-negative disease. CTCs and DTCs were nearly equally prevalent in both LN-positive women and LN-negative women. There was no significant correlation between the occurrence of CTCs or DTCs with tumor classification (T1 vs T2), tumor histologic grade, positive ER status, positive PR status, or positive HER2 status, and axillary LN status.
Metastasis is the most common cause of death by far in patients with breast cancer. The complicated process of metastasis entails a cascade of events that begins with the dislodgement of tumor cells from the primary site, followed by their entry into lymphovascular spaces, and their eventual proliferation in a wide variety of homing sites. The dissemination of tumor cells to any of the homing sites can occur very early in the course of the disease, at a point when overt metastasis is not apparent, even with the use of high-resolution imaging modalities. Husemann et al1 demonstrated in human epidermal growth factor 2 (HER2) and polyomavirus middle T-antigen (PyMT) transgenic mouse models that tumor cells can disseminate systemically from the earliest epithelial alterations in the mammary gland. Those authors also reported that the number of disseminated cells and their karyotypic abnormalities were similar irrespective of the size of the primary tumor both in mouse models and in patients with cancer. It is believed that release from dormancy of early disseminated cancer cells is the most likely cause of metachronous metastasis.
Investigation of disseminated tumor cells (DTCs) in bone marrow from patients with breast cancer has been pursued actively in the last 2 decades by several investigators using mainly immunocytochemical staining of bone marrow aspirations with different epithelial markers for the detection of these rare cells, which occur in 1 of 106 to 107 hematopoietic cells. Braun et al2 conducted a pooled analysis of 4703 patients with stage I, II, and III breast cancer who were followed for >10 years (median follow-up, 5.2 years). Those investigators observed the presence of micrometastatic disease in bone marrow in 30.6% of patients at the time of breast cancer diagnosis and reported that such presence portends a poor prognosis. The American Society of Clinical Oncology, in its recommendation for the use of tumor markers in breast cancer, does not advocate the evaluation of DTCs in bone marrow, citing the lack of sufficient data to recommend such evaluation.3 In contrast to several studies that investigated DTCs in bone marrow in patients with operable breast cancer, it is only recently that circulating tumor cells (CTCs) have been investigated actively in breast cancer. Rapid advancements in the methodologies for the detection of CTCs are facilitating our understanding of CTCs, particularly their prognostic and predictive implications. The clinical relevance of detecting CTCs in different populations of patients with breast cancer is evolving rapidly. It has been demonstrated that the detection of CTCs with the CellSearch CTC test (Veridex, Warren, NJ) provides significant prognostic information before and after the initiation of chemotherapy in patients with metastatic breast cancer.4, 5 It also has been demonstrated that measuring CTCs is superior to conventional imaging for evaluating response to chemotherapy.6 However, there have been very few studies of CTCs in early breast cancer.
In this study, we investigated the occurrence of DTCs in bone marrow and CTCs in peripheral blood from patients with early stage breast cancer (American Joint Committee on Cancer [AJCC] tumor classification T1 and T2) and evaluated the correlation of their presence with standard prognostic markers, including tumor size (AJCC T1 vs T2), tumor histologic grade, estrogen receptor (ER) status, progesterone receptor (PR) status, human epidermal growth factor receptor 2 (HER2)/neu gene amplification and the status of axillary lymph nodes.
MATERIALS AND METHODS
Patients with early stage breast cancer (AJCC T1 and T2) were enrolled during their preoperative surgical consultation visit from February 2005 to September 2008 in this institutional review board-approved protocol (M. D. Anderson Cancer Center [MDACC] LAB 04-657, MDACC LAB 04-698; Principal Investigator, A. Lucci) for the detection of minimal residual disease in peripheral blood and bone marrow. All patients provided informed written consent before enrollment. None of the patients in the study received preoperative (neoadjuvant) chemotherapy. Bone marrow aspirates and peripheral blood samples were obtained from the patients under general anesthesia at the time of primary surgery before resection of the breast tumor.
Bone marrow aspirations (10 mL each) were collected from bilateral anterior superior iliac crests into tubes that contained ethylene diamine tetracetic acid. Enrichment of tumor cells was performed by density gradient separation using Ficoll-Hypaque solution. Ten cytospins were prepared from each bone marrow site. One of the cytospin preparation was used for Papanicolaou staining, and the remaining preparations were used for immunocytochemical staining. The Cytospin specimens of bone marrow aspirates, enriched for epithelial cells by density gradient separation, were immunostained using a pancytokeratin cocktail of antibodies, including AE1/AE3, CAM5.2, MNF116, cytokeratin 8 (CK8), and CK18. Distinct cytoplasmic and/or membranous staining of cells and clusters with cytomorphologic features consistent with a malignant cell were identified as DTCs.
Peripheral blood was collected in 3 tubes, each containing 7.5 mL of blood, for the detection of CTCs by the CellSearch CTC test. The CellSearch test was performed at MDACC and was evaluated by the same individual observer who was blinded to any clinical information regarding the patients in all of the cases. The CellSearch test uses human epithelial adhesion molecule (EpCAM)-coated ferrofluids for enriching CTCs, which are identified further using a combination of CK and CD45 fluorescent antibodies. CTCs were identified as nucleated cells that were positive for CK and negative for CD45. The presence of 1 or more CK-positive cells in 1 or more tubes was regarded as positive for CTCs in a given patient. The performance of the CellSearch test was evaluated using the CellSearch epithelial cell control kit before any of the patient samples were tested. This kit provides high and low populations of fixed cells obtained from a breast carcinoma cell line (SK-BR-3) that serve as the positive control of the test.
Primary Tumor Specimens
Samples of the primary breast tumor were fixed in Pen-fix buffered alcohol formalin fixative, processed routinely, and embedded in paraffin for cutting. Five-micrometer sections of the paraffin-embedded tissue blocks were obtained for hematoxylin and eosin (H&E) staining and were used for conventional pathologic examination and for the evaluation of prognostic and predictive markers. The primary breast tumors were graded histologically on a scale from 1 to 3 based on a modified, combined Nottingham histologic grading system. Immunostaining for ER, PR, and HER2 was performed using a Bond Max machine (Vision Biosystems, Norwell, Mass). Primary antibodies were used against ER (clone 6FII, dilution 1:35; Novocastra Laboratories Ltd., Newcastle upon Tyne, United Kingdom), PR (clone PgR 1294, dilution 1:200; Dako, Carpenteria, Calif), and HER2/neu (Lab Vision, clone neu Ab8, dilution 1:300; Lab Vision, Fremont, Calif), all with epitope retrieval in citrate buffer, pH 6.0. Primary breast tumors that expressed nuclear staining in ≥1% of tumor cells were regarded as positive for ER and PR. Immunostaining results for HER2 were scored as 1+ when <10% of the tumor cells had complete membranous staining; as 2+ when weak-to-moderate, complete membranous staining was present in >10% of tumor cells; and as 3+ when strong, complete membranous staining was present in >30% of tumor cells. All 2+ and 3+ cases were evaluated by fluorescence in situ hybridization (FISH) for HER2 gene amplification using the Abbott PathVysion HER2 DNA probe kit (Abbott Laboratories, Abbott Park, Ill). A HER2/CEP17 ratio >2.2 was regarded as positive for HER2 gene amplification.
Sentinel lymph nodes were sliced at 2-mm intervals, processed routinely, embedded in paraffin, and cut at 5 μm thickness for H&E staining of the first level and for CK immunostaining of the third level. Lymph nodes with metastatic tumor that measured >0.2 mm alone were regarded as positive for metastatic disease. Lymph nodes with isolated tumor cells that measured <0.2 mm were regarded as negative for metastatic disease in all evaluations. Nonsentinel lymph nodes were evaluated by H&E staining only.
The presence of CTCs in peripheral blood and DTCs in bone marrow was correlated with primary tumor characteristics, including tumor size, tumor histologic grade, ER, PR, HER2, and lymph node status by using the Pearson chi-square test in STATA10 software (StataCorp, College Station, Tex) and NQuery Advisor 6.01 (Statistical Solutions, Saugus, Mass).
The Fisher exact test was used when any 1 of the observed frequencies in the 2 × 2 contingency table was <5. P values <.05 were considered statistically significant.
There were 92 patients in the study, including 49 patients with T1 tumors and 43 patients with T2 tumors. The median patient age was 53.7 years, and the median tumor size was 2.1 cm. Most primary tumors were ductal in type (81%). The primary breast tumor was categorized as low grade in 4 patients, intermediate grade in 43 patients, and high grade in 45 patients. Invasive tumors were positive for ER in 66% of patients, positive for PR in 54% of patients, and positive for HER2/neu in 5% of patients.
Either sentinel or nonsentinel axillary lymph nodes were positive in 39 patients (42%) with early stage breast cancer; micrometastasis (0.2 mm to 2 mm in size) was present in 10 patients (26%), and macrometastasis was present in 29 patients (74%). Lymph node involvement was limited to 1 lymph node in 17 patients, to 2 or 3 lymph nodes in 13 patients, and to >3 lymph nodes in 9 patients. The clinicopathologic features of the patients with early stage breast cancer who were included in the study are summarized in Table 1.
Table 1. Clinicopathologic Features of Patients in the Study (n=92)
No. of Patients
NA indicates not applicable; HER2, human epidermal growth factor receptor 2.
Median age, y
Primary tumor histologic type
Hormone receptor status
Estrogen receptor positive
Progesterone receptor positive
Lymph node positive
Both CTCs and DTCs could be evaluated in the same patient in 63 of 92 patients who were included in the study. DTCs alone could be evaluated in 67 patients, and CTCs alone could be evaluated in 81 patients.
Circulating Tumor Cells
One or more CTCs were detected in 25 of 81 patients (31%) with early stage breast cancer. Figure 1 is an illustration of a CTC in peripheral blood from a patient who had T2 disease detected by the CellSearch assay. CTCs occurred in 13 of 43 patients (30%) with T1 tumors and in 12 of 38 patients (32%) with T2 tumors. There was no correlation between the occurrence of CTCs in patients who had T1 tumors and patients who had T2 tumors (P = .6) and with tumor histologic grade (P = .7). CTCs were equally prevalent in patients with the presence and absence of standard prognostic and predictive markers. CTCs were noted in 18 of 54 patients (33%) with ER-positive primary tumors, in 7 of 27 patients (26%) with ER-negative tumors (P = .5), in 14 of 44 patients (32%) with PR-positive tumors, in 11 of 37 patients (30%) with PR-negative tumors (P = .9), in 1 of 4 patients (25%) with HER2-positive tumors, and in 24 of 77 patients (31%) with HER2-negative tumors (P = .6). CTCs occurred in 10 of 35 patients (29%) with lymph node-positive disease compared with 15 of 46 patients (33%) patients with lymph node-negative disease (P = .7). CTCs were detected in 2 of 7 patients (29%) with micrometastases and in 7 of 28 patients (25%) with macrometastases in the lymph nodes. There was no statistically significant difference between the occurrence of CTCs and the number of lymph nodes involved with metastatic disease or between the presence of micrometastasis versus macrometastasis in the lymph nodes. Therefore, there was no correlation between the detection of CTCs and pathologic characteristics, including tumor histologic grade, ER status, PR status, HER2 positivity, and metastasis to axillary lymph nodes. The details of the correlation of CTCs with standard prognostic features are summarized in Table 2.
Table 2. Circulating Tumor Cells in Early Stage Breast Cancer (n=81)
No. of Patients
HER2 indicates human epidermal growth factor receptor 2.
P value was determined using the Fisher exact test.
Papanicolaou-stained direct smears of the density gradient-separated specimens of bone marrow aspirates did not reveal any recognizable tumor cells in any of the cases. Immunocytochemical staining of the bone marrow cytospin slides usually revealed a single DTC or rarely revealed DTC clusters in 18 of 67 patients (27%) with early stage breast cancer. Figure 2 shows a single CK-positive DTC in a patient who had T1 breast cancer. DTCs were noted in 12 of 37 patients (32%) with T1 disease and in 6 of 30 patients (20%) with T2 disease. The occurrence of DTCs in patients with T1 disease did not differ from that in patients with T2 disease (P = .5). There was no correlation between tumor histologic grade and the presence of DTCs (P = .8). DTCs were equally prevalent in patients with ER-positive and ER-negative tumors (11 of 48 patients [23%] vs 7 of 19 patients [37%], respectively; P = .3), in patients with PR-positive and PR-negative tumors (8 of 36 patients [22%] vs 10 of 31 patients [32%]; P = .4), and in patients with HER2-positive and HER2-negative tumors (0 of 4 patients [0%] vs 18 of 63 patients [29%]; P = .3). DTCs were noted in 8 of 31 patients (26%) patients who had positive axillary lymph node status compared with 10 of 36 patients (28%) who had no evidence of metastasis in the lymph nodes. DTCs were detected in 3 of 7 patients (43%) with micrometastases and in 6 of 24 patients (25%) with macrometastases in the lymph nodes. There was no statistically significant difference between the occurrence of DTCs with the number of lymph nodes that were positive for metastatic disease or with micrometastasis versus macrometastasis in the lymph nodes. Details of the correlations between DTCs and standard prognostic features are summarized in Table 3.
Table 3. Disseminated Tumor Cells in Early Stage Breast Cancer (n=67)
No. of Patients
HER2 indicates human epidermal growth factor receptor 2.
P value was determined using the Fisher exact test.
The detection of both CTCs and DTCs could be performed in only 63 of 92 patients in this study, because bone marrow could not be collected successfully from all patients. Five of 63 patients demonstrated the presence of both CTCs and DTCs. The clinicopathologic features of these 5 patients did not differ significantly from the features of patients who had either CTCs or DTCs or from the features of patients who were negative for minimal residual disease. We detected CTCs alone in 13 of 63 patients and DTCs alone in 15 of 63 patients. Essentially, there was no correlation between the occurrence of CTCs and the occurrence of DTCs in this study (P = .7). Thirty of 63 patients did not have any evidence of CTCs or DTCs. The clinicopathologic features of these 30 patients did not differ significantly from the features of patients who had minimal residual disease.
In the last 2 decades, several investigators have evaluated DTCs in bone marrow from patients with operable breast cancer. In the pooled analysis of 4703 patients reported by Braun et al, 2507 patients had T1 disease, and 1706 patients had T2 disease, and the incidence of DTCs in bone marrow for the 2 groups was 30.5% and 22.5%, respectively. These results are very similar to the incidence of 27% observed in our study. Our finding of no correlation between DTCs in bone marrow and standard prognostic factors in breast cancer is similar to findings from a few previous reports but differ from most.
In the pooled analysis of DTCs in bone marrow reported by Braun et al, patients with micrometastases in the bone marrow had larger tumors, higher histologic tumor grade, negative hormone receptor status, and positive axillary lymph node status. It is noteworthy that the correlation of DTCs with standard prognostic factors in that study was evaluated for the whole group of operable breast tumors (T1-T3) and not strictly for the T1 and T2 tumors that were analyzed in our current study. Similarly, Diel et al7 observed a significant correlation between DTCs evaluated by immunostaining of bone marrow aspirates for antitumor-associated glycoprotein-12 and tumor size, lymph node status, tumor grade, and postmenopausal status. The Ludwig Cancer Institute Group also reported that the presence of DTCs immunopositive for epithelial membrane antigen was associated significantly with tumor size, lymph node involvement, peritumoral vascular invasion, and tumor size.8
Two studies by Braun et al,9, 10 however, indicated the lack of a correlation of DTCs in bone marrow with axillary lymph node metastasis and pathologic tumor characteristics of the primary tumor, which we also observed in our study. One of those studies was conducted in 552 patients with T1, T2, and T3 disease, the other study included 150 lymph node-negative patients who had T1 and T2 primary breast cancer. The incidence of DTCs in those 2 studies was 36% and 29%, respectively.
The availability of standardized techniques for the detection of CTCs in peripheral blood has created tremendous advances in the investigation and research of CTCs in breast cancer. Currently, 2 commercial systems are available for detecting CTCs: the CellSearch test and the products by AdnaGen (Langenhagen, Germany). Currently, the CellSearch test is approved by the US Food and Drug Administration for detecting CTCs in metastatic breast cancer. This is because most studies related to CTCs have been conducted in metastatic breast cancer, and very few studies have evaluated CTCs in early stage breast cancer. Most of those studies used reverse transcriptase-polymerase chain reaction (RT-PCR) to identify the messenger RNA of selected markers that detect CTCs in the peripheral blood from patients with early stage breast cancer.11-14 The results of previous studies of CTCs in early stage breast cancer compared with our results are summarized in Table 4. The lack of studies that directly compared the CellSearch test with RT-PCR assays for different molecular markers precludes making a stringent comparison of the results from our studies in early stage breast cancer with the results from other studies that used molecular markers.
Table 4. Studies of Circulating Tumor Cells in Early Stage Breast Cancer
The CellSearch system has been used to investigate CTCs in peripheral blood from patients with primary breast cancer in very few prospective studies to date. Lang et al15 detected CTCs with the CellSearch system in 38% of patients who had operable breast cancer (T1-T4 tumors; 33% received neoadjuvant chemotherapy) and reported a median of 1 CTC in each patient who was positive for CTCs. In their patient population, which included patients with early stage disease as well as patients who received preoperative chemotherapy, the HER2 status of the primary tumor alone was associated significantly with the detection of CTCs in the peripheral blood. Rack et al16 evaluated the role of CTCs in peripheral blood from 1767 patients with lymph node-positive and lymph node-negative breast cancer before systemic treatment and, in 852 of their patients, during adjuvant chemotherapy, endocrine treatment, and bisphosphonate treatment using the CellSearch system in the German trial monitoring CTCs in adjuvant therapy for breast cancer (the SUCCESS trial). Those authors detected more than 1 CTC in 10% of patients at the time of primary diagnosis before adjuvant therapy, which is much lower than the 31% observed in our study. The inclusion of T1, T2, and T3 disease compared with T1 and T2 disease alone in our study precludes strict comparison of the results of the 2 studies. Although CTCs did not correlate with any of the primary tumor characteristics, there was a significant association with positive axillary lymph node status.
In the current study, both CTCs in peripheral blood and DTCs in bone marrow occurred simultaneously in only 7.9% of patients with early stage breast cancer. The clinicopathologic characteristics of these patients did not differ significantly from the characteristics of patients who demonstrated either CTCs or DTCs alone or from those of patients who were entirely negative for minimal residual disease. The lack of a correlation between CTCs and DTCs and either CTCs or DTCs with standard prognostic and predictive markers in our study raises the possibility of independent modes of dissemination to the different homing sites. The median follow-up for our study is approximately 20 months; therefore, we did not attempt to assess survival data related to CTCs and DTCs in the current report. However, we plan to perform and report survival analyses once we have longer follow-up.
Our study was limited by the relatively small sample size of patients with early stage breast cancer who were investigated for minimal residual disease. The very low rate of HER2 positivity (5%) in this population particularly limited the power of statistical analyses that were performed to evaluate the correlation of this important prognostic and predictive marker with CTCs and/or DTCs. The results of our study suggesting independent modes of dissemination to blood, bone marrow, and lymph nodes require further validation in a larger cohort of patients. The detection of CTCs and DTCs in patients with early stage breast cancer who were negative for locoregional metastasis in our study indicates the potential utility of CTCs and DTCs as factors that may influence the selection of these patients for adjuvant chemotherapy. The clinicopathologic features of these patients were not particularly different from those of the other lymph node-positive patients in the study. In this regard, larger prospective studies of minimal residual disease in blood and bone marrow from patients with lymph node-negative, early stage breast cancer may provide valuable prognostic information. In addition, correlation of the presence of CTCs and DTCs in lymph node-negative, early stage breast cancer with other currently available molecular tests, such as Oncotype Dx (Genomic Health, Redwood City, Calif) and MammaPrint (Agendia, Amsterdam, the Netherlands), which provide prognostic information and facilitate the decision-making process regarding the use of adjuvant therapy, needs serious consideration.
In summary, we observed CTCs in 31% and DTCs in 27% of patients with early stage T1 and T2 breast cancer. The lack of correlation between minimal residual disease in blood and bone marrow with primary tumor characteristics in our study suggests that there are independent modes of dissemination to different homing sites. However, in view of the conflicting reports in this regard, larger, prospective, multi-institutional trials using standardized detection techniques are certainly warranted to establish unequivocally the association of different modes of tumor spread to homing sites and the role of CTCs and DTCs as prognostic indicators in this patient population. In addition, the possible effect of the presence of CTCs and DTCs in lymph node-negative, early stage breast cancer and their correlation with the results of genomic testing of the primary tumor may provide valuable insight into their definite role in the selection of patients for adjuvant therapy.
CONFLICT OF INTEREST DISCLOSURES
Supported by the Department of Defense Breast Cancer Research Program (Award DAMD 17-03-01-0669), Society of Surgical Oncology Clinical Investigator Award, and The M. D. Anderson Institute for Personalized Cancer Therapy (to A.L.). Also supported by a grant from the State of Texas Rare and Aggressive Breast Cancer Research program (to M.C.) and the Morgan Welch Inflammatory Breast Cancer Research program and clinic.