Breast Cancer Research and Treatment

, Volume 131, Issue 2, pp 645–651

Prognostic impact of isolated tumor cells in breast cancer axillary nodes: single tumor cell(s) versus tumor cell cluster(s) and microanatomic location

Authors

  • Johanna H. Vestjens
    • Department of Internal Medicine, Division of Medical Oncology, GROW—School for Oncology and Developmental BiologyMaastricht University Medical Centre
  • Maaike de Boer
    • Department of Internal Medicine, Division of Medical Oncology, GROW—School for Oncology and Developmental BiologyMaastricht University Medical Centre
  • Paul J. van Diest
    • University Medical Centre Utrecht
  • Carolien H. van Deurzen
    • University Medical Centre Utrecht
  • Jos A. van Dijck
    • Radboud University Nijmegen Medical Centre
  • George F. Borm
    • Radboud University Nijmegen Medical Centre
  • Eddy M. Adang
    • Radboud University Nijmegen Medical Centre
  • Peter Bult
    • Radboud University Nijmegen Medical Centre
    • Department of Internal Medicine, Division of Medical Oncology, GROW—School for Oncology and Developmental BiologyMaastricht University Medical Centre
Epidemiology

DOI: 10.1007/s10549-011-1771-0

Cite this article as:
Vestjens, J.H., de Boer, M., van Diest, P.J. et al. Breast Cancer Res Treat (2012) 131: 645. doi:10.1007/s10549-011-1771-0

Abstract

In breast cancer, it has been shown that pN0(i+) and pN1mi have a comparable negative impact on disease-free survival, compared with pN0. However, pN0(i+) is considered to be a heterogeneous group. We determined the effect of metastatic size and microanatomic location within the pN0(i+) group on breast cancer recurrence. We included all Dutch breast cancer patients diagnosed in 1998–2005 with favorable primary tumor characteristics and a final nodal status of pN0(i+). For this analysis, only patients without adjuvant systemic therapy were eligible (n = 513). Presence of single tumor cells versus cell clusters, metastatic size and microanatomic location were recorded. Primary endpoint was disease-free survival. Analyses were adjusted for age at diagnosis, tumor size, tumor grade, axillary treatment and hormone receptor status. The 5-year disease-free survival of patients with single tumor cell(s) (n = 93) was 78.6% and with tumor cell cluster(s) (n = 404) 77.1%. The hazard ratio for disease events was 1.05 (95% CI 0.63–1.76) for cell cluster(s) compared with single cell(s). In a Cox regression model, doubling of metastatic tumor size corresponded to a hazard ratio of 1.21 (95% CI 1.02–1.43). The adjusted hazard ratio was 0.90 (95% CI 0.54–1.50) for parenchymal (n = 112) versus sinusoidal location (n = 395). Single tumor cells bear similar prognostic information as small tumor cell clusters, even though results do suggest that within the pN0(i+) group, increasing size of nodal involvement is associated with reduced survival. Microanatomic location does not seem to have prognostic relevance.

Keywords

Nodal isolated tumor cellsBreast cancer

Introduction

It is uniformly accepted that the axillary lymph node status is still one of the most important prognostic features in breast cancer. However, with new techniques and more aggressive treatment approaches, the landscape of primary breast cancer treatment has markedly changed, posing us for new dilemma’s.

With the introduction of the sentinel node (SN) procedure, the pathologic work-up has been intensified, resulting in an increased detection frequency of small nodal metastases [1]. The MIRROR study is the first large cohort study in the SN era with central pathology review and restaging according to the Sixth AJCC Cancer Staging Manual [2, 3]. Its results showed that the presence of isolated tumor cells, pN0(i+), as final nodal status, is an independent adverse prognostic factor as well as micrometastatic disease, pN1(mi), and that patients with pN0(i+) or pN1(mi) benefitted from adjuvant systemic therapy [2].

However, pN0(i+) is a heterogeneous group [4]. Obviously, isolated tumor cells can be the first sign of metastasis. However, it has also been hypothesized that in some patients, a few single tumor cells resemble benign or malignant cells that lack outgrowth potential, as can be found as a result of trauma [4, 5]. In melanoma and breast cancer studies, it has been shown that microanatomic location and penetrative depth of SN metastasis predict non-SN metastases [610]. Although the presence of pN0(i+) status was associated with reduced disease-free survival in the MIRROR study, part of this category may consist of patients with merely non-significant (benign or malignant) epithelial cells. Therefore, we determined the effect of extent (single cells vs. cell clusters), size and microanatomic location of nodal involvement on disease-free survival.

Methods

Patients

Women with invasive breast cancer, who had undergone an SN biopsy between 1998 and 2005, were identified from the Netherlands Cancer Registry. Only women with favorable primary tumor characteristics according to the guidelines at that time (i.e., tumors ≤1 cm in diameter, irrespective of grade, or tumors >1 to ≤3 cm, grade 1 or 2) were included. The methods and results were reported in detail by de Boer et al. [2]. For this analysis, only women with a final nodal status of pN0(i+), after an SN procedure with or without completion axillary lymph node dissection, and who had not undergone adjuvant systemic treatment, were selected.

Pathology

All available original slides from the SNs and of positive nodes obtained in the axillary lymph node dissection were reviewed by three pathologists (Carolien H. van Deurzen, Paul J. van Diest, Peter Bult) specialized in breast cancer. These included slides stained with hematoxylin and eosin, slides stained for immunohistochemical analysis and frozen sections. Almost all participating pathology laboratories used a protocol in which the SN was serially sectioned at at least every 150 μm and at a minimum of three levels, with the use of keratin immunohistochemical staining if the hematoxylin and eosin staining was negative. The non-SNs were macroscopically sectioned every 2–5 mm and one section per slice was stained with hematoxylin and eosin. During central review, presence of single tumor cell(s) versus tumor cell cluster(s) and the microanatomic location (sinus vs. parenchyma) were recorded. In addition, metastatic size was measured by an oculometer. Classification occurred according to the sixth edition of the AJCC Cancer Staging Manual [3].

Statistical analysis

The primary endpoint was disease-free survival, defined as the time from diagnosis to locoregional recurrence, distant metastases, contralateral breast cancer or ductal carcinoma in situ, other malignancy or death, whichever occurred first.

Hazard ratios for single cell(s) versus cluster(s), metastatic size and microanatomic location were estimated in multivariate models adjusted for age at diagnosis, tumor size, tumor grade, axillary treatment and hormone receptor status. In the analyses, we used a transformation and included log2(metastatic size) as independent variable, because the regression coefficient of this covariate then corresponds to the hazard ratio associated with a doubling of the size. P-values <0.05 were considered significant.

Results

Patients

In total, 513 patients were included. The median duration of follow up was 4.2 years (range 0.2–8.8).

Single tumor cell(s) versus tumor cell cluster(s)

For this analysis, 16 patients were excluded because of missing relevant data. The baseline characteristics according to single tumor cell(s) or tumor cell cluster(s) are shown in Table 1. In patients with single tumor cell(s), axillary lymph node dissection was less often performed. Also, axillary treatment by irradiation was less often performed in this group. The median metastatic size was larger in the tumor cell cluster(s) group. Single tumor cell(s) were found more often in lobular carcinoma. The 5-year disease-free survival in patients with single tumor cell(s) (n = 93) was 78.6% and 77.1% in patients with tumor cell cluster(s) (n = 404). The difference was not significant (P = 0.58) (Fig. 1). In a multivariate analysis, including age at diagnosis, tumor size, tumor grade, axillary treatment and hormone receptor status, the hazard ratio of tumor cell cluster(s) compared with single cell(s) was 1.05 (95% CI 0.63–1.76).
Table 1

Baseline clinicopathologic characteristics of 497 study patients with invasive breast cancer with micrometastatic axillary lymph node disease

Characteristics

All (n = 497)

Single cell(s) (n = 93)

Cell cluster(s) (n = 404)

P value

Age at diagnosis (yrs)

   

0.59

 Median

57

58

57

 

 Range

34–90

36–89

34–90

 

Tumor size-no. (%)

   

0.98

 ≤2 cm

429 (87.0)

81 (87.1)

348 (87.0)

 

 >2 to ≤3 cm

64 (13.0)

12 (12.9)

52 (13.0)

 

Tumor grade-no. (%)

   

0.84

 1

188 (37.8)

37 (39.8)

151 (37.4)

 

 2

281 (56.5)

50 (53.8)

231 (57.2)

 

 3

19 (3.8)

4 (4.3)

15 (3.7)

 

 Unknown

9 (1.8)

2 (2.2)

7 (1.7)

 

Tumor type-no. (%)

   

<0.001

 Ductal

362 (72.8)

53 (57.0)

309 (76.5)

 

 Lobular

71 (14.5)

23 (24.7)

48 (11.9)

 

 Other

64 (13.1)

17 (18.3)

47 (11.6)

 

Expression of estrogen receptors, progesterone or both-no. (%)

   

0.99

 Yes

443 (89.1)

80 (86.0)

363 (89.9)

 

 No

39 (7.8)

7 (7.5)

32 (7.9)

 

 Unknown

15 (3.0)

6 (6.5)

9 (2.2)

 

Type of surgery-no. (%)

   

0.066

 Breast-conserving

354 (71.2)

59 (63.4)

295 (73.0)

 

 Mastectomy

143 (28.7)

34 (36.6)

109 (27.0)

 

No. of SN removed

   

0.14

 Median

2

2

2

 

 Range

1.0–14.0

1.0–14.0

1.0–13.0

 

Metastatic size (mm)

   

<0.001

 Median

0.06

0.01

0.08

 

 Range

0.01–0.20

0.01–0.04

0.01–0.20

 

Axillary lymph node dissection-no. (%)

   

<0.001

 No

278 (55.9)

68 (73.1)

210 (52.0)

 

 Yes

219 (44.1)

25 (26.9)

194 (48.0)

 

Axillary lymph-node dissection, axillary irradiation or both-no. (%)

   

<0.001

 No

252 (50.7)

64 (68.8)

188 (46.5)

 

 Yes

245 (49.3)

29 (31.2)

216 (53.5)

 
https://static-content.springer.com/image/art%3A10.1007%2Fs10549-011-1771-0/MediaObjects/10549_2011_1771_Fig1_HTML.gif
Fig. 1

Disease-free survival according to disease extent, within the group of patients classified as pN0(i+)

Metastatic size

When including log2 metastatic size within the pN0(i+) group as a continuous variable in a Cox regression model, increasing metastatic size was significantly associated with a reduced disease-free survival with an adjusted hazard ratio of 1.21 (95% CI 1.00–1.61) for doubling of metastatic size.

Micro-anatomic location of SN metastases

Regarding the influence of microanatomic location of SN-metastasis, six patients were excluded because of missing relevant data. The baseline characteristics are shown in Table 2. The 5-year disease-free survival was 77.6% for patients with tumor cells located only in the sinus of the SN (n = 395), and 75.3% for patients with isolated tumor cells in the parenchyma of the SN (n = 112) (P = 0.98, Fig. 2). In multivariate analysis, the adjusted hazard ratio of parenchymal location compared with sinusoidal was 0.90 (95% CI 0.54–1.50).
Table 2

Comparison of clinicopathologic features between patients with different microanatomic location of isolated tumor cells

Characteristics

All

Microanatomic location

P value

Sinus (n = 395)

Parenchyma (n = 112)

Age at diagnosis (year)

   

0.34

 Median

57

57

57

 

 Range

34–93

35–93

34–89

 

Tumor size-no. (%)

   

0.67

 ≤2cm

434 (86.3)

336 (85.9)

98 (87.5)

 

 >2 to ≤3 cm

69 (13.7)

55 (14.1)

14 (12.5)

 

Tumor grade-no. (%)

   

0.81

 1

192 (37.9)

152 (38.5)

40 (35.7)

 

 2

287 (56.6)

222 (56.2)

65 (58.0)

 

 3

19 (3.7)

14 (3.5)

5 (4.5)

 

 Unknown

9 (1.8)

7 (1.8)

2 (1.8)

 

Tumor type-no. (%)

   

0.33

 Ductal

370 (73.0)

293 (74.2)

77 (68.7)

 

 Lobular

73 (14.4)

52 (13.2)

21 (18.7)

 

 Other

64 (12.6)

50 (12.7)

14 (12.5)

 

Expression of estrogen receptors, prog both-no. (%)

   

0.75

 Yes

451 (88.9)

351 (88.9)

100 (89.3)

 

 No

40 (7.9)

32 (8.1)

8 (7.1)

 

 Unknown

16 (3.2)

12 (3.0)

4 (3.6)

 

Type of surgery-no. (%)

   

0.17

 Breast-conserving

361 (71.2)

287 (72.7)

74 (66.1)

 

 Mastectomy

146 (28.8)

108 (27.3)

38 (33.9)

 

No of SN removed-no.

   

0.33

 Median

2

2

2

 

 Range

1–14

1–14

1–9

 

Metastatic size (mm)

   

0.22

 Median

0.06

0.06

0.05

 

 Range

0.01–0.20

0.01–0.20

0.01–0.20

 

Axillary lymph node dissection-no. (%)

   

0.26

 No

286 (56.4)

228 (57.7)

58 (51.8)

 

 Yes

221 (43.6)

167 (42.3)

54 (48.2)

 

Axillary lymph-node dissection, axillary irradiation or both-no. (%)

   

0.18

 No

259 (51.1)

208 (52.7)

51 (54.5)

 

 Yes

248 (48.9)

187 (47.3)

61 (45.5)

 
https://static-content.springer.com/image/art%3A10.1007%2Fs10549-011-1771-0/MediaObjects/10549_2011_1771_Fig2_HTML.gif
Fig. 2

Disease-free survival according to microanatomic location, within the group of patients classified as pN0(i+)

Discussion

Previously, we reported that in breast cancer patients with favorable primary tumor characteristics, pN0(i+) and pN1mi had a comparable negative impact on disease-free survival, compared with pN0 [2]. However, pN0(i+) can be regarded as a heterogeneous group, including both patients with single tumor cells and patients with tumor cell clusters, which may either be located in the sinus or in the parenchyma of the nodes. We, therefore, questioned whether there would be a differential impact on prognosis of size and microanatomic location of nodal metastases within the group of patients with pN0(i+) as final nodal status. In this study, we show that within the pN0(i+) group, patients with single tumor cells had a 5-year disease-free survival of 78.6%, not different from the 5-year disease-free survival of 77.1% for patients with tumor cell clusters. On the other hand, we noticed that increasing metastatic size as a continuous variable was associated with a reduced disease-free survival at a hazard ratio of 1.21 (95% CI 1.02–1.43), yet indicating a small impact of size within the pN0(i+) group. Notably, the microanatomic location did not influence prognosis (adjusted hazard ratio 0.90; 95% CI 0.54–1.50).

The prognostic relevance of pN0(i+) has been studied before. Five small studies including 21–61 patients from the pre-SN era examined the relevance of isolated tumor cells found after step sectioning and immunohistochemical staining in lymph nodes defined as negative after hematoxylin and eosin staining [1115]. In three of these, no significant adverse prognosis in the pN0(i+) group compared with the pN0 group was seen. However, in two of these studies, no adverse prognosis even for larger occult metastases was found at all, probably because of small number of patients evaluated [11]. One study reported an unadjusted hazard ratio for adverse events of 2.51 (95% CI 1.55–4.07) in the group of pN0(i+) (n = 24) as compared with pN0, which was also seen for patients having pN1mi [14]. The adjusted hazard ratio for pN0(i+) and pN1mi combined versus pN0 was 2.16 (95% CI 1.42–3.28). In another study, the adverse prognostic significance of pN0(i+) (relative risk 1.7, 95% CI 1.0–2.7) was comparable to or larger than primary tumor size, histological grade and lymphovascular invasion [15].

The prognostic effect of isolated tumor cells has also been analyzed in 11 SN studies, but in 9 of these, only 17–151 patients with pN0(i+) were included [2, 1625]. Only four studies corrected for use of adjuvant systemic therapy in multivariate analysis [2, 19, 21]. In eight studies, the presence of isolated tumor cells did not seem to influence prognosis [1619, 2124]. In one of these, however, overall survival was lower in patients who had not undergone an additional axillary lymph node dissection as compared with patients who had undergone an axillary lymph node dissection (P = 0.02) [17]. On the other hand, in three studies, including the two by far largest studies, the presence of isolated tumor cells did have a negative impact on breast cancer outcome [2, 20, 25]. In one study (n = 63), the 5-year breast cancer–specific survival (95.7% vs. 98.4%, P = 0.035) was decreased and the 5-year rate of distant metastasis (8.1% vs. 1.9%, P = 0.001) was increased for patients with pN0(i+) as compared with pN0 patients, even though adjuvant systemic therapy was significantly more often administered to patients with pN0(i+) as final status [20]. The MIRROR study not only included a large group of patients (n = 2,707) but also analyzed patients without adjuvant systemic therapy [2]. The 5-year disease-free survival was significantly higher (85.7 vs. 77.2%) in patients with final stage pN0 as compared with pN0(i+) patients, all without adjuvant systemic therapy. Separate analyses showed that patients with pN0(i+) who had received adjuvant systemic therapy had an improved survival as compared with those who did not receive adjuvant systemic therapy (83.0 vs. 77.2%). Similar results were seen in a recently published large prospective study including 431 patients with isolated tumor cells [25]. The hazard ratio of isolated tumor cell clusters versus no tumor cell clusters was 1.27 (95% CI 1.04–1.54) for death.

It has been hypothesized that single tumor cells not necessarily represent metastatic disease, but merely displaced tumor cells that lack outgrowth potential, or even benign epithelial cells [2632]. The first hypothesis presented is traumatic displacement of cells. It has been shown that after primary tumor biopsy, tumor cells were located in the needle tract as well as in the stroma [26, 27]. It has also been shown that epithelial cells were found more often in the SN after biopsy or breast massage [2830]. This hypothesis was strengthened by finding epithelial cells accompanied by hemosiderin-laden macrophages and damaged erythrocytes suggesting traumatic displacement [31]. A second hypothesis could be the inadequate use of cytokeratin immunohistochemistry with staining of nonepithelial cells [33, 34]. A third hypothesis is the use of tumor cell–contaminated instruments in sectioning the SNs. Furthermore, nodal entrapment of epithelial remains during embryonic development resulting in ectopic breast tissue has been suggested a possible cause for finding of pN0(i+) [32]. Morphometric and immunohistochemical studies indeed presented clues that some of the cells may not be (vital) malignant cells [4, 5].

Hypothetically, breast cancer cells arrive through the afferent lymph vessel in the subcapsular sinus. The cells outgrow in the marginal sinus and into the cortical parenchyma. After extending the deeper parenchyma into the medullary sinus, further routing takes place through the efferent lymph vessel [35]. Exploring this hypothesis, microanatomic location of tumor cells could be an indicator of relevant metastatic disease as studied in melanoma [69]. Results of breast cancer studies on microanatomic location are conflicting. In three studies, including 45–167 patients with isolated tumor cells and/or micrometastases, sinusoidal deposits, in contrast to another location, were shown to be a significant adverse prognostic factor [13, 36, 37]. However, in five other studies, including 50–136 patients, the opposite was found [10, 11, 3840]. All these studies are heterogeneous with respect to (small) size and extent of metastatic disease. Our study shows that the hazard ratio of events was comparable for sinusoidal and parenchymal location.

Another discriminative feature could be the number of tumor cells. Isolated tumor cells can either represent truly single isolated cells or cell clusters. It can be hypothesized that the presence of only a few tumor cells might less severely affect the outcome than the presence of many tumor cells. Indeed, old studies have shown that the number of positive lymph nodes is associated with outcome, in essence also representing a measurement of nodal size involvement [4143]. In our study, we showed that even within a small range of metastatic size, the number of tumor cells does matter as the prognosis declines with increasing number of cells. But of note, we have also shown that single tumor cell(s) bear similar prognostic information as tumor cell cluster(s). Notably, in our study, in the group of patients with single tumor cells, less frequently axillary treatment was performed. However, adjusting for axillary treatment in multivariate analysis did not change the hazard ratio.

In daily practise, the decision to administer adjuvant systemic therapy is obviously not solely based on the lymph node involvement, because other risk factors co-influence this decision. It has been hypothesized that micro-arrays such as the Mammaprint® or Oncotype DX® assessing primary tumor characteristics might replace the information by nodal status, but so far scientific evidence shows that nodal involvement remains an independent prognosticator [44].

In conclusion, microanatomic location does not seem to influence prognosis of nodal status classified as pN0(i+). Single tumor cells bear similar prognostic information as tumor cell clusters, although results suggest a small impact of size within the pN0(i+) group when analyzed as a continuous variable.

Acknowledgments

We thank Wim A.J.G. Lemmens for his assistance with statistical analyses. This project has received funding from the Netherlands Organization for Health Research and Development (ZonMw 945-06-509) and supported by the Breast Research Group of The Netherlands (BOOG)

Conflict of interest

None.

Copyright information

© Springer Science+Business Media, LLC. 2011