Prognostic relevance of disseminated tumour cells from the bone marrow of early stage breast cancer patients – Results from a large single-centre analysis

European Journal of Cancer (2014) 50, 2550– 2559

Available at www.sciencedirect.com

ScienceDirect journal homepage: www.ejcancer.com

Original Research

Prognostic relevance of disseminated tumour cells from the bone marrow of early stage breast cancer patients – Results from a large single-centre analysis Andreas D. Hartkopf a,⇑,1, Florin-Andrei Taran a,1, Markus Wallwiener b, Markus Hahn a, Sven Becker c, Erich-Franz Solomayer d, Sara Y. Brucker a, Tanja N. Fehm e, Diethelm Wallwiener a a

Department of Obstetrics and Gynaecology, University of Tuebingen, Calwerstrasse 7, 72076 Tuebingen, Germany Department of Obstetrics and Gynaecology, University of Heidelberg, Im Neuenheimer Feld 110, Heidelberg, Germany c Department of Obstetrics and Gynaecology, University of Frankfurt, Theodor-Stern-Kai 7, 60590 Frankfurt, Germany d Department of Obstetrics and Gynaecology, Saarland University, Kirrberger Strasse 100, 66424 Homburg, Germany e Department of Obstetrics and Gynaecology, University of Duesseldorf, Moorenstrasse 5, 40225 Duesseldorf, Germany b

Received 22 April 2014; received in revised form 6 June 2014; accepted 27 June 2014 Available online 2 August 2014

KEYWORDS Breast cancer Disseminated tumour cells Micrometastasis Prognosis Bisphosphonates

Abstract Background: This is the largest single-centre study to determine the prognostic relevance of disseminated tumour cells (DTCs) from the bone marrow (BM) of stage I-III breast cancer patients. Additionally, we aimed to analyse the impact of DTC detection on adjuvant bisphosphonate (BP) treatment efficacy. Methods: BM aspirates were collected during primary surgery for early breast cancer (EBC; T1–4, N0–2, M0) at Tuebingen University, Germany, between January 2001 and January 2013. DTCs were identified by immunocytochemistry (pancytokeratin antibody A45/B-B3) and cytomorphology. We retrospectively estimated the influence of DTC detection and BP treatment on disease-free survival (DFS) and overall survival (OS) using univariate (log-rank test) and multivariate (cox regression) analysis. Findings: BM aspirates were available from 3141 patients. In 803 (26%) of these, DTCs were detectable. As compared to DTC-negative patients, DTC-positive patients more frequently had larger tumors (p < 0.001), lymph node involvement (p < 0.001), hormonal receptor positive tumours (p < 0.001) and HER2-positive tumours (p = 0.048). DTC-positive patients were at an increased risk of relapse (hazard ratio (HR) 1.74, 95% confidence interval (CI) 1.34–2.25, p < 0.001) and death (HR 1.44 95% CI 1.13–1.86, p = 0.004). In the multivariate analysis

⇑ Corresponding author: Address: Department of Obstetrics and Gynaecology, Tuebingen University Hospital, Calwerstrasse 7, 72076 Tuebingen, Germany. Tel.: +49 7071 2980791; fax: +49 7071 294805. E-mail address: [email protected] (A.D. Hartkopf). 1 Equal contributors and joint first authors.

http://dx.doi.org/10.1016/j.ejca.2014.06.025 0959-8049/Ó 2014 Elsevier Ltd. All rights reserved.

A.D. Hartkopf et al. / European Journal of Cancer 50 (2014) 2550–2559

2551

DTCs were an independent predictor of DSF and OS. Additionally, BP treatment had no significant influence on DFS or OS in DTC-negative patients, while it was significantly associated with increased DFS (p < 0.001) and OS (p = 0.006) in DTC-positive patients. Interpretation: These data confirm the clinical validity of DTCs from the BM for prognostication of early breast cancer patients. Further studies are warranted to determine whether DTCs are predictive for adjuvant treatment efficacy using bisphosphonates. Ó 2014 Elsevier Ltd. All rights reserved.

1. Introduction Early breast cancer (EBC) can relapse even years after successful treatment of the primary tumour. It has therefore been hypothesised that individual tumour cells spread to secondary sites in the body where they can persist for long periods of time before initiating metastatic growth [1]. This phenomenon is termed minimal residual disease (MRD), and the aim of all adjuvant therapies is to eradicate MRD before it becomes clinically evident. Evidence has been increasing that cytokeratin (CK)positive disseminated tumour cells (DTCs) in the bone marrow (BM) of breast cancer patients may reflect the presence of MRD [2]. DTCs are detected in the BM of 30–40% of EBC patients and associated with a worsened prognosis [3–7]. The strong independent prognostic significance of DTCs at primary diagnosis was confirmed in a large pooled analysis of more than 4700 patients published by Braun and colleagues in 2005 [7]. Additionally, the persistence of DTCs during adjuvant treatment of stages I–III breast cancer predicts an increased risk of disease relapse and death [8,9]. Thus, the detection of DTCs might serve as an indicator of systemic treatment efficacy, helping to identify patients in need of additional adjuvant treatment. However, chemotherapy regimes often fail to eliminate DTCs, which can persist in the BM in a dormant, non-proliferative state for years [10]. Another approach to eradicating DTCs from the BM might be the use of bisphosphonates (BPs), which primarily target the skeletal system [11]. However, the role of BPs in EBC treatment remains unclear. A large meta-analysis investigating the addition of BPs to adjuvant therapy in patients with EBC found that menopausal status was a predictor of BP treatment efficacy [12]. Additionally, recent findings indicate that successful eradication of DTCs with BP therapy improves prognosis in patients with EBC [11,13]. Although the results highlighting the potential usefulness of DTC detection in the BM of patients with EBC have been encouraging, DTC determination has not thus far been recommended for routine clinical use by guidelines or expert panels due to a lack of consensus on certain methodological and institutional problems. The pooled analysis by Braun et al. [7] covered seven different techniques for DTC detection. Moreover, while patient enrolment in the largest published studies ended

in 2002, the past decade has seen a number of significant changes in the treatment of patients with EBC, including primary systemic therapy and HER2-targeted therapy. Against this background, we conducted the present large single-centre study to investigate the impact of DTCs on prognosis in patients with early stage I–III breast cancer. DTCs were determined according to a standardised method recommended by expert consensus [14]. We moreover determined the impact of DTC detection on adjuvant BP efficacy by analysing survival outcomes in DTC-negative and DTC-positive patients treated with or without BPs. 2. Methods 2.1. Study population, setting, design, and ethics Women undergoing primary surgery for EBC (T1–4, N0–2, M0) at the Department of Obstetrics and Gynaecology at Tuebingen University Hospital, Tuebingen, Germany, between January, 2001, and January, 2013, were eligible for this retrospective study. Patients with recurrent or metastatic disease, bilateral breast cancer, R1 resection or a previous history of secondary malignancy were excluded. All patients provided written informed consent for BM aspiration. The analysis was approved by the ethics committee of the University of Tuebingen (reference number 560/2012R). 2.2. Systemic treatment Systemic treatment was based on national treatment guidelines (www.ago-online.de). Most patients were treatment-naı¨ve at the time of DTC determination. However, as bone marrow aspiration was performed exclusively during primary surgery, a small proportion of patients received preoperative treatment (ie, neoadjuvant chemotherapy) before being evaluated for DTC status. In patients who were treatment-naı¨ve at the time of surgery, tumour stage (ie, tumour size and nodal status) was determined by pathology. If patients received neoadjuvant chemotherapy, tumour size and nodal status were determined by clinical examination and imaging modalities before the first treatment cycle. The preferred imaging modality was ultrasound, but magnetic resonance imaging and mammography served as alternatives if ultrasound findings appeared not to be valid.

2552

A.D. Hartkopf et al. / European Journal of Cancer 50 (2014) 2550–2559

As presented in Supplementary Table 1, adjuvant BP treatment was performed in the context of clinical trials that included BP treatment (GAIN (NCT00196872), SUCCESS A (www.success-studie.de), or NATAN (NCT00512993)). Additionally, based on existing clinical evidence and in accordance with German guidelines (www.ago-online.de), patients were offered treatment with zoledronate at 4 mg every six months outside of clinical trials. 2.3. DTC studies BM aspirates (10–20 ml per patient) were collected during primary surgery and processed within 24 h. Briefly, mononuclear cells were isolated by density centrifugation (Ficoll, 1.077 g/ml, Biochrom, Berlin, Germany), spun down onto a glass slide (cytocentrifuge, Hettich, Tuttlingen, Germany), and fixed in 4% formalin. The presence of DTCs (DTC status) was determined by immunostaining using the DAKO Autostainer (Dako, Glostrup, Denmark), the mouse monoclonal antibody A45-B/B3 directed against pancytokeratin (Micromet, Munich, Germany) and the DAKO-APAA detection kit (Dako, Glostrup, Denmark). Two slides (1  106 cells, each) per patient were evaluated, in accordance with consensus recommendations for standardised tumour cell detection [14,15]. An additional slide was stained with an unspecific isotype-matched antibody. Moreover, each batch of samples was analysed together with leukocytes from healthy volunteers as negative controls and the human breast cancer cell lines MCF-7 and SKBR-3 as positive controls. To assess the specificity of our method for DTC detection, we analysed BM samples from 100 patients without evidence of malignant disease, of whom one was DTC-positive [8]. 2.4. Statistical analysis Associations between categorical variables (DTC status and patient characteristics) were analysed using the chi-square test. Predictors of BM involvement were identified by multivariate logistic regression analysis. Factors that achieved statistical significance at p < 0.1 in the univariate analysis were considered for multivariate analysis. Variables were entered according to strength of crude association, using a step-wise forward technique. Odds ratios (ORs) and corresponding 95% confidence intervals (CIs) were calculated. To determine survival, times from BM aspiration to any recurrence of disease (disease free survival, DFS) and death of any cause (overall survival, OS), were investigated separately. If no event occurred, data were censored at last follow-up. The influence of DTC status on survival was determined by univariate analysis and expressed as a hazard ratio (HR) and 95% CI. Kaplan–Meier curves were plotted and compared by the log-rank test. Multivariate analysis of survival

used a Cox proportional regression model. Variables were entered stepwise backward, using a significance level of 0.1 to exclude a variable from the model. The initial model included menopausal status, histology, grading, nodal status, tumour size, oestrogen/progesterone/epidermal growth factor receptor 2 (ER/PR/HER2) status, BP treatment and DTC status. The effect of each variable was assessed using the Wald test and expressed as an HR and 95% CI. All statistical tests were performed using PASW Statistics 21 (SPSS Inc., Chicago, IL, USA), and all reported p values are for two-sided tests with the significance level set at p < 0.05. 3. Results 3.1. Patient characteristics In total, 3141 EBC patients were eligible for inclusion. Patient characteristics are detailed in Table 1. Most patients had invasive ductal carcinoma (79%). Tumours were mostly grade 2 (67%), pT1 (61%) or pT2 (31%). Most patients were node-negative (65%) and oestrogen receptor (ER; 83%) or progesterone receptor (PR; 75%) positive. Epidermal growth factor receptor 2 (HER2) was overexpressed by the primary tumour in 470 (16%) patients. More than half of patients received neoadjuvant (12%) or adjuvant (41%) chemotherapy or adjuvant endocrine treatment (79%). As shown in Supplementary Table 2, data on adjuvant BP treatment were available in 3113 patients, and 876 patients (28%) received BP treatment. 3.2. DTC studies DTCs were detected in the BM of 803/3141 (26%) patients at the time of primary surgery (Table 1). The incidence of DTC-positive patients was highly associated with tumour size (p < 0.001). Of 95 patients with T4 tumours (invasion of contiguous structures or inflammatory breast cancer) 37 (39%) were DTC-positive as compared to only 418/1868 (22%) patients with T1 tumours. DTC positivity was also increased in patients with axillary lymph node involvement (319/ 1092 (29%) compared to 482/2030 node-negative patients (24%), p < 0.001). With respect to tumour biology, we found the incidence of DTCs to be higher in more aggressive (ER/PR/HER2-negative) tumours; 507/2176 (23%) patients with luminal-like (ie, hormone receptor-positive/HER2-negative) tumours, 132/469 (28%) patients with HER2-positive and 80/265 (30%) patients with basal-like (ie, hormone receptor-negative and HER2-negative) were DTC-positive (p = 0.008). Moreover, we found a strong association between neoadjuvant cytotoxic treatment, and DTC status. Of the 391 neoadjuvantly treated patients, 160 (41%) were DTC-positive whereas only 305/1277 (25%) adjuvantly treated

A.D. Hartkopf et al. / European Journal of Cancer 50 (2014) 2550–2559

2553

Table 1 Patient characteristics. Total n

DTC positive n (%)

p value

Total Menopausal status Premenopausal Postmenopausal

3141 938 2203

803 (26) 251 (27) 552 (25)

0.317

Histology Invasive ductal Invasive lobular Other Unknown

2454 501 162 24

638 (26) 120 (24) 37 (23) 8 (33)

0.459

Tumour grade I II III Unknown

351 2059 653 78

77 (22) 520 (25) 184 (28) 22 (28)

0.087

Tumour size T1 T2 T3 T4 Unknown

1868 963 137 95 78

418 (22) 264 (27) 52 (38) 37 (39) 32 (41)

<0.001

Nodal status Negative Positive Unknown

2030 1092 19

482 (24) 319 (29) 2 (11)

0.001

ER status Negative Positive

547 2594

161 (29) 642 (25)

0.022

PR status Negative Positive Unknown

785 2347 9

240 (31) 562 (24) 1 (11)

<0.001

HER2 status Negative Positive Unknown

2442 470 229

586 (24) 133 (28) 84 (37)

0.048

Molecular status Basal-like (HormR /HER2 ) Luminal-like (HormR+/HER2 ) HER2 type (HER2+) Unknown

265 2176 470 230

80 (30) 507 (23) 132 (28) 84 (37)

0.008

Chemotherapy No Yes (neoadjuvant) Yes (adjuvant)

1473 391 1277

338 (23) 160 (41) 305 (24)

<0.001

DTC = disseminated tumour cell. ER = oestrogen receptor. PR = progesterone receptor. HER2 = epidermal growth factor receptor 2. HormR = hormone receptor (ie, ER, PR, or both). Bold p values indicate statistical significance.

patients and 338/1473 (24%) patients who were not treated with chemotherapy harboured DTCs (p < 0.001). Predictors of DTC positivity were determined by multivariate logistic regression analysis. Table 2 shows that cytotoxic treatment before BM sampling was the strongest risk factor for DTC detection (OR 1.69, 95% CI 1.22–2.36), followed by negative PR status (OR 1.34, 95% CI 1.10–1.64). The remaining factors had no impact on DTC status in the multivariate analysis.

3.3. Survival analysis Patients followed up for less than six months were not included in the survival analysis. Hence, DFS and OS were based on follow-up data for 2809 and 2931 patients, respectively. Median follow-up (95% CI) was 43 (41–44) months for DSF and 53 (52–55) months for OS. By the time of the present analysis, 18 February 2014, 255 patients had developed recurrent disease and

2554

A.D. Hartkopf et al. / European Journal of Cancer 50 (2014) 2550–2559

Table 2 Multivariate logistic regression analysis of predictors of bone marrow involvement. Parameter

OR

95% CI

p value

PR status Positive Negative

1.00 1.34

1.10–1.64

0.004

Chemotherapy None Adjuvant Neoadjuvant

1.00 0.96 1.69

0.79–1.17 1.22–2.36

0.676 0.002

OR = odds ratio. CI = confidence interval. receptor. Bold p values indicate statistical significance.

PR = progesterone

289 patients had died. Fig. 1 shows that DTC-positive patients had a significantly greater risk of relapse and death than did DTC-negative patients. The HR (95% CI) for relapse was 1.74 (1.34–2.25) (p < 0.001) and the HR for death was 1.44 (1.13–1.86) (p = 0.004). In the multivariate analysis (Table 3), independent predictors of reduced DFS were DTC status, tumour size, nodal status, ER status and whether patients received BPs or not. Independent predictors of reduced OS were DTC status, menopausal status, tumour size, nodal status and whether patients received BPs or not. Subsequently we performed a subgroup analysis to determine whether neoadjuvant cytotoxic treatment or adjuvant BPs would influence the prognostic value of determining DTC status. As illustrated in Figs. 2 and 3, adjuvant BP treatment appeared to reduce the influence of DTC status on DFS and OS. In patients who did not receive BPs, the respective HRs (95% CIs) of relapse and death were 2.17 (1.62–2.88) and 1.67 (1.26–2.21). Patients treated with BPs had respective HRs (95% CIs) of relapse and death of 1.36 (0.70–2.67) and 1.16 (0.58–2.30). As shown in Fig. 4, BP treatment had no significant influence on DFS or OS in DTC-negative patients. In contrast, BP treatment was significantly associated with increased DFS (p < 0.001) and OS (p = 0.006) in

DTC-positive patients. Supplementary Fig. 1 shows the effect of BP on the survival of DTC-positive patients, stratified by menopausal status. DFS of DTC-positive patients was significantly longer after adjuvant BP treatment in both premenopausal (p = 0.018) and postmenopausal (p = 0.014) patients. In contrast, OS of DTC-positive patients given adjuvant BP treatment was improved only in postmenopausal women (p = 0.009). 4. Discussion Numerous studies have shown that EBC with DTCs in their BM have worse outcomes than those without DTCs [3–5,16]. In a large pooled analysis, DTCs were found in 31% of 4703 patients with stage I–III breast cancer [7]. Detection of DTCs was associated with poor overall and poor breast-cancer–specific survival, with more than twofold increased HR. Multivariate analysis revealed positive DTC status to be the strongest independent prognostic factor. However, the pooled analysis included data from several institutions using methods of DTC detection that had not been standardised. In 2006, the German, Austrian, and Swiss Societies of Senology reached a consensus on standardised DTC detection [14]. Using these criteria, we conducted what we believe to be the largest single-centre study to date on DTC determination in EBC. Our data confirm the strong evidence supporting the independent adverse prognostic significance of DTCs in the BM of EBC patients. Moreover, subgroup analysis of different adjuvant treatment modalities yielded evidence that BP therapy might improve the prognosis of DTC-positive patients. Although DTCs were more frequent in patients with larger tumours, positive nodal status and aggressive, HER2-positive or hormone receptor-negative tumours, multivariate analysis revealed that the prognostic significance of DTC presence was independent of these risk factors. Notably, we found the incidence of DTC positivity to be almost twice as high in patients receiving cytotoxic therapy before DTC determination. To eliminate a

100

80 60 40

p<0.001 DTC-negative (events/n = 165/2103) DTC-positive (events/n = 90/706)

20

Overall survival (%)

Disease-free survival (%)

100

80 60 40

p=0.004 DTC-negative (events/n = 198/2199) DTC-positive (events/n = 91/732)

20 0

0 0

20

40

Months

60

80

0

20

40

60

80

Months

Fig. 1. Disease-free survival (left) and overall survival (right) by disseminated tumour cell status (all patients). DTC = disseminated tumour cell. n = number of patients.

A.D. Hartkopf et al. / European Journal of Cancer 50 (2014) 2550–2559

2555

Table 3 Multivariate Cox regression analysis of disease free survival (DFS) and overall survival (OS). Parameter

DFS

OS

HR

95% CI

p value

HR

95% CI

p value

DTC status Negative Positive

1.00 1.81

1.37–2.43

<0.001

1.00 1.45

1.08–1.94

0.013

Menopausal status Premenopausal Postmenopausal

–

–

NS

1.00 2.26

1.55–3.27

<0.001

Tumour size T1 T2–4

1.00 1.70

1.28–2.25

<0.001

1.00 1.63

1.23–2.15

0.001

Nodal status Negative Positive

1.00 1.96

1.48–2.60

<0.001

1.00 1.74

1.32–2.28

<0.001

ER status Positive Negative

1.00 1.99

1.43–2.60

<0.001

–

–

NS

BP treatment No Yes

1.00 0.53

0.36–0.79

0.002

1.00 0.60

0.40–0.90

0.013

HR = hazard ratio. CI = confidence interval. DTC = disseminated tumour cell. PR = progesterone receptor. BP = bisphosphonates. Bold p values indicate statistical significance.

Events/n

HR and 95% CI

DTC-pos.

DTC-neg.

90/706

165/2103

No CT

23/284

55/975

Neoadjuvant CT

33/141

30/208

Adjuvant CT

34/281

80/920

No

69/397

145/1590

Yes

17/194

20/304

All patients

DTC-positive : DTC-negative

Chemotherapy

BP therapy

DTC-positive better

DTC-negative better

Fig. 2. Influence of disseminated tumour cell (DTC) status on disease-free survival (DFS) by type of systemic treatment. n = number of patients. HR = hazard ratio. CI = confidence interval. DTC = disseminated tumour cell. CT = chemotherapy. BP = bisphosphonate.

potential bias in the increased rate of DTC-positive patients associated with tumour characteristics favouring neoadjuvant cytotoxic treatment, we performed a multivariate logistic regression analysis. After adjusting for other factors associated with DTC detection in the univariate analysis, neoadjuvant cytotoxic treatment before BM aspiration emerged as the strongest predictor of positive DTC status, followed by negative PR status. This finding suggests that chemotherapy may be capable

of mobilising tumour cells to the bone marrow, a phenomenon previously described by others [11]. We recently demonstrated that persistence of DTCs after neoadjuvant chemotherapy is independent of the primary tumour’s response to treatment [17]. As DTC detection was associated with impaired DFS in patients receiving neoadjuvant systemic therapy, even patients with pathological complete remission but persistent DTCs might benefit from additional adjuvant therapy.

2556

A.D. Hartkopf et al. / European Journal of Cancer 50 (2014) 2550–2559

Events/n DTC-pos. All patients

HR and 95% CI x

DTC-neg.

91/732

198/2199

No CT

46/320

104/1082

Neoadjuvant CT

21/153

28/221

Adjuvant CT

24/259

66/896

DTC-positive : DTC-negative x

Chemotherapy

BP therapy No

70/427

177/1637

Yes

19/296

17/544

DTC-negative better

DTC-positive better

Fig. 3. Influence of disseminated tumour cell (DTC) status on overall survival (OS) by type of systemic treatment. n = number of patients. HR = hazard ratio. CI = confidence interval. DTC = disseminated tumour cell. CT = chemotherapy. BP = bisphosphonate.

DTC-negative

DTC-negative 100

Overall survival (%)

Disease-free survival (%)

100 80 60 40

p=0.235 20

80 60 40

p=0.121 20

BP treatment (events/n = 17/494) No BP treatment (events/n = 145/1590)

BP treatment (events/n = 17/ 544) No BP treatment (events/n = 177/1637)

0

0 0

20

40

60

0

80

20

DTC-positive

60

80

DTC-positive 100

Overall survival (%)

100

Disease-free survival (%)

40

Months

Months

80 60 40

p<0.001 20

BP treatment (events/n = 69/397) No BP treatment (events/n = 20/304)

0

80 60 40

p=0.006 20

BP treatment (events/n = 19/296) No BP treatment (events/n = 70/427)

0 0

20

40

Months

60

80

0

20

40

60

80

months

Fig. 4. Disease-free survival (left) and overall survival (right) of DTC-positive (top) and DTC-negative (bottom) patients with or without bisphosphonate treatment. n = number of patients. BP = bisphosphonate.

A.D. Hartkopf et al. / European Journal of Cancer 50 (2014) 2550–2559

Recent results on the prognostic impact of DTC detection during follow-up of EBC patients are in line with these findings and highlight the role of DTC determination in monitoring the efficacy of systemic therapy [8,9]. With DTCs being in a non-proliferative state, however, chemotherapy might not be an appropriate route to their elimination, which is in agreement with the high incidence of DTC positivity we observed in patients receiving neoadjuvant cytotoxic treatment [10]. As regards the anticancer effects of BPs in DTC-positive breast cancer patients, earlier studies have suggested that clodronate can reduce the incidence and number of bone and visceral metastases in these patients [11,18]. More recently, several pilot and phase II studies have reported that zoledronic acid, administered at 4 mg every three or four weeks for six to 24 months in combination with standard neoadjuvant or adjuvant systemic therapy, improves the elimination of DTCs and reduces the number of DTCs in the BM compared with standard therapy alone [19–21]. Additionally, in a follow-up of the MRD-1 study, Banys et al. reported that zoledronic acid had a positive influence on survival in DTC-positive breast cancer patients [13]. This BP-induced reduction in DTC persistence might be one of the mechanisms underlying the improved clinical outcome observed in two large randomised clinical trials of adjuvant BP therapy. The ABCSG-12 and ZO-FAST trials demonstrated a significantly prolonged disease-free survival in EBC patients receiving zoledronic acid [22,23]. In contrast to these two trials, the AZURE trial could not confirm these results [24]. A meta-analysis of individual patient data from randomised studies on the effects of BP treatment on recurrence and cause-specific mortality in 22982 EBC patients showed that adjuvant BPs reduce bone metastases and improve survival only in postmenopausal women [12]. However, BM biopsies to assess the DTC status were not performed in any of these trials. Our subgroup analysis revealed that BP treatment interacts with the prognostic value of DTC detection. This prompted us to evaluate retrospectively the impact on prognosis that adding BPs to adjuvant therapy might have compared with DTC status and menopausal status. Whereas BP treatment improved prognosis in both premenopausal and postmenopausal DTC-positive women, it had no detectable impact on DFS or OS in DTC-negative patients. However, these findings are limited by the retrospective nature of our study and subgroup analyses should generally be interpreted with caution. Especially non-significant findings might be due to small sample sizes. Moreover, the decision to use adjuvant BPs was either due to enrolment in studies that included adjuvant BP therapy or based on the decision of the treating physician. Hence, as shown in Supplementary Table 2, BPs were mostly used in postmenopausal, hormonal receptor positive and DTC positive patients. Although the prog-

2557

nostic benefit that patients achieved from receiving BPs remained highly significant after adjustment for other prognostic factors, the non-randomisation may result in a significant selection bias. Therefore, further prospective trials are needed to confirm the clinical benefit of BPs in DTC-positive patients and to clarify whether DTC detection predicts BP treatment efficacy. A nonrandomised phase II study is currently underway to assess the impact of the monoclonal antibody denosumab on DTCs in patients with early-stage breast cancer (NCT01545648). Another approach to specifically eradicate MRD might be the use of targeted therapy. Generally, established biomarkers for individualised breast cancer therapy, including ER and HER2 status, are used to characterise primary tumour tissue to predict response to treatment. This approach assumes that the primary tumour is representative of total tumour burden, including MRD, and that its initial phenotype will not change during the course of the disease. However, DTCs can exhibit characteristics that differ from those of the primary tumour, especially with respect to ER and HER2 expression. We recently found that patients with HER2-positive DTCs at diagnosis were at an increased risk of disease relapse, indicating that these patients might benefit from additional HER2-directed treatment [25]. A major limitation in assessing BM for DTCs is the invasiveness of BM aspiration. Recent research has focused on examining peripheral blood samples for circulating tumour cells (CTCs). Whereas CTC enumeration has been demonstrated to be of prognostic relevance in patients with metastatic breast cancer, low detection rates make the use of this technology challenging in EBC [26,27]. The translational research programme of the German SUCCESS trial (www. success-studie.de) evaluated peripheral blood samples from high-risk breast cancer patients for CTCs using the CellSearch system and reported that disease-free survival was significantly reduced in women with CTC counts of one or more CTCs per 7.5 ml blood [27]. Molloy et al. recently found that CTC detection in EBC patients was associated with the presence of DTCs [28]. However, the prognostic information that DTCs provide appears to be stronger than with CTCs, and other reports comparing CTC and DTC detection do not confirm these results, likely reasons being methodological differences between studies and the low sensitivity of CTC detection [6,29,30]. Hence, future studies should analyse DTCs and CTCs in parallel. In conclusion, our data confirm the prognostic relevance of DTC detection in patients with early stage I-III breast cancer. The standardised method we used to detect DTCs is suitable for prospective clinical trials of adjuvant breast cancer treatment. Such studies could include additional targeted therapy or BP treatment to eradicate DTCs from the BM.

2558

A.D. Hartkopf et al. / European Journal of Cancer 50 (2014) 2550–2559

5. Research in context

Appendix A. Supplementary data

5.1. Systematic review

Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/ 10.1016/j.ejca.2014.06.025.

Based on results from several trials that provided evidence on the prognostic validity of disseminated tumour cell (DTC) detection in early breast cancer, we started bone marrow sampling as a routine procedure at our department in January 2001. Before conducting this retrospective analysis, we searched PubMed with the following term: (‘disseminated tumour cells’ or ‘micrometastasis’) and ‘breast cancer’. Various methods were described to detect DTCs. Hence, our aim was to confirm the prognostic value of DTC determination in the largest single centre study so far, while using a standardised method for DTC detection. Additionally, recent trials found a potential impact of the DTC status on the efficacy of bisphosphonate treatment. Therefore, we performed subgroup analyses of patients that received bisphosphonates versus not. 5.2. Interpretation Our study confirms the independent prognostic validity of the DTC status on progression-free and overall survival. Additionally, the DTC status seemed predictive for bisphosphonate treatment efficacy. The standardised method we used is suitable for clinical trials of adjuvant breast cancer treatment. Such studies could include additional targeted therapy or BP treatment to eradicate DTCs from the BM. Funding None. Contributors ADH, FAT and DW conceived of the study and designed it. ADH and FAT wrote the draft manuscript. MW, MH, SB, EFS, SYB and TF participated in patient recruitment, patient management, clinical data collection and sample collection and analysis. Conflict of interest statement None declared. Acknowledgments We are grateful to Silke Duerr-Sto¨rzer, Ingrid Teufel, Sabine Hofmeister, Angelika Amman and Brigitte Neth for excellent technical assistance.

References [1] Klein CA. Parallel progression of primary tumours and metastases. Nat Rev Cancer 2009;9(4):302–12. [2] Pantel K, Brakenhoff RH, Brandt B. Detection, clinical relevance and specific biological properties of disseminating tumour cells. Nat Rev Cancer 2008;8(5):329–40. [3] 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 1996;88(22):1652–8. [4] Gebauer G, Fehm T, Merkle E, Beck EP, Lang N, Jager W. Epithelial cells in bone marrow of breast cancer patients at time of primary surgery: clinical outcome during long-term follow-up. J Clin Oncol 2001;19(16):3669–74. [5] Mansi JL, Gogas H, Bliss JM, Gazet JC, Berger U, Coombes RC. Outcome of primary-breast-cancer patients with micrometastases: a long-term follow-up study. Lancet 1999;354(9174):197–202. [6] Wiedswang G, Borgen E, Karesen R, et al. Detection of isolated tumor cells in bone marrow is an independent prognostic factor in breast cancer. J Clin Oncol 2003;21(18):3469–78. [7] Braun S, Vogl FD, Naume B, et al. A pooled analysis of bone marrow micrometastasis in breast cancer. N Engl J Med 2005;353(8):793–802. [8] Janni W, Vogl FD, Wiedswang G, et al. Persistence of disseminated tumor cells in the bone marrow of breast cancer patients predicts increased risk for relapse–a European pooled analysis. Clin Cancer Res 2011;17(9):2967–76. [9] Gruber I, Fehm T, Taran FA, et al. Disseminated tumor cells as a monitoring tool for adjuvant therapy in patients with primary breast cancer. Breast Cancer Res Treat 2014;144(2):353–60. [10] 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 2000;18(1):80–6. [11] Diel IJ, Jaschke A, Solomayer EF, et al. Adjuvant oral clodronate improves the overall survival of primary breast cancer patients with micrometastases to the bone marrow: a long-term follow-up. Ann Oncol 2008;19(12):2007–11. [12] Coleman R, Gnant M, Paterson A, et al. Effects of bisphosphonate treatment on recurrence and cause-specific mortality in women with early breast cancer: a meta-analysis of individual patient data from randomised trials. San Antonio Breast Cancer Symp 2013:S4–07. [13] Banys M, Solomayer EF, Gebauer G, et al. Influence of zoledronic acid on disseminated tumor cells in bone marrow and survival: results of a prospective clinical trial. BMC Cancer 2013;13:480. [14] Fehm T, Braun S, Muller V, et al. A concept for the standardized detection of disseminated tumor cells in bone marrow from patients with primary breast cancer and its clinical implementation. Cancer 2006;107(5):885–92. [15] Borgen E, Naume B, Nesland JM, et al. Standardization of the immunocytochemical detection of cancer cells in BM and blood: I. establishment of objective criteria for the evaluation of immunostained cells. Cytotherapy 1999;1(5):377–88. [16] 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(8):525–33. [17] Hartkopf AD, Taran FA, Wallwiener M, et al. The presence and prognostic impact of apoptotic and nonapoptotic disseminated tumor cells in the bone marrow of primary breast cancer patients after neoadjuvant chemotherapy. Breast Cancer Res 2013;15(5):R94.

A.D. Hartkopf et al. / European Journal of Cancer 50 (2014) 2550–2559 [18] Diel IJ, Solomayer EF, Costa SD, et al. Reduction in new metastases in breast cancer with adjuvant clodronate treatment. N Engl J Med 1998;339(6):357–63. [19] Solomayer EF, Gebauer G, Hirnle P, et al. Influence of zoledronic acid on disseminated tumor cells in primary breast cancer patients. Ann Oncol 2012;23(9):2271–7. [20] Aft R, Naughton M, Trinkaus K, et al. Effect of zoledronic acid on disseminated tumour cells in women with locally advanced breast cancer: an open label, randomised, phase 2 trial. Lancet Oncol 2010;11(5):421–8. [21] Rack B, Juckstock J, Genss EM, et al. Effect of zoledronate on persisting isolated tumour cells in patients with early breast cancer. Anticancer Res 2010;30(5):1807–13. [22] Coleman R, de Boer R, Eidtmann H, et al. Zoledronic acid (zoledronate) for postmenopausal women with early breast cancer receiving adjuvant letrozole (ZO-FAST study): final 60-month results. Ann Oncol 2013;24(2):398–405. [23] Gnant M, Mlineritsch B, Stoeger H, et al. Adjuvant endocrine therapy plus zoledronic acid in premenopausal women with earlystage breast cancer: 62-month follow-up from the ABCSG-12 randomised trial. Lancet Oncol 2011;12(7):631–41. [24] Coleman RE, Marshall H, Cameron D, et al. Breast-cancer adjuvant therapy with zoledronic acid. N Engl J Med 2011;365(15): 1396–405.

2559

[25] Hartkopf AD, Banys M, Meier-Stiegen F, et al. The HER2 status of disseminated tumor cells in the bone marrow of early breast cancer patients is independent from primary tumor and predicts higher risk of relapse. Breast Cancer Res Treat 2013;138(2):509–17. [26] Cristofanilli M, Budd GT, Ellis MJ, et al. Circulating tumor cells, disease progression, and survival in metastatic breast cancer. N Engl J Med 2004;351(8):781–91. [27] Rack BK, Schindlbeck C, Andergassen U, et al. Use of circulating tumor cells (CTC) in peripheral blood of breast cancer patients before and after adjuvant chemotherapy to predict risk for relapse: The SUCCESS trial. J Clin Oncol 2010;28(Suppl. 15s). [28] Molloy TJ, Bosma AJ, Baumbusch LO, et al. The prognostic significance of tumour cell detection in the peripheral blood versus the bone marrow in 733 early-stage breast cancer patients. Breast Cancer Res 2011;13(3):R61. [29] Pierga JY, Bonneton C, Vincent-Salomon A, et al. Clinical significance of immunocytochemical detection of tumor cells using digital microscopy in peripheral blood and bone marrow of breast cancer patients. Clin Cancer Res 2004;10(4):1392–400. [30] 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(5): 672–80.