Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
VOLUME 24 䡠 NUMBER 22 䡠 AUGUST 1 2006 JOURNAL OF CLINICAL ONCOLOGY O R I G I N A L R E P O R T Breast Cancer–Specific mRNA Transcripts Presence in Peripheral Blood After Adjuvant Chemotherapy Predicts Poor Survival Among High-Risk Breast Cancer Patients Treated With High-Dose Chemotherapy With Peripheral Blood Stem Cell Support Miguel Quintela-Fandino, Joaquín Martínez López, Ricardo Hitt, Soledad Gamarra, Antonio Jimeno, Rosa Ayala, Javier Hornedo, Cecilia Guzman, Florinda Gilsanz, and Hernán Cortés-Funes From the Medical Oncology Department; Molecular Biology Division, Hematology Department, University Hospital 12 de Octubre; Roche Pharma, Madrid, Spain; and The Sidney Kimmel Comprehensive Cancer Center, John Hopkins University, Baltimore, MD. Submitted September 3, 2005; accepted June 6, 2006. Supported by Roche Pharma, Madrid, Spain. Presented in part at the 41st Annual Meeting of the American Society of Clinical Oncology, Orlando, FL, May 13-17, 2005. Authors’ disclosures of potential conflicts of interest and author contributions are found at the end of this article. Address reprint requests to Miguel Quintela-Fandino, MD, PhD, Medical Oncology Department, University Hospital 12 de Octubre, Avenida de Córdoba Km 5.4, 28041 Madrid, Spain; e-mail: [email protected]. © 2006 by American Society of Clinical Oncology 0732-183X/06/2422-3611/$20.00 DOI: 10.1200/JCO.2005.04.0576 A B S T R A C T Purpose To study the prognostic significance of the presence of breast cancer–specific mRNA transcripts in peripheral blood (PB), defined by serial analysis of gene expression, in high-risk breast cancer (HRBC) patients undergoing high-dose chemotherapy after receiving adjuvant chemotherapy. Methods From 1994 to 2000, 84 HRBC patients (median age, 44 years; ⬎ 10 nodes; 74%) received adjuvant chemotherapy (fluorouracil, epirubicin, and cyclophosphamide for six cycles [83%] or doxorubicin and cyclophosphamide followed by paclitaxel) before undergoing one course of cyclophosphamide plus thiotepa plus carboplatin (STAMP V). Radiotherapy or hormone therapy was administered whenever indicated. Aliquots of apheresis-mononuclear blood cells were frozen from each patient. mRNA was isolated using an automatic nucleic acid extractor based on the magnetic beads technology; reverse transcription was performed using random hexamers. Cytokeratin 19, HER-2, P1B, PS2, and EGP2 transcripts were quantified to B-glucuronidase by real-time polymerase chain reaction (RT-PCR) using a linear DNA probe marked with a quencher and reporter fluorophores used in RT-PCR. Presence of PB micrometastases, estrogen receptor and progesterone receptor status, tumor size, age, tumor grade, number of nodes affected, and treatment with paclitaxel were included in the statistical analysis. Results Median follow-up was 68.3 months (range, 6 months to 103 months). Forty-seven relapses (56%) and 35 deaths (41.7%) were registered. Both tumor size and presence of micrometastases reached statistical significance according to the Cox multivariate model. Relapse hazard ratio (HR) for those patients with PB micrometastases was 269% (P ⫽ .006); death HR, 300% (P ⫽ .011). Time relapse was 53 months longer for patients without micrometastases: 31.3 v 84.2 months (P ⫽ .021). Conclusion PB micrometastases presence after adjuvant chemotherapy predicts both relapse and death more powerful than classical factors in HRBC patients undergoing high-dose chemotherapy. Micrometastases search using a gene panel appears to be a more accurate procedure than classical approaches involving only one or two genes. J Clin Oncol 24:3611-3618. © 2006 by American Society of Clinical Oncology INTRODUCTION Among localized breast cancer patients, clinically occult micrometastatic disease (MD) leads to relapse. HER-2 positivity is linked to worse prognosis.1-4 Positive HER-2 tumor cell selection takes place during tumor progression.5-7 It could be hypothesized that patients harboring HER-2–positive MD after adjuvant chemotherapy are at high risk of relapse. The conventional approach to screen for MD in breast cancer involved the detection of a protein present in breast tumor tissue and absent in hematopooietic tissue. Many studies have analyzed MD in the adjuvant setting with variable results.8-26 The following issues may explain such variation and might be taken into account for the study design: sample collection timing, in order to avoid detection of cells nonspecifically released during the surgical trauma27,28; tumor cell dormancy29; the decrease of 3611 Downloaded from ascopubs.org by 78.47.27.170 on November 17, 2016 from 078.047.027.170 Copyright © 2016 American Society of Clinical Oncology. All rights reserved. Quintela-Fandino et al MD contamination during adjuvant chemotherapy30; MD deposits existence in other organs that bone marrow; higher sensitivity and specificity are obtained with polymerase chain reaction (PCR) techniques rather than with immunohistochemistry procedures; and cytokeratin 19 (K19) was the most widely used protein for breast MD detection. K19 is present in breast tissue/cancer, but is exchanged for vimentin and experiences downregulation as tumor progresses, being mainly expressed in breast tumor cells unable to produce metastases.31-35 Breast MD could be detected by searching for the mRNA expression of a gene that expresses in every breast cancer cell, but at the same time it is absent in hematopooietic tissue. Such a gene does not exist. However, Bosma et al36 designed a highly accurate gene panel to detect circulating tumor cells, increasing specificity/sensitivity over singlegene assays and sorting out the main concern of background signal in mRNA-based assays. Two genes (P1B/PS2) expressed in breast cancer and almost absent in other tissues were found by serial analysis of gene expression. K19 and EGP237 were subsequently added to the panel.37 We hypothesize that the former issues might be overcome with the following design, and relapse risk of breast cancer patients can be assessed by indirect detection of clinically occult micrometastases located in any parenchima that are releasing tumor cells to the peripheral blood. We have performed a retrospective study to assess the prognostic role of MD, searching for P1B, PS2, EGP2, K19, and HER-2 (HER-2 added to the panel as it is not constitutively expressed in hematopooietic tissue) expression in peripheral blood of high-risk breast cancer patients undergoing high-dose chemotherapy (HDCT). We used real-time (RT) polymerase chain reaction (PCR) assays in samples collected during blood apheresis (several months after surgery and after conventional-dose adjuvant chemotherapy). The aims of this study were to assess the prognostic value of breast cancer– specific mRNA transcripts in peripheral blood, and determine if MD study deserves prospective validation; to test whether the study design resolves the inherent difficulties of assessing the prognostic implications of MD; and to study the HER-2 status of MD. METHODS Patients and Study Design This retrospective study was initiated in June 2003. Between 1993 and 1999 patients undergoing HDCT for breast cancer at two hospitals located in Madrid, Spain were screened. Inclusion criteria were: histopathologically proven, locally advanced breast cancer; the absence of distant metastases in chest abdominal computed tomography scan and bone scan both before adjuvant or neoadjuvant treatment and before HDCT; negative histopathologic examination for tumor cells in bone marrow biopsy before HDCT; previous adjuvant/neoadjuvant treatment with a standard anthracycline/taxane-based regimen; followed by adjuvant cyclophosphamide plus thiotepa plus carboplatin (STAMP V) course; and availability of frozen mononuclear cells obtained during apheresis and consequently collected and stored (⫺80°C; 10% dimethyl sulfoxide) for research purposes. Patients were required to give written informed consent before inclusion in the study. The study protocol was approved by the institutional review board at each study center. Molecular Biology Procedures Assays were conducted using 2-mL aliquots of mononuclear cells obtained during apheresis (one to six per patient; median three; total samples 226). Total mRNA was isolated and RNA integrity was tested by Table 1. Patient Characteristics Value Characteristic Age, years Median Range Sex Male Female Hormone status Premenopausal Menopausal Tumor size (surgical specimen) Tx T1 T2 T3 T4 Not available No. of positive nodes Median Range 1-3ⴱ 4-9 ⬎ 10 Primary metastatic disease Previous breast cancer history Histopathologic subtype Ductal Lobular Other Histologic grade I II III Hormone receptor status ER⫹/PR⫹ ER⫹/PR⫺ ER⫺/PR⫹ ER⫺/PR⫺ HER-2/NEU Negative (0 or 1⫹) Positive (2⫹ or 3⫹) Not available Surgery† Mastectomy ⫹ ALN Lumpectomy ⫹ ALN Adjuvant chemotherapy FEC AC-paclitaxel Adjuvant STAMP V Toxic deaths Adjuvant radiotherapy Adjuvant tamoxifen Time, months Median Range‡ No. % 44.28 24-69 0 100 72 12 85.5 14.5 3 7 32 24 15 3 3.6 8.5 38.3 28.9 18.8 3.6 13.37 2-38 4 18 62 0 0 4.7 21.4 73.8 0 0 72 7 5 85.7 8.3 6 1 32 51 1.1 38.1 60.8 43 9 1 31 51.19 10.71 1.19 36.90 23 24 37 27 28 44 59 25 70.3 29.8 70 14 84 1 84 50 83 17 100 1.2 100 59.5 48 1-60 Abbreviations: ER, estrogen receptor; PR, progesterone receptor; ALN, axillary lymph nodes; FEC, fluorouracil, epirubicin, and cyclophosphamide; AC, doxorubicin and cyclophosphamide; STAMP V, cyclophosphamide, thiotepa, and carboplatin. ⴱ These four patients had received neoadjuvant chemotherapy. †Surgery performed after neoadjuvant treatment in T4 patients. ‡Thirty-one patients stopped tamoxifen at disease progression and one patient stopped due to intolerance. 3612 JOURNAL OF CLINICAL ONCOLOGY Downloaded from ascopubs.org by 78.47.27.170 on November 17, 2016 from 078.047.027.170 Copyright © 2016 American Society of Clinical Oncology. All rights reserved. Micrometastases in Blood Predicts Poor Survival Table 2. Median RNA Yield Among Patients and Negative Controls Ct GUS Efficiency Patients Healthy donors with neutrophils Healthy donors without neutrophils 1.99 1.90 1.94 Range Positive Samples Negative Samplesⴱ 27.86 27.32 22.83-38.44 25.54-30.45 72 13 12 2 27.25 25.52-30.68 14 1 Median Abbreviations: Ct, threshold cycle; GUS, b-glucuronidase. ⴱ A sample is considered that has not yielded enough RNA when GUS Ct is ⬎ 29. amplification of the b-glucuronidase (GUS) housekeeping gene. c-DNA was synthesized using random-hexamer oligonucleotides. P1B, PS2, EGP2, HER-2, and K19 were amplified and quantified for each patient by RT-PCR with TaqMan probes (Applied Biosystems, Foster City, CA) in a LightCycler Instrument using the LC Fast Start DNA Master Hybridization Probe kit (Roche Diagnostics). RT-PCR efficiency was calculated using the LightCycler Software version 3.5 (Roche Applied Science, Mannheim, Germany) according to the equation: E ⫽ 10(⫺1/S), where S is the slope of the cumulative fluorescence during the exponential phase. Quantification was calculated relative to GUS38 using the equation: Ratio ⫽ (Etarget)(Ctref – Cttarget), where target represents each transcript and ref, GUS.39 Ct is the threshold cycle (PCR cycle at which a significant increase in target-specific fluorescence is detected due to exponential accumulation of PCR products, expressed in arbitrary units). Cell lines HCC1143 and MCF-7 were used as positive controls. As apheresis samples lack neutrophils, two types of negative controls were used: peripheral blood from 15 neutrophil-containing and 15 neutrophil-depleted healthy donors, respectively. For each patient, all available aliquots were pooled into a single sample (number of samples ⫽ 84). Those with a GUS-Ct above cycle 29 (n ⫽ 12) were discarded, as the low mRNA amount could yield false-negatives. All the procedures were repeated twice. The technicians neither had access to follow-up data nor took part in the analysis. More detailed information regarding treatment and patient selection, specific molecular biology procedures, and verification of classic prognostic factors are available at the Appendix. Statistical Analysis The follow-up period started the day surgery was performed. The main variable analyzed was presence of micrometastatic (MM) activity. Breast cancer cells do not harbor the same genetic expression profile, as opposed to hematologic malignant cells, and consequently one target may be expressed in one tumor cell and absent in another one. In order to increase the chance of finding circulating tumor cells, a high number of specific mRNA targets for micrometastatic detection were utilized. MM was categorized as positive/yes if either target was detected in peripheral blood regardless of the value, and as negative/no if no target was detected. K19 was excluded from MM as K19 was present at similar ratios in samples from patients and negative controls. The Kaplan-Meier method and Cox hazard regression model were used for statistical analysis. The clinical end points under analysis were time to relapse or time to death (Kaplan-Meier) and relapse/death (Cox). Age, tumor grade, tumor size, treatment with paclitaxel, estrogen receptor (ER) and progesterone receptor (PR) status, and number of nodes were included in the Cox analysis to test the independent prognostic impact of MM. HercepTest for Immunoenzymatic Staining (Dako Corp, Carpinteria, CA) was not routinely performed at our institution until 1998. These data were not available from all patients and therefore HER-2 in the primary tumor was removed from the multivariate analysis. The overall significance and goodness of fit of each model was calculated with the Rao test and Cox-Snell R2. Both the log-linear and the multiplicative conditions for the application of the Cox model were tested. SPSS version 12.0 for Windows (SPSS Inc, Chicago, IL) was used for the statistical analysis. All statistical tests were two sided. RESULTS Patients Patient characteristics are depicted in Table 1 (n ⫽ 84). After a median follow-up of 68.3 months (range, 6 to 103; follow-up period 1993 to 2004), 47 patients (56.0%; 95% CI, 45.4% to 66.6%) have relapsed, with a median TTR of 66 months. Median overall survival has not been reached. At the end of follow-up, 36 patients (42.8%; 95%CI, 37.3% to 48.3%) were alive and disease-free, 13 patients Table 3. Efficiency and Relative Values for Each Gene Among Patients and Negative Controls Blood From Healthy Donors (containing neutrophils) Patients Blood From Healthy Donors (neutrophil-depleted) Efficiency M/R (all) M/R (⫹) ⫾ Efficiency M/R (all) M/R (⫹) ⫾ P1B 1.96 1.79 0 0 — 2.01 — 7/8 — 0 — 0 (0-0) 0 (0-0) — 0/15 HER2 1.74 7/65 2.02 3/12 0 0 (0-0) 0 (0-0) 0/15 K19 2.05 3/69 1.96 0.48 (0.48-0.48) — 0.0028 (0.0018-0.0058) 0.029 (0.015-0.048) 0.00024 (5.7E-5-0.00039) 0 — 24/48 0.003 (0-0.48) — 0.00013 (0-0056) 0.006 (0-0.048) 8,1E-5 (0-0.00039) 1/14 — 2.21 0.36 (0.02-1.79) — 0.0014 (0.0001-0.006) 0.036 (0.012-0.087) 0.00062 (0.00058-0.00068) 12/60 PS2 EGP2 0.051 (0-1.79) — 0.0005 (0-0.006) 0.0035 (0-0.087) 2.6E-5 (0-0.00062) 5/10 1.78 4.8E-5 (0-0.00053) 0.00029 (5E-5-0.00053) 2/13 Gene Efficiency M/R (all) M/R (⫹) ⫾ 0/15 — NOTE. Mean values are presented instead of median values as many of the samples are negative and thus the median value would be “0” in most of the boxes and thus poorly orientative. Abbreviations: M/R (all), mean relative values and range considering all patients; M/R (⫹), mean relative values and range considering only those patients with positive values for the given transcript; ⫾, samples positive and negative for the given transcripts among the 72 patients with valid samples. 3613 www.jco.org Downloaded from ascopubs.org by 78.47.27.170 on November 17, 2016 from 078.047.027.170 Copyright © 2016 American Society of Clinical Oncology. All rights reserved. Quintela-Fandino et al Table 4. Relative Values for Each Transcript in Peripheral Blood (containing neutrophils) From Healthy Donors Control P1B EGP2 HER-2 K19 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 0 0 0 0 0 0 0 0.489 0 0 0 0 0 0 0.0024 0.0019 0.0026 0 0.0057 0 0.0027 0 0 0 0 0 0.0029 0.0019 0 0 0.015 0 0 0 0 0.048 0 0 0 0 0 0 0 0.023 0.00037 0.00039 5.7E-05 0 0 0 0 0.00013 0 0 0 0 0 0.00026 0 NOTE. Values are represented as relative rates to b-glucuronidase. (15.5%; 95%CI, 7.8% to 23.2%) were alive but had relapsed, and 35 patients (41.7%; 95%CI, 31.2% to 52.2%) have died. RNA Yield, RT-PCR Efficiency, Target Gene Positivity, and Relative Values for Each Transcript Median Ct and Ct range for GUS, efficiency and number of positive and negative samples for patients, and negative controls are depicted in Table 2. The last Ct (intercept) that was considered positive for each gene was 38.51 (P1B), 40.42 (EGP2), 47.73 (HER-2), and 41.60 (K19). PS2 Ct was higher than 40 and higher than 35 in nondiluted cell line samples of HCC1143 and MCF7, respectively. Since Ct for the other targets was approximately 14, this yields a 30,000-fold lower expression of PS2 compared with the other targets. Therefore, PS2, although detectable in nondiluted tumor cell lines, as apheresis tumor cell contamination is about 1/106 nontumor cells, remained undetectable in every patient. Table 3 presents the efficiency and relative values for each gene among patients and negative controls. The individual results for each transcript in blood from healthy donors containing neutrophils are presented in Table 4. MM and Prognosis Patient flow diagram and Kaplan-Meier plots for time-torelapse and overall survival according to MM status are depicted in Figures 1 and 2. Multivariate analysis for relapse and survival is summarized in Table 5. Tumor size and presence of MM remained as independent factors on relapse (13.6% increase in hazard ratio [HR] per centimeter increase in tumor size and 269.6% increase for MM–positive). MM status maintained an independent prognostic influence (HR, 3) in overall survival. Considering the highest risk subgroup (ⱖ 10 nodes; n ⫽ 42), the HR among MM–positive patients was 2.55 for relapse (P ⫽ .031) and 3.05 for death (P ⫽ .030). Figure 3 contains the Kaplan-Meier plots for TTR among patients with circulating mRNA of K19–positive versus K19–negative; HER-2– positive versus HER-2–negative; EGP2–positive versus EGP2–negative; and P1B–positive versus P1B–negative. Log-rank P values were .22 (K19), .61(HER-2), and .64 (EGP2), respectively; P1B–positive versus P1B–negative median TTR was 23.2 months (95% CI, 6.2 to 40.2 months) versus 76.6 months (95% CI, 54.6 to 98.5 months; P ⫽ .01). The median overall survival for P1B–positive versus P1B–negative was 32.3 months (95% CI, 0 to 87.4 months) versus not reached (P ⫽ .03). Considered separately, K19, HER-2, and EGP2 (log-rank P ⫽ .17, .91, and .75, respectively) did not influence overall survival (plots not shown). When analyzing age, grade, tumor size, PR/ER status, treatment with paclitaxel, nodes, and relative values of circulating K19, HER-2, EGP2 and P1B mRNA in a multivariate Cox model for TTR (Rao P ⫽ .016; R2 ⫽ 0.0027), the value of circulating EGP2 and P1B mRNA showed a statistically significant independent impact (P ⫽ .003 and .017, respectively; HR ⫽ 103 and HR ⫽ 1.021). However, the extremely low R2 values invalidate these results. With regard to overall survival, the final model had an overall significance of .201; thus, no conclusions can be drawn regarding individual transcripts. Fig 1. Patient flow through the study. MM, micrometastases; TTR, time to relapse (in months); TTD, time to death (in months). 3614 JOURNAL OF CLINICAL ONCOLOGY Downloaded from ascopubs.org by 78.47.27.170 on November 17, 2016 from 078.047.027.170 Copyright © 2016 American Society of Clinical Oncology. All rights reserved. Micrometastases in Blood Predicts Poor Survival Table 5. Multivariate Analysis for Relapse and Survival Relapse Factor Age, years Tumor size, cm Grade Fictitious grade 1 Fictitious grade 2 ER PR Paclitaxel reception Micro Nodes Significance of the model/R 2 P Hazard Ratio .289 1.023 .0470 1.136 .924 .896 6219 .894 7100 .455 0.666 .724 0.841 .088 2.311 .006 2.696 .115 1.041 .030/.221 Overall Survival P Hazard Ratio .580 1.013 .276 1.088 .709 .923 3918 .920 5523 .235 0.484 .623 0.749 .394 1.723 .011 3.000 .122 1.044 .047/.191 NOTE. Grade, ER, PR, treatment with paclitaxel and micrometastatic activity were considered categorical variables. Age, tumor size and number of affected nodes were considered quantitative variables. The variables that reached statistically significant impact in relapse and overall survival are depicted in bold. Categorical variables with ⬎ two categories (grade) were automatically decomposed in fictitious variables for Cox’s analysis by the SPSS program (SPSS Inc, Chicago, IL). Abbreviations: ER, estrogen receptor; PR, progesterone receptor; micro, micrometastatic disease. The presence of one transcript was not found to be a predictor for any other. An association (r ⫽ 0.536; P ⫽ .000) was found between circulating K19 and EGP2. However the limited number of K19positive cases (n ⫽ 3) precluded any interpretation. DISCUSSION Fig 2. Kaplan-Meier plots for (A) time to relapse and (B) overall survival according to micrometastatic (MM) status. Relationship Between Conventional Prognostic Factors and MM Categoric variables (ER, PR, grade, HER-2) did not predict MM presence (P ⫽ .071, .358, .603, and .847, respectively). The correlation between the quantitative variables (age, tumor size, number of affected nodes) and the presence/absence of MM as well as the relative values of each transcript was also analyzed using logistic regression and Pearson analysis, respectively. The B coefficients for the logistic regression test were close to the null value (⫺0.003, 0.017, and ⫺0.011), with P values of .91, .87, and .746, respectively. No correlation was found in the Pearson analysis (data not shown), although there was a positive correlation (r ⫽ 0.279; P ⫽ .019) between age and amount of circulating EGP2. MD has been the subject of extensive research in breast cancer and the detection of K19 in hematopooietic tissue has been used as an indicator of the presence of residual tumor cells.8-26 Gene panels were first used in the study by Bosma et al,36 where samples from 103 metastatic patients were analyzed using immunohistochemistry to find circulating breast cancer cells. In our study a similar expression (rate and percentage) of the transcripts in the blood of patients and healthy donors has been found, but no expression in negative controls after removing neutrophils, suggesting that these genes may indicate the presence of MD in neutrophil-depleted samples. Even though the question of whether EGP2 (EP-CAM), P1B (TFF-3), HER-2, or K19 are expressed by neutrophils warrants further study, we performed the negative control analysis in neutrophildeplected peripheral blood samples as apheresis samples from patients did not contain neutrophils. K19 was present in neutrophil-depleted negative control (2 of 15) and patient (3 of 72) samples. Whether this can be explained by contamination from skin during venopuncture or from neutrophils cannot be concluded from our study, but led us to exclude K19 from the panel. Previous studies using K19 state neither whether their samples were neutrophil-depleted nor the K19-positive rate among negative controls.8-26 Even though K19 should not be expressed in neutrophils, future studies assessing MM with a genepanel should be performed with neutrophil-depleted samples in order to avoid misclassification, and results among negative controls should be presented. 3615 www.jco.org Downloaded from ascopubs.org by 78.47.27.170 on November 17, 2016 from 078.047.027.170 Copyright © 2016 American Society of Clinical Oncology. All rights reserved. Quintela-Fandino et al Fig 3. Kaplan-Meier plots for time to relapse according to presence/absence of each target transcript in peripheral blood. (A) HER-2, (B) K19, (C) EGP2, (D) P1B. Although in the study by Bosma et al36 the predictive power of the panel could be checked by immunohistochemistry, as the experiments were conducted with samples from active disease patients, we can sustain that the panel indicates MD, based on our prognostic results and the results form neutrophil-depleted negative controls. Previous studies have worked with epithelial markers, many of them showing no prognostic impact despite a high number of patients,8-26 and only two obtained significant prognostic impact in the multivariate analysis.24,25 Overall, the average K19 detection in those studies ranges 20% to 40%; however, the majority of patients included matched the low- or intermediate-risk category. The keratins are mainly expressed in the primary tumor and are associated with better prognosis than that attributed to keratinnegative breast cancer cells, while the tumor cell downregulates the epithelial markers when it reaches the blood flow.31-35 In case of accepting K19 as a MM marker in our series, we would have found MM only in 4.2% of the patients, in contrast to 59.7% using metastatic breast cancer markers found by serial analysis of gene expression by Bosma et al.36 Although our work was conducted in patients in the adjuvant setting, many of them were actually (micro)metastatic. Among the main studies performed in adjuvant breast cancer, to our knowledge, this is the first in where the prognostic impact for the patients with MD is found for patients with high-risk breast cancer undergoing high-dose chemotherapy; in addition, it is the first in where a gene panel appears yields more accurate prognostic information than the classic K19-based approach. Although in our series the prognostic information yielded by the panel is conserved both in the patients with four to nine nodes or morethan10nodes,whatreinforcesitsvalueoverK19-basedapproaches, the fact that the K19 is downregulated in highly metastatic breast cancer cells, should be taken into account. In our series, the lower-risk patients still have a high risk (four to nine positive nodes). These facts, and the results depicted in Tables 3 and 4, suggest that K19 alone may not be the ideal target for MM detection. 3616 JOURNAL OF CLINICAL ONCOLOGY Downloaded from ascopubs.org by 78.47.27.170 on November 17, 2016 from 078.047.027.170 Copyright © 2016 American Society of Clinical Oncology. All rights reserved. Micrometastases in Blood Predicts Poor Survival A recent meta-analysis by Braun et al40 set the HR for death among more than 4,000 breast cancer patients at 1.93 for those with bone marrow contamination. We obtained an HR attributable to MM of 3.0 for death in the Cox model, and a 55-month survival advantage in the Kaplan-Meier plots for patients without MM. The study design (which overcomes MM assessment difficulties: samples obtained months after surgery/chemotherapy; avoiding dormant cells in bone marrow and using several targets) and the high number of events have probably contributed to the ability to obtain these results with 84 patients. The similarity of the HR to that obtained by Braun led us to believe that the MM assessment in peripheral blood is accurate. When the panel is separated into individual transcripts, the extremely low percentage of variability in relapse explained by the Cox model (R2), the high P value of the model for death, and the results shown in Figure 3 suggest that the MM study should be performed with several transcripts rather than one. REFERENCES 1. Slamon DJ, Clark GM, Wong SG, et al: Human breast cancer: Correlation of relapse and survival with amplification of the HER-2/neu oncogene. Science 235:177-182, 1987 2. Muss HB, Thor AD, Berry DA, et al: c-erbB-2 expression and response to adjuvant therapy in women with node-positive early breast cancer. N Engl J Med 330:1260-1266, 1994 3. Paik S, Bryant J, Park C, et al: erbB-2 and response to doxorubicin in patients with axillary lymph node-positive, hormone receptor-negative breast cancer. J Natl Cancer Inst 90:1361-1370, 1998 4. Paik S, Bryant J, Tan-Chiu E, et al: HER2 and choice of adjuvant chemotherapy for invasive breast cancer: National Surgical Adjuvant Breast and Bowel Project Protocol B-15. J Natl Cancer Inst 92:19911998, 2000 5. Lacroix H, Iglehart JD, Skinner MA, et al: Overexpression of erbB-2 or EGF receptor proteins present in early stage mammary carcinoma is detected simultaneously in matched primary tumors and regional metastases. Oncogene 4:145-151, 1989 6. Niehans GA, Singleton TP, Dykoski D, et al: Stability of HER-2/neu expression over time and at multiple metastatic sites. J Natl Cancer Inst 85: 1230-1235, 1993 7. Braun S, Schlimok G, Heumos I, et al: ErbB2 overexpression on occult metastatic cells in bone marrow predicts poor clinical outcome of stage I-III breast cancer patients. Cancer Res 61:1890-1895, 2001 8. Pantel K, Schlimok G, Braun S, et al: Differential expression of proliferation-associated molecules in individual micrometastatic carcinoma cells. J Natl Cancer Inst 85:1419-1424, 1993 9. Vredenburgh JJ, Peters WP, Rosner G, et al: Detection of tumor cells in the bone marrow of stage IV breast cancer patients receiving high-dose chemotherapy: The role of induction chemotherapy. Bone Marrow Transplant 16:815-821, 1995 10. Diel IJ, Kaufmann M, Costa SD, et al: Micrometastatic breast cancer cells in bone marrow at primary surgery: Prognostic value in comparison with nodal status. J Natl Cancer Inst 88:1652-1658, 1996 11. Vannucchi AM, Bosi A, Glinz S, et al: Evaluation of breast tumour cell contamination in the bone marrow and leukapheresis collections by RT-PCR for We failed to demonstrate that the late relapse is due to HER-2– positive circulating tumor cells. We found that the rate of mRNA expression was almost 1000-fold lower than the other transcripts in the positive controls, what may preclude interpretations of the fact that HER-2 transcripts were not found in peripheral blood due to the low rate of cell contamination. The main limitations of our study are its retrospective nature and the sample size, although it has been the largest one to assess MM using RT-PCR in patients undergoing HDCT. This study was performed in order to ensure that the present approach merits further prospective validation. Given the facts that MM has been the most powerful prognostic factor in this series and that available therapies like antiangiogenic agents could be effective against MD, a prospective study should be conducted in a large, average-risk cohort in order to validate our results. cytokeratin-19 mRNA. Br J Haematol 103:610-617, 1998 12. Cooper BW, Moss TJ, Ross AA, et al: Occult tumor contamination of hematopoietic stem-cell products does not affect clinical outcome of autologous transplantation in patients with metastatic breast cancer. J Clin Oncol 16:3509-3517, 1998 13. Braun S, Hepp F, Sommer HL, et al: Tumorantigen heterogeneity of disseminated breast cancer cells: Implications for immunotherapy of minimal residual disease. Int J Cancer 84:1-5, 1999 14. Mansi JL, Gogas H, Bliss JM, et al: Outcome of primary-breast-cancer patients with micrometastases: A long-term follow-up study. Lancet 354:197202, 1999 15. Braun S, Pantel K, Muller P, et al: Cytokeratinpositive cells in the bone marrow and survival of patients with stage I, II, or III breast cancer. N Engl J Med 342:525-533, 2000 16. Lyda MH, Tetef M, Carter NH, et al: Keratin immunohistochemistry detects clinically significant metastases in bone marrow biopsy specimens in women with lobular breast carcinoma. Am J Surg Pathol 24:593-1599, 2000 17. Gerber B, Krause A, Muller H, et al: Simultaneous immunohistochemical detection of tumor cells in lymph nodes and bone marrow aspirates in breast cancer and its correlation with other prognostic factors. J Clin Oncol 19:960-971, 2001 18. Gebauer G, Fehm T, Merkle E, et al: Epithelial cells in bone marrow of breast cancer patients at time of primary surgery: Clinical outcome during long-term follow-up. J Clin Oncol 19:3669-3674, 2001 19. Schneeweiss A, Diel I, Hensel M, et al: Micrometastatic bone marrow cells at diagnosis have no impact on survival of primary breast cancer patients with extensive axillary lymph node involvement treated with stem cell-supported high-dose chemotherapy. Ann Oncol 15:1627-1632, 2004 20. Kasimir-Bauer S, Oberhoff C, Schindler AE, et al: A summary of two clinical studies on tumor cell dissemination in primary and metastatic breast cancer: Methods, prognostic significance and implication for alternative treatment protocols. Int J Oncol 20:1027-1034, 2002 21. Witzig TE, Bossy B, Kimlinger T, et al: Detection of circulating cytokeratin-positive cells in the blood of breast cancer patients using immunomagnetic enrichment and digital microscopy. Clin Cancer Res 8:1085-1091, 2002 22. Wong IH, Yeo W, Chan AT, et al: Quantitative correlation of cytokeratin 19 mRNA level in periph- eral blood with disease stage and metastasis in breast cancer patients: Potential prognostic implications. Int J Oncol 18:633-638, 2001 23. Hu XC, Chow LW: Detection of circulating breast cancer cells with multiple-marker RT-PCR assay. Anticancer Res 21:421-424, 2001 24. Xenidis N, Vlachonikolis I, Mavroudis D, et al: Peripheral blood circulating cytokeratin-19 mRNApositive cells after the completion of adjuvant chemotherapy in patients with operable breast cancer. Ann Oncol 14:849-855, 2003 25. Stathopoulou A, Vlachonikolis I, Mavroudis D, et al: Molecular detection of cytokeratin-19-positive cells in the peripheral blood of patients with operable breast cancer: Evaluation of their prognostic significance. J Clin Oncol 20:3404-3412, 2002 26. Silva JM, Dominguez G, Silva J, et al: Detection of epithelial messenger RNA in the plasma of breast cancer patients is associated with poor prognosis tumor characteristics. Clin Cancer Res 7:28212825, 2001 27. Fidler IJ, Ellis LM: The implications of angiogenesis for the biology and therapy of cancer metastasis. Cell 79:185-188, 1994 28. Folkman J: Angiogenesis in cancer, vascular, rheumatoid and other disease. Nat Med 1:27-31, 1995 29. Barz H, Barz D: The prevalence of single and synchronous multiple tumours in autopsy series and its dependence on age. Arch Geschwulstforsch 47:428-436, 1977 30. Braun S, Kentenich C, Janni W, et al: Lack of effect of adjuvant chemotherapy on the elimination of single dormant tumor cells in bone marrow of high-risk breast cancer patients. J Clin Oncol 18:8086, 2000 31. Thomas PA, Kirschmann DA, Cerhan JR, et al: Association between keratin and vimentin expression, malignant phenotype, and survival in postmenopausal breast cancer patients. Clin Cancer Res 5:2698-2703, 1999 32. Fuchs IB, Lichtenegger W, Buehler H, et al: The prognostic significance of epithelial-mesenchymal transition in breast cancer. Anticancer Res 22:34153419, 2002 33. Zajchowski DA, Bartholdi MF, Gong Y, et al: Identification of gene expression profiles that predict the aggressive behavior of breast cancer cells. Cancer Res 61:5168-5178, 2001 34. Woelfle U, Cloos J, Sauter G, et al: Molecular signature associated with bone marrow micrometastasis in human breast cancer. Cancer Res 63: 5679-5684, 2003 3617 www.jco.org Downloaded from ascopubs.org by 78.47.27.170 on November 17, 2016 from 078.047.027.170 Copyright © 2016 American Society of Clinical Oncology. All rights reserved. Quintela-Fandino et al 35. Woelfle U, Sauter G, Santjer S, et al: Downregulated expression of cytokeratin 18 promotes progression of human breast cancer. Clin Cancer Res 10:2670-2674, 2004 36. Bosma AJ, Weigelt B, Lambrechts AC, et al: Detection of circulating breast tumor cells by differential expression of marker genes. Clin Cancer Res 8:1871-1877, 2002 imal residual disease follow up in leukemia patients. Blood 98:4467-4475, 2001 39. Pfaffl MW: A new mathematical model for relative quantification in real-time RT-PCR. Nucleic Acids Res 29:2002-2007, 2001 40. Braun S, Vogl FD, Naume B, et al: A pooled analysis of bone marrow micrometastasis in breast cancer. N Engl J Med 353:793-802, 2005 37. Helfrich W, ten Poele R, Meersma GJ, et al: A quantitative reverse transcriptase polymerase chain reaction-based assay to detect carcinoma cells in peripheral blood. Br J Cancer 76:29-35, 1997 38. Pallisgaard N, Hokland P, Bi W, et al: Selection of referente genes for the European standardization and quality control program of real-time quantitative RT-PCR analysis of fusion gene transcripts for min■ ■ ■ Appendix The Appendix is included in the full-text version of this article, available online at www.jco.org. It is not included in the PDF version (via Adobe® Reader®). Authors’ Disclosures of Potential Conflicts of Interest Although all authors completed the disclosure declaration, the following authors or their immediate family members indicated a financial interest. No conflict exists for drugs or devices used in a study if they are not being evaluated as part of the investigation. For a detailed description of the disclosure categories, or for more information about ASCO’s conflict of interest policy, please refer to the Author Disclosure Declaration and the Disclosures of Potential Conflicts of Interest section in Information for Contributors. Authors Employment Leadership Consultant Stock Honoraria Research Funds Testimony Other Miguel QuintelaFandino Roche (A) Roche (A) Ricardo Hitt Roche (A) Roche (A) Soledad Gamarra Roche (A) Cecilia Guzman Roche (N/R) Hernan Cortes Funes Roche (A) Dollar Amount Codes (A) ⬍ $10,000 (B) $10,000-99,999 (C) ⱖ $100,000 (N/R) Not Required Author Contributions Conception and design: Miguel Quintela-Fandino, Joaquin Martinez, Ricardo Hitt, Antonio Jimeno Financial support: Cecilia Guzman, Hernan Cortes Funes Administrative support: Ricardo Hitt, Hernan Cortes Funes Provision of study materials or patients: Ricardo Hitt, Javier Hornedo, Cecilia Guzman, Hernan Cortes Funes Collection and assembly of data: Miguel Quintela-Fandino, Soledad Gamarra, Rosa Ayala Data analysis and interpretation: Miguel Quintela-Fandino, Joaquin Martinez, Ricardo Hitt Manuscript writing: Miguel Quintela-Fandino, Joaquin Martinez, Hernan Cortes Funes Final approval of manuscript: Miguel Quintela-Fandino, Joaquin Martinez, Ricardo Hitt, Soledad Gamarra, Antonio Jimeno, Rosa Ayala, Javier Hornedo, Florinda Gilsanz, Cecilia Guzman, Hernan Cortes Funes 3618 JOURNAL OF CLINICAL ONCOLOGY Downloaded from ascopubs.org by 78.47.27.170 on November 17, 2016 from 078.047.027.170 Copyright © 2016 American Society of Clinical Oncology. All rights reserved.