Prognostic Significance of Disseminated Tumor Cells as Detected by Quantitative Real-Time Reverse-Transcriptase Polymerase Chain Reaction in Patients with Breast Cancer

Prognostic Significance of Disseminated Tumor Cells as Detected by Quantitative Real-Time Reverse-Transcriptase Polymerase Chain Reaction in Patients with Breast Cancer

original contribution Prognostic Significance of Disseminated Tumor Cells as Detected by Quantitative Real-Time Reverse-Transcriptase Polymerase Chai...

136KB Sizes 2 Downloads 33 Views

original

contribution Prognostic Significance of Disseminated Tumor Cells as Detected by Quantitative Real-Time Reverse-Transcriptase Polymerase Chain Reaction in Patients with Breast Cancer Ina H. Benoy,1 Hilde Elst,1 Marita Philips,1 Hilde Wuyts,1 Peter Van Dam,1 Simon Scharpé,2 Eric Van Marck,1 Peter B. Vermeulen,1 Luc Y. Dirix1 Abstract Background: In this study we have validated the feasibility of detecting disseminated tumor cells (DTC) by real-time reverse-transcriptase polymerase chain reaction (RT-PCR) analysis. Bone marrow samples from a large cohort of patients with breast cancer were analyzed for the presence of DTC by immunocytochemistry (ICC) or a molecular-based method. Patients and Methods: Bone marrow samples were collected from 170 patients with breast cancer with stage I-IV disease before the initiation of any local or systemic treatment. Staining for cytokeratin (CK)–positive cells was performed with the Epimet® kit. Disseminated tumor cells were also quantified by measuring relative gene expression for CK19 and mammaglobin (MAM) using a quantitative RT-PCR detection method. The mean follow-up time was 30 months. Kaplan-Meier analysis was used for predicting overall survival. Results: Despite an excellent quantitative correlation and qualitative concordance between ICC and RT-PCR, survival analysis suggested an improved prognostic significance of DTC as detected by quantitative RT-PCR. Univariate survival analysis computed a relative risk of death of 2.87 for women with ICC-positive cells in the bone marrow, as compared with those without positive cells. The relative risk for women with RT-PCR–positive bone marrow was even higher: 3.5 (CK19) and 3.39 (MAM). In multivariate analysis, bone marrow CK19 was a stronger prognostic factor than bone marrow ICC. Conclusion: Reversetranscriptase polymerase chain reaction–detected DTC is shown to be prognostically significant in untreated patients with breast cancer. Furthermore, it seems to be a more sensitive method for detecting DTC in bone marrow samples when compared with ICC.

Clinical Breast Cancer, Vol. 7, No. 2, 146-152, 2006 Key words: Cytokeratin 19, Immunocytochemistry, Mammaglobin

Introduction In Europe and the United States, breast cancer remains an important public health problem. Because of increasing awareness and the implementation of screening methods, more breast tumors are being diagnosed at an early 1Translational Cancer Research Group Antwerp 2Medical Biochemistry, University of Antwerp Wilrijk, Belgium

Submitted: Apr 30, 2006; Revised: May 23, 2006; Accepted: Jun 1, 2006 Address for correspondence: Luc Y. Dirix, MD, Oncology Centre St-Augustinus, Oosterveldlaan 24, 2610 Antwerp (Wilrijk), Belgium Fax: 32-3-4433009; e-mail: [email protected]

stage. Even among these small tumors, however, there is a remarkable heterogeneity in outcome. The association of prognosis in patients with lymph node–negative breast cancer with specific gene expression profiles has highlighted the heterogeneity of breast cancer as it is grouped by TNM stage.1-5 Adjuvant treatment regimes need to become patient tailored.6 Whole genome expression profile of the primary tumor is one method to achieve this goal. The initiation of the MINDACT (Microarray in Node Negative Disease May Avoid Chemotherapy) trial attempts to achieve this goal.7 An alternative line of thinking is not to search for the hidden message in the primary tumor but rather to demonstrate the presence of disseminated tumor cells

Electronic forwarding or copying is a violation of US and International Copyright Laws. Authorization to photocopy items for internal or personal use, or the internal or personal use of specific clients, is granted by CIG Media Group, LP, ISSN #1526-8209, provided the appropriate fee is paid directly to Copyright Clearance Center, 222 Rosewood Drive, Danvers, MA 01923 USA 978-750-8400.

146 • Clinical Breast Cancer June 2006

(DTCs).6,8 Detection of minimal disseminated disease might be important in assessing this risk. The increasing use of immunohistochemistry and molecular biologic techniques has enabled pathologists to detect microscopic lesions down to the level of isolated tumor cells. Molecular diagnostics have, for the first time, been integrated in the revised TNM staging system for breast cancer in order to collect data on minimal disease in lymph nodes that could affect treatment policies in the future.9 The staging system, however, does not take into account the presence of DTCs at distant sites. Since the 1980s, studies on the detection of low-level disease in patients with operable breast cancer have been met with increasing interest. Bone marrow is the most frequently studied organ to be examined for the presence of DTCs. The prognostic relevance of disseminated epithelial cells in bone marrow is clearly demonstrated by a number of large studies.10-12 In pooled analyses on bone marrow micrometastasis in breast cancer, Braun et al showed that the presence of micrometastasis was a significant prognostic factor with respect to poor overall survival (OS), breast cancer–specific survival, disease-free survival, and distant disease-free survival during the 10-year observation period.12 Two distinct approaches are widely used to detect isolated tumor cells: immunocytochemistry (ICC) and molecular biologic–based methods. Immunocytochemistry is the standard method for tumor cell detection, but in recent years, we have witnessed a rapid expansion in the application of molecular polymerase chain reaction (PCR)–based methods to detect isolated tumor cells in hematologic fluids. The major limitation of the use of reverse-transcriptase PCR (RT-PCR) in detecting micrometastatic disease in bone marrow of patients with common solid tumors has been the lack of gene markers expressed exclusively and uniformly by the tumor cells. In this study, we attempted to quantify transcripts of a marker considered relatively specific for epithelial cells, cytokeratin (CK) 19 (CK19), and of a breast cell–related marker, mammaglobin (MAM), in the bone marrow of patients with breast cancer using quantitative realtime RT-PCR according to the Taqman® method. In a pilot study, we demonstrated the quantitative relationship between ICC and RT-PCR for the detection of DTCs in the bone marrow of patients with metastatic breast cancer.13 The aim of this study was to confirm this relationship in a large cohort of patients with breast cancer and to evaluate and compare the prognostic value of both methods. Therefore, this study is conducted on the ICC- and RT-PCR– based detection of DTCs in bone marrow samples from 170 patients with stage I-IV breast cancer before the initiation of any systemic treatment.

Patients and Methods Sample Collection Bone marrow samples were collected from 170 unselected patients with breast cancer (stage I, II, and III before initial surgery or therapy, stage IV at diagnosis) and from 17

patients with a nonmalignant breast lesion or hematologic diseases (control patients). Bone marrow (12-20 mL) was aspirated from the posterior iliac crest under local or general anaesthesia into syringes containing 5000 IU heparin as anticoagulant. Mononuclear cells (MNCs) were isolated by density-gradient centrifugation through Ficoll-Paque and washed twice with phosphate-buffering saline. The samples were then divided into 2 aliquots, 1 for each method. After centrifugation, the cell pellets were resuspended in a guanidine isothiocyanate–containing buffer or in phosphatebuffering saline for RT-PCR and ICC, respectively. The corresponding bone marrow volume for each method was documented to enable the calculation of target messenger RNA (mRNA) concentration later. The study protocol was approved by the ethical committees of the Faculty of Medicine, University of Antwerp, and of the General Hospital Sint-Augustinus. Patients gave informed consent to the study. Patient characteristics were prospectively recorded in a database. All analyses were conducted blind of the patient’s clinicopathologic status.

Immunocytochemistry The bone marrow MNC suspension was counted and cytocentrifuged onto glass slides at a concentration of 5 × 105 cells per spot. The cytospin preparations were air-dried overnight and stored at –80°C. Immunostaining to detect CK-positive cells was done with the Epimet® kit on an automatic immunostainer. This kit uses the monoclonal antibody A45-B/B3, a pancytokeratin marker. A total of 2 million cells per patient were screened microscopically. Cells were identified as disseminated epithelial cells according to the European ISHAGE (International Society for Hematology and Graft Engineering) Working Group for Standardization of Tumor Cell Detection.14 Results were expressed as the number of positive cells per million MNCs.

Real-Time Reverse-Transcriptase Polymerase Chain Reaction From the bone marrow MNCs, total RNA was extracted with the RNeasy® midi kit according to the protocol. The exact elution volume (range, 150-600 μL) was documented to enable the calculation of target mRNA concentration later. The concentration and purity of the total RNA was determined by measuring the absorbance at 260 nm (A260) and 280 nm (A280) in a spectrophotometer. The RNA integrity was tested on the Agilent 2100 Bioanalyzer. The profile of the electropherogram, 18S/28S ratio and RNA integrity number was used for integrity interpretation. Only samples with lack of degradation on the electropherogram and a 18S/28S ratio > 1.75 and a RNA integrity number > 7 were used for further analysis. For the generation of first-strand complementary DNA (cDNA), 2 μg of total RNA was reverse-transcribed in a final volume of 100 μL with the High-Capacity cDNA Archive Kit. Cytokeratin 19, MAM, `-actin, and glyceraldehyde-3phosphate dehydrogenase mRNA expression was measured

Clinical Breast Cancer June 2006 • 147

Prognostic Significance of Disseminated Epithelial Cells Table 1 Primer and Probe Information Cytokeratin 19

Characteristic

GenBank accession number Forward primer

NM_002276 CCCGCGACTACAGCCACTA

Reverse primer

CTCATGCGCAGAGCCTGTT

Probe

FAM-ACCATTGAGAACTCCA GGATTGTCCTGCA-TAMRA

Amplicon size

163 Base pairs

Mammaglobin GenBank accession number

AF_015224 ATGAAGTTGCTGATGGTCCTCAT

Reverse primer

GTCTTAGACACTTGTGGATTGATTGTCT

Amplicon size

Total

FAM-CGGCCCTCTCCCAGCACTGC-TAMRA 119 Base pairs

as previously described.13,15 All PCR reactions were performed on the ABI Prism® 7700 Sequence Detection System using the fluorescent Taqman® method. Primer and probe sequences are presented in Table 1. Ten microliters of the reverse-transcriptase volume was used for each PCR reaction in a total volume of 50 μL. The target mRNA quantities were analyzed in triplicate, and mean Ct levels were used for further analyses. As described by Livak and Schmittgen, results per PCR reaction are calculated as relative gene expression (RGE) using the bb CT method.16 The calibrator was produced from blood of a healthy volunteer spiked with 5 MDA-MB361 cells per 106 blood cells. The calibrator was given an RGE value of 100.

Statistics Data were analyzed with the data package SPSS for Windows. Survival was calculated from the day of bone marrow aspiration until the day of death or last follow-up. Survival curves were calculated using the Kaplan-Meier estimates. Statistical significance between groups was assessed using the log-rank test. Univariate and multivariate analysis was carried out to assess the relative influence of prognostic factors on OS using the Cox proportional hazards model in a forward stepwise procedure. A P value < 0.05 was considered significant.

Number of Patients

ICCPositive Rate (%)

CK19Positive Rate (%)

MAMPositive Rate (%)

170

28

29

21

Age (Years) Range

27-88

Median

61

Menopausal Status Premenopausal

41

29

21

16

Postmenopausal

123

24

31

24

6

33

33

33

I

36

21

19

3

II

44

21

23

28

III

43

13

24

12

IV

47

55

45

40

T1

53

22

29

12

T2

44

30

20

25

T3

21

21

30

20

T4

46

26

32

27

TX

6

100

60

40

pN0

58

21

21

15

pN+

82

26

33

24

pNX*

30

47

37

41

Ductal

143

25

28

20

Lobular

24

43

32

23

Other

3

33

33

67

1

39

16

29

9

2

74

30

23

26

3

57

31

37

27

Positive

119

28

25

21

Negative

51

27

38

23

Positive

94

26

23

16

Negative

76

29

35

28

Unknown

Forward primer Probe

Table 2 Clinicopathologic Characteristics of Patients

Disease Stage

Tumor Status

Lymph Node Status

Histology

Histologic Grade

ER Status

PgR Status

Results Patients Clinicopathologic characteristics of the patients are summarized in Table 2.

HER2 Expression (FISH)

Immunocytochemistry of Bone Marrow Samples

Negative

138

26

29

19

For 11 patients with breast cancer, the sample volume of the bone marrow aspirate was insufficient for ICC and RT-PCR analyses, and RT-PCR analysis was preferred. Also for 8 patients, the cytospin quality was inappropriate for conclusions about the presence of DTCs. Excluding these patients,

Positive

24

30

28

29

Unknown

8

25

25

25

148 • Clinical Breast Cancer June 2006

*pNx: patients who received previous neoadjuvant chemotherapy are considered as being Nx for the purpose of this study. Abbreviations: ER = estrogen receptor; FISH = fluorescence in situ hybridization; PgR = progesterone receptor

Ina H. Benoy et al Figure 1 Correlation Between Number of Mononuclear Cells and Total RNA in Bone Marrow Samples

Figure 2 Equation for Calculating CK19 and MAM Gene Expression per Polymerase Chain Reaction 1 nRGE = RGE × CRNA × VRNA × Vext CcDNA × VPCR Reaction

250

Total RNA (μg)

200 Abbreviations: CcDNA = concentration cDNA, typically 2 μg/100μL; CRNA = concentration total RNA extracted per sample; nRGE = normalized RGE expressed as relative target concentration per mL peripheral blood or bone marrow (relative gene expression per mL sample); RGE = relative gene expression of target determined by sequence detector per PCR reaction; Vext = volume of bone marrow extracted, typically 9 mL; VPCR = volume of cDNA solution used for PCR amplification, typically 10 μL; VRNA = elution volume of total RNA obtained after extraction, typically 300 μL per Qiagen RNeasy® Midi Extraction

150 100 50 0 0

2  107 4  107 6  107 8  107 10  108 12  107 Number of Mononuclear Cells

151 bone marrow samples were used for further ICC analyses. Cytokeratin-positive cells were not detected in any of the 17 bone marrow aspirates from negative control patients The median number of positive cells in bone marrow aspirates from patients with breast cancer was 0 cells per 106 MNCs (range, 0-5000 cells; mean, 55 cells). In 28% (42 of 151) of patients with breast cancer, * 1 cell were immunostained.

Quantification of Messenger RNA Results Bone marrow samples of 4 negative control patients were excluded for analysis because of insufficient sample collection. RNA from good quality could be isolated in 158 of 170 aspirates from patients with breast cancer. These samples were used for further analysis. The number of disseminated epithelial cells in the bone marrow of patients with breast cancer is low (if they are present at all) and requires a sensitive method for detection. Accordingly, CK19 and MAM mRNA molecules can be rare compared with MNC-derived mRNA. For the quantitative detection of circulating tumor cells in peripheral blood samples and for the quantitative detection of disseminated epithelial cells in bone marrow samples, we are only interested in the number of tumor cells per mL sample, regardless of the amount of MNCs. Using the method described earlier, gene expression is calculated in 10 μL cDNA derived from 0.2 μg total RNA per sample. This total RNA contains RNA derived from possible tumor cells, with the majority derived from the existing MNCs. The number of MNCs in each sample differs from patient to patient. This implies that the amount of total RNA extracted from the same sample’s volume differs from patient to patient. When analyzing the 171 (158 patients plus 13 negative controls) bone marrow samples, a strong correlation was found between the number of MNCs and amount of total RNA recovered: Spearman l = 0.884; P < 0.0001. Figure 1 represents the correlation observed between the number of MNCs and the amount

of total RNA extracted (μg). Because of this relationship, results for CK19 and MAM gene expression per PCR reaction in bone marrow samples were recalculated using the equation in Figure 2.

CK19 and MAM Messenger RNA Expression in Bone Marrow Mammaglobin expression was not measurable by RT-PCR in any of the 13 negative control bone marrow samples. On the other hand, CK19 mRNA was quantified in all control samples, with a median normalized RGE (nRGE) of 9.94 (range, 4-36 nRGE). With the 95th percentile from the CK19 nRGE (26.3) of the negative control group as a cutoff, 45 of 158 (28.5%) bone marrow aspirates of patients with breast cancer had an increased CK19 expression. Mammoglobin expression was measurable in 34 of 158 samples (21.5%).

Correlation Between Reverse-Transcriptase Polymerase Chain Reaction and Immunocytochemistry in Bone Marrow Samples In 140 of 170 patients with breast cancer, results for detection of DTCs in bone marrow were obtained by 2 techniques: ICC and RT-PCR. Quantitatively, a strong correlation was found between the number of ICC-detected CK-positive cells and CK19 or MAM nRGE in the bone marrow (CK19: Spearman l = 0.45, P < 0.0001, 95% CI, 0.31-0.59; MAM: Spearman l = 0.33, P < 0.0001, 95% CI, 0.17-0.47). In 45% (19 of 42) of the patient samples with bone marrow CK19 nRGE above the cutoff, no CK-positive cells were detected by ICC. On the other hand, 13 patients had a positive ICC result but a bone marrow nRGE of CK19 below the cutoff (Table 3). Overall, there is a concordance of 77% (108 of 140 samples) between ICC and CK19 RTPCR bone marrow DTC positivity. Statistically, there is an association of 44% between the row and column variables (Cramer’s V = 0.44; P < 0.0001) and a fair g value of 0.43. Also, according to the McNemar test, there is an association between ICC and CK19 RT-PCR (P = 0.377). For MAM expression, a concordance of 71% (100 of 140 samples) was found with ICC (Table 3). The g value for MAM expression was marginal (g = 0.22), and there was no difference in positivity according to the McNemar test (P = 0.636).

Clinical Breast Cancer June 2006 • 149

Prognostic Significance of Disseminated Epithelial Cells Figure 3 Overall Survival in Patients According to the Presence of DTCs in Bone Marrow A

Bone Marrow Number of Patients Number of Patients Status ICC Negative ICC Positive

100

CK19 Status*

Cumulative Survival (%)

BM ICC– 80 60

Cumulative Survival (%)

B

Cumulative Survival (%)

C

10

20

30 Months

40

50

60

13

Positive

19

23

Negative

86

13

Positive

18

14

Table 4 Univariate Analysis for Overall Survival in Patients with Breast Cancer* All Patients Bone Marrow Status

100

Patients Who Died/ Total Number

Relative Risk of Death (95% CI)

P Value

Positive

11/40

2.87 (1.28-6.44)

0.01

Negative

13/107

1

Positive

14/45

3.5 (1.62-7.6)

Negative

12/110

1

Positive

10/33

3.39 (1.53-7.5)

Negative

16/122

1

Positive

10/24

2.32 (1.54-3.5)

Negative

14/123

1

BM CK19–

80

ICC

60 BM CK19+ 40 20

CK19 0.001

MAM

10

20

30 Months

40

50

60

BM MAM–

80

BM MAM+

60 40 20

10

20

30 Months

40

50

60

Abbreviation: BM = bone marrow

Survival Analysis Follow-up data were available for 166 of 170 patients with breast cancer. The median follow-up was 30 months, and 28 deaths have occurred in these 170 patients. In the bone marrow ICC-positive group, 11 of 40 patients died, compared with 13 of

150 • Clinical Breast Cancer June 2006

0.003

ICC, CK19, and MAM

100

0

85

*Number of patients with positive and negative CK19 nRGE and ICC in the bone marrow. †Number of patients with positive and negative MAM nRGE and ICC in the bone marrow.

20

0

Negative MAM Status†

BM ICC+

40

0

Table 3 Comparison Between ICC and RT-PCR for the Detection of Disseminated Epithelial Cells in Bone Marrow*

< 0.0001

*Results of univariate analysis (Cox regression) for overall survival in patients with breast cancer.

107 in the bone marrow ICC-negative group. When analyzing survival data according to the MAM or CK19 mRNA expression in bone marrow, 14 of 45 CK19-positive patients and 10 of 33 bone marrow MAM-positive patients died, compared with 12 of 110 bone marrow CK19-negative patients and 16 of 122 in the bone marrow MAM-negative patient group (Table 4). KaplanMeier survival analyses demonstrate a markedly reduced OS among the bone marrow–positive patients (OS: log-rank test, P = 0.0073 [ICC]; P = 0.0007 [CK19] and P = 0.0014 [MAM]; Figure 3). Among women with ICC-positive cells in the bone marrow compared with those without positive cells, the relative risk of death was 2.87. The relative risk for patients with increased CK19 or MAM mRNA expression in the bone marrow was 3.5 (CK19) and 3.39 (MAM) compared with those with normal CK19 or MAM mRNA expression (Table 4). The statistic of risk ratio for patients who are positive for all 3 markers (ICC, CK19, and MAM) was 2.32.

Ina H. Benoy et al Multivariate analysis using Cox regression model, including disease stage, tumor size, lymph node status, menopausal status, hormone receptor status, histologic subtype, histologic grade, HER2 expression, and bone marrow status (both methods), revealed bone marrow CK19 positivity, disease stage, and histologic grade to be the strongest independent prognostic factors for OS (Table 5). Analyzing the data separately for the M0 subgroup (n = 123), visual discrimination was better for RT-PCR analysis. Analyses of the patients in the M0 subgroup revealed a trend toward significance in OS according to the bone marrow CK19 or MAM status but not according to the ICC status (OS: log-rank test, P = 0.125 [CK19] and P = 0.077 [MAM]; P = 0.428 [ICC]; data not shown).

Discussion In this study, we have validated the feasibility of detecting disseminated epithelial cells by real-time RT-PCR analysis. Bone marrow samples from a large cohort of patients with breast cancer (N = 170) were analyzed for the presence of DTCs by the golden standard method ICC or by a molecularbased method. For the first time, a large prospective analysis of clinical relevance of the detection of DTCs by real-time RT-PCR was reported. We have confirmed our previous pilot study by demonstrating the quantitative correlation between the number of ICC-detected CK-positive cells and CK19 or MAM mRNA expression using the bone marrow of a large group of untreated patients with breast cancer at various stages of the disease.13 Quantitatively, a concordance of 77% (CK19-ICC) and 71% (MAM-ICC) was observed between the 2 methods. We found some samples that were positive by RT-PCR but negative by ICC. This could have been a result of the superior sensitivity of PCR. By PCR analysis, each target will be amplified a million times before detection. The presence of multiple copies of mRNA can provide a more reliable target. In some samples, RT-PCR failed to detect ICC-positive cells. Because the ICC assay uses a pancytokeratin marker (A45-B/B3), it might be that these cells express CKs other than CK19. Or as suggested by Schoenfeld et al, these ICC-detected cells were not viable, or they were dormant with low metabolic activity as defined by their inability to synthesize CK19 or MAM mRNA.17 Other possibilities for the discordant results observed between ICC and RT-PCR detection are the false-positive staining reactions or the false-positive mRNA expression. Hematopoietic cells can occasionally be stained positive by the ICC technique used.14 Morphologic evaluation improves the specificity of ICC detection, but optimal separation of isolated tumor cells from hematopoietic cells is not always clear.18 By taking the 95th percentile of target mRNA expression in negative control bone marrow samples as a cutoff value for the RT-PCR analysis, by definition, 5% of patients’ RT-PCR results are considered false-positive. Also the phenomenon of illegitimated expression can induce false-positive results.19,20

Table 5 Multivariate Analysis for Overall Survival in Patients with Breast Cancer Bone Marrow

Risk Ratio

P Value

Bone Marrow CK19 mRNA Expression

4.74

0.004

1.6-13.61

Disease Stage

3.97

0.001

1.79-8.84

Histologic Grade

3.26

0.007

1.39-7.69

95% CI

Results of multivariate analysis (Cox regression) for overall survival in 148 patients with breast cancer (only variables significant on multivariate analysis are shown).

Despite the excellent quantitative correlation and qualitative concordance between ICC and RT-PCR, survival analyses suggest a superior prognostic significance of DTCs as detected by quantitative RT-PCR. Univariate survival analysis computed a relative risk of death of 2.87 for women with ICC-positive cells in the bone marrow compared with those without positive cells. The relative risk for women with RT-PCR–positive bone marrow was even higher: 3.5 (CK19) and 3.39 (MAM). In a multivariate analysis, bone marrow CK19 was a stronger prognostic factor than bone marrow ICC. The prognostic relevance of disseminated epithelial cells in bone marrow confirms previous studies.10-12

Conclusion Even with this low number of events, the prognostic significance of RT-PCR–detected DTCs trended toward significance in OS in the cohort of patients with disease stage M0. Reverse-transcriptase PCR–detected DTCs are shown to be prognostically significant in untreated patients with breast cancer. Furthermore, it seems to be a more sensitive method for detecting DTCs in bone marrow samples when compared with ICC.

References 1. Perou CM, Sorlie T, Eisen MB, et al. Molecular portraits of human breast tumours. Nature 2000; 406:747-752. 2. Van’t Veer L, Dai H, Van de Vijver MJ, et al. Gene expression profiling predicts clinical outcome of breast cancer. Nature 2002; 415:530-536. 3. Van de Vijver MJ, He YD, Van’t Veer L, et al. A gene-expression signature as a predictor of survival in breast cancer. N Engl J Med 2002; 347:1999-2009. 4. Huang E, Cheng SH, Dressman H, et al. Gene expression predictors of breast cancer outcomes. Lancet 2003; 361:1590-1596. 5. Foekens JA, Atkins D, Zhang Y, et al. Multicenter validation of a gene expression-based prognostic signature in lymph node-negative primary breast cancer. J Clin Oncol 2006; 24:1665-1671. 6. Dirix L, van Dam P, Vermeulen P. Genomics and circulating tumor cells: promising tools for choosing and monitoring adjuvant therapy in patients with early breast cancer? Curr Opin Oncol 2005; 17:551-558. 7. Loi S, Sotiriou C, Buyse M, et al. Molecular forecasting of breast cancer: time to move forward with clinical testing. J Clin Oncol 2006; 24:721-722. 8. Martin KJ, Graner E, Li Y, et al. High-sensitivity array analysis of gene expression for the early detection of disseminated breast tumor cells in peripheral blood. Proc Natl Acad Sci U S A 2001; 98:2646-2651. 9. Singletary SE, Allred C, Ashley P et al. Revision of the American Joint Committee on Cancer staging system for breast cancer. J Clin Oncol 2002; 20:3628-3636. 10. Braun S, Pantel K, Muller P, et al. Cytokeratin-positive cells in the bone marrow and survival of patients with stage I, II, or III breast cancer. N Engl J Med 2000; 342:525-533.

Clinical Breast Cancer June 2006 • 151

Prognostic Significance of Disseminated Epithelial Cells 11. Weinschenker P, Soares HP, Clark O, et al. Immunocytochemical detection of epithelial cells in the bone marrow of primary breast cancer patients: a meta-analysis. Breast Cancer Res Treat 2004; 87:215-224. 12. Braun S, Vogl FD, Naume B, et al. A pooled analysis of bone marrow micrometastasis in breast cancer. N Engl J Med 2005; 353:793-802. 13. Benoy IH, Elst H, Van der Auwera I, et al. Real-time RT-PCR correlates with immunocytochemistry for the detection of disseminated epithelial cells in bone marrow aspirates of patients with breast cancer. Br J Cancer 2004; 91:1813-1820. 14. Borgen E, Naume B, Nesland J, et al. Standardization of the immunocytochemical detection of cancer cells in BM and blood: establishment of objective criteria for the evaluation of immunostained cells. Cytotherapy 1999; 1:377-388. 15. Benoy IH, Elst H, Philips M, et al. Real-time RT-PCR detection of disseminated tumour cells in bone marrow has superior prognostic significance in comparison with circulating tumour cells in patients with breast cancer. Br J Cancer 2006; 94:672-680.

152 • Clinical Breast Cancer June 2006

16. Livak KJ, Schmittgen TD. Analysis of relative gene expression data using real-time quantitative PCR and the 2(-delta delta C(T)) method. Methods 2001; 25:402-408. 17. Schoenfeld A, Kruger KH, Gomm J, et al. The detection of micrometastases in the peripheral blood and bone marrow of patients with breast cancer using immunohistochemistry and reverse transcriptase polymerase chain reaction for keratin 19. Eur J Cancer 1997; 33:854-861. 18. Naume B, Wiedswang G, Borgen E, et al. The prognostic value of isolated tumor cells in bone marrow in breast cancer patients: evaluation of morphological categories and the number of clinically significant cells. Clin Cancer Res 2004; 10:3091-3097. 19. Chelly J, Concordet JP, Kaplan JC, et al. Illegitimate transcription: transcription of any gene in any cell type. Proc Natl Acad Sci U S A 1989; 86:2617-2621. 20. Ko Y, Grunewald E, Totzke G, et al. High percentage of false-positive results of cytokeratin 19 RT-PCR in blood: a model for the analysis of illegitimate gene expression. Oncology 2000; 59:81-88.