| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Article |
1 Department of Surgery, Radboud University Nijmegen Medical Center (MC), Nijmegen, The Netherlands
2 Department of Epidemiology and Biostatistics, Radboud University Nijmegen, Nijmegen, The Netherlands
3 Department of Pathology, Radboud University Nijmegen MC, Nijmegen, The Netherlands
4 Department of Pathology, Canisius-Wilhelmina Hospital, Nijmegen, The Netherlands
5 Department of Pathology, Viecuri MC, Venlo, The Netherlands
6 Department of Pathology, Rijnstate Hospital, Arnhem, The Netherlands
7 Department of Surgery, Canisius-Wilhelmina Hospital, Nijmegen, The Netherlands
8 Department of Surgery, Viecuri MC, Venlo, The Netherlands
9 Department of Surgery, Rijnstate Hospital, Arnhem, The Netherlands
10 Department of Medical Oncology, Radboud University Nijmegen MC, Nijmegen, The Netherlands
11 Department of Medical Oncology, University Hospital Maastricht, P.O. Box 5800, NL-6202 AZ, Maastricht, The Netherlands
Correspondence: Address correspondence and reprint requests to: V. C. G. Tjan-Heijnen, MD, PhD; E-mail: V.Tjan{at}sint.azm.nl
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Methods: In this multi-institutional study, 541 eligible breast cancer patients were included prospectively.
Results: The occurrence of non-SN metastases was related to the size of the SN metastasis (P = .02), primary tumor size (P = .001), and lymphovascular invasion (P = .07). The adjusted odds ratio was 3.1 for SN micro-metastasis compared with SN isolated tumor cells, 4.0 for SN macro-metastasis versus SN isolated tumor cells, 3.1 for tumor size (>3.0 cm compared with 
3.0 cm), and 2.0 for lymphovascular invasion (yes versus no). There were no positive non-SNs when the primary tumor size was 1.0 cm (n = 24) [95% confidence interval (95% CI) 0%–14.0%]. The proportion of positive non-SNs ranged in a prognostic logistic regression model from 9.7% (95% CI 4.0%–23.0%) for patients with SN isolated tumor cells, tumor size of 1.1–3.0 cm, and without vessel invasion, to 72.6% (95% CI 47.0%–89.0%) for patients with SN macro-metastasis, tumor size >3.0 cm, and with vessel invasion.
Conclusion: We identified three predictive factors for non-SN metastases in breast cancer patients with a positive SN: size of the SN metastasis; primary tumor size; and vessel invasion. We were not able to identify a specific group of patients with a positive SN in whom the risk for non-SN metastases was less than 5%.
Key Words: Breast cancer • Sentinel lymph node • Micro-metastasis • Non-sentinel lymph node metastasis • Risk factors
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The SN is the first lymph node(s) in which the primary tumor drains. In cases where the SN shows tumor involvement, a completion ALND will still be performed. However, in patients with a negative SN, completion ALND can be avoided, as, in that situation, the incidence of non-SN metastases is very low.1 According to the Netherlands Cancer Registry, approximately 60% of breast cancer patients had node-negative disease before the introduction of the SN biopsy.2 Hence, it was expected that in these 60% of patients with resectable breast cancer, a completion ALND could be avoided by carrying out a SN biopsy.
Obviously, a reliable examination of the SN by the pathologist is crucial, as a false-negative finding may result in under-treatment. Consequently, pathologists have intensified the examination of the SN by using serial sectioning (SS) and immunohistochemistry (IHC), whereas previously, the axillary lymph nodes were examined by hematoxylin–eosin (H&E) in one or two slides only. However, intensified examination of the SN may result in increased detection of isolated tumor cells and micro-metastases, the clinical significance of which is unclear.3
We postulated that the SN biopsy leads to an increased detection of isolated tumor cells and micro-metastases due to the intensified work-up of the SN by the pathologist. If, upon the detection of isolated tumor cells and micro-metastases a completion ALND is performed, this could partially offset the expected reduction in the rate of ALND, while, in fact, it is uncertain whether a SN with isolated tumor cells or micro-metastases justifies a completion ALND.
In summary then, we hypothesized that the incidence of non-sentinel lymph node (non-SN) metastases would be lower in patients with isolated tumor cells or micro-metastases in the SN versus those with macro-metastases and, that in cases where the incidence of non-SN metastases would be 5% or less in specific subgroups of patients, a completion ALND would not likely justified.
|
PATIENTS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The prospectively collected data included the lymph node status and number of nodes examined, number of positive nodes, size of metastases, classification according to the tumor node metastasis (TNM) categories defined in the sixth edition of the TNM Classification of Malignant Tumors4, and the detection method (H&E/IHC). These items were separately registered for SNs and non-SNs. Also, details of primary tumor characteristics (localization, tumor size, histology, histological grade, lymph and/or blood vessel invasion, hormone receptor status), patient characteristics (age) and information on the surgical procedure (SN biopsy with or without ALND, lumpectomy or mastectomy and various combinations) were collected.
The surgical procedure and the pathological examination were in accordance with the Dutch guidelines for treatment of breast cancer.5 SN localization was performed using the combined technique of blue dye and radioisotope in all patients. At least three levels at, at least, 150 µm intervals were examined with H&E. In the absence of apparent metastases with H&E examination, IHC examination was performed. In the presence of isolated tumor cells, micro-, or macro-metastases in the SN, a completion ALND was recommended. The nodes in the ALND specimen were examined at one to two levels with H&E staining.
According to the international TNM classification 2002, isolated tumor cells, micro-metastases, and macro-metastases were classified as follows: isolated tumor cells [pN0(i+)] are defined as solitary tumor cells or tumor cell clusters with a size of .2 mm or less. Micro-metastases (pN1mi) are more than .2 mm and maximally 2.0 mm in size. Macro-metastases are >2.0 mm in size. For the SN findings, "sn" was added between brackets [pN(sn)]. In this present paper we added pN1+, which refers to pN1a and higher pN positive stages. Further, we added the term "pNtotal", which refers to the final pTNM stage including both the SN and, if applicable, the non-SNs findings.
Statistical Analyses
We tried to identify a subgroup of breast cancer patients in whom the incidence of metastatic disease in the non-SNs (obtained after completion ALND) had to be reliably predicted to be 5 % or less. In such a group we considered that omitting completion ALND was justified.
The following variables were explored for prognostic significance with respect to occurrence of non-SN metastases in patients in whom non-SNs were removed: age (
50 years versus <50 years), SN findings (isolated tumor cells versus micro-metastases versus macro-metastases), tumor size (<1 cm, 1.1–2 cm, etc., up to >5 cm), histological grade (I versus II versus III), hormone-receptor status (ER-positive and/or PgR-positive versus both negative) and lymph and/or blood vessel invasion (yes or no). Those variables yielding a P value of less than .10 for the chi-square test were incorporated in a multiple logistic regression model. Subgroups were formed, based on combinations of these variables. For each subgroup the model-based predicted probability for non-SN metastases was compared with the proportion of women who actually had positive non-SN nodes. Goodness of fit was assessed by the deviance statistic. The discrimination of the model was measured by the area under the receiver operating characteristic (ROC) curve.
The strength of a risk factor for non-SN metastases was expressed by the odds ratio.
The relation between pathological tumor size and nodal status was also assessed with a logistic regression model.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
.
|
).
|
).
|
SN and Final Nodal Classification According to TNM
In 54 out of 541 (10.0%) patients the SN contained isolated tumor cells [pN0(i+)(sn)]; in 53 (9.8%) patients the SN contained micro-metastases [pN1mi(sn)] and, in 96 (17.7%) patients, macro-metastases [pN1+(sn)] (Fig. 3).
|
). The final nodal stage was classified as node-negative in 335 (61.9%) patients. Forty-seven (8.7%) patients had isolated tumor cells, 49 (9.1%) patients had micro-metastases, and 110 (20.3%) patients had macro-metastases (Table 2).
|
).
|
Tumor size was also significantly associated with the incidence of positive non-SNs. None of 24 patients with a tumor size of 1.0 cm or smaller had non-SN metastases. Twenty-seven percent (27.3%) of patients with a tumor size 1.1–2 cm had a positive non-SN compared with 30.4% with a tumor size 2.1–3 cm, compared with 56.2% with a tumor size 3.1–4 cm, compared with 42.9% with a tumor size 4.1–5 cm, and compared with 66.7% with a tumor size >5 cm (chi-square test, P = .001).
In addition, patients without lymph and/or blood vessel invasion had an incidence of 25.6% of positive non-SNs compared with an incidence of 38.3% in patients with vessel invasion (chi-square test, P = .07) (Table 3).
|
).
Probabilities for Non-SN Metastases
There were no positive non-SNs when the primary tumor size was <1.0 cm (n = 24). However, because of the small number of observations, the confidence interval (CI) was quite large and included the pre-defined upper tolerable limit of 5% (95% CI 0%–14%). Of these 24 patients, 12 had solely isolated tumor cells in the SN of whom 11 had no lymph and/or blood vessel invasion. Six patients had micro-metastasis in the SN, and six patients had macro-metastasis in the SN.
For tumors >1 cm the probability for positive non-SNs was determined with a prognostic logistic regression model including pN(sn), tumor size and lymph and/or blood vessel invasion. Subgroups were formed, based on the combinations of these variables. The categories for tumor size 1.1–2 cm and 2.1–3 cm were joined, because there was hardly any difference in predicted probability of positive non-SNs (less than .8%) between these two categories for tumor size. Tumors >5 cm (n = 6) were excluded, because the fit of the model improved considerably from P = .05 to P = .18 for the deviance statistic. The discrimination of the model, measured by the area under the ROC curve, was .67. For each subgroup the model-based predicted probability for non-SN metastases was compared with the proportion of patients who actually had positive non-SNs.
In patients with tumor size 1.1–3 cm and no lymph and/or blood vessel invasion, the incidence of positive non-SNs was 7.1% in the case of isolated tumor cells in the SN, 29.6% in the case of micro-metastasis in the SN and 30.0% in the case of macro-metastasis in the SN. The predicted probability of positive non-SNs according to our model of these three groups was, for isolated tumor cells in the SN, 9.7% (95% CI 4%–23%), for micro-metastasis in the SN, 25.0% (95% CI 14%–41%), and for macro-metastasis in the SN, 30.0% (95% CI 20%–42%). See Table 4 and Fig. 5 for further incidences.
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
In this prospective study, 541 patients with primary breast cancer who successfully underwent a SN biopsy were included. In agreement with others, we demonstrated that the SN biopsy is an excellent tool to make a first selection for the omission of a completion ALND, with 62.5% of patients having a negative SN. These were all patients with clinically T1 or T2 tumors. Of interest, of 186 patients with a positive SN who subsequently underwent a completion ALND, only 30% had one or more positive non-SN. According to a prior meta-analysis, approximately 50% of patients who had a positive SN would be expected to have residual disease in the axilla.1 This striking difference may be explained by the currently accepted intensified pathology protocol, with the detection of more and smaller metastases in the SN. Indeed, in nearly 50% of our patients with a positive SN, the SN contained only low-volume metastasis, i.e., isolated tumor cells or micro-metastasis. This may support our initial hypothesis, that some of the patients having small SN metastasis might not benefit from a completion ALND.
In a recent meta-analysis of 25 publications, it was concluded that the risk of non-SN metastases with low-volume metastasis in the SN is around 10–15%.3 This meta-analysis was not based on individual patient data, and, therefore, as not all studies contained details of the subgroup categorization, the investigators were not able to perform any other quantitative analyses that might have predicted the occurrence or absence of metastases in the non-SNs.
In our study, we observed that 14.6% of patients with isolated tumor cells in the SN did have non-SN involvement and 28.6% of patients with micro-metastasis, compared to 38.0% of patients with macro-metastasis in the SN. Obviously, patient selection (T1 versus T1/2 tumors) and differences in pathology protocols may account for some of the differences between reported detection rates.
In this present study, we demonstrated that, apart from the size of SN metastasis, the primary tumor size and the presence or absence of lymph and/or blood vessel invasion were also associated with the occurrence of metastases in the non-SNs. However, no single variable predicted non-SN metastases with sufficient accuracy that ALND might safely be omitted. Others also reported similar associations, but only few performed additional analyses on whether a combination of factors would better distinguish between high-risk and low-risk categories.6–10
In two small studies the combination of two risk factors was analyzed, showing that none of the patients who had small SN metastases (<2 mm, including isolated tumor cells) in combination with a small primary tumor (<2 cm in diameter) did have non-SN metastases.11,12 In a large study by Weiser et al., none of 24 patients with three predictive factors (tumor size <1 cm, SN with isolated tumor cells or micro-metastasis, and absence of lymphovascular invasion) had non-SN metastases, whereas 58% of patients with none of the favorable factors had disease in the non-SNs.13 Unfortunately, 95% CIs were not reported, so it remains unclear whether upfront decision making based on these three factors is fully reliable. In this latter study, they reported that selection by the two aforementioned variables only was not possible, as 26% of patients with favorable SN metastasis and tumor size had non-SN metastases.
As shown in Table 4 and Fig. 5, one extra risk factor multiplied the percentage of the observed as well as the model-based predicted proportion of positive non-SNs, on average, by three times. The OR for SN micro-metastases compared to SN isolated tumor cells was 3.1, and, for SN macro-metastases versus SN isolated tumor cells, it was 4.0. The OR for tumors larger than 3 cm compared with tumors smaller than 3 cm was 3.1, and the OR for lymph and/or blood vessel invasion (presence versus absence) was 2.0.
However, despite a significant association between the presence of non-SN metastases and size of SN metastases, primary tumor size and lymph and/or blood vessel invasion, we were not able to identify a specific group of patients with a "positive" SN in whom the incidence of positive non-SNs could be reliably predicted to be less than 5%. So, although we observed that the risk of positive non-SNs decreased with reduced number of risk factors, the predicted proportion of patients with positive non-SNs was still 9.7% for patients who had as the only risk factor the presence of isolated tumor cells. It can be hypothesized that, with a larger study population, the confidence intervals may become smaller, supporting in due time the concept that completion ALND may safely be omitted in patients with isolated tumor cells in the SN who have small primary tumors (possibly <1.0 cm) without lymph and/or blood vessel invasion, but, definite proof could not be given in this present study, which included 541 eligible patients. Of note, the number of patients who might theoretically be good candidates for the omission of completion ALND was estimated to be 5.0% (11 out of 203) of patients with a positive SN.
In our study, we hypothesized that an expected detection frequency of non-SN metastases of 5% or less would not justify a completion ALND. The limit of 5% is arbitrary, and, in fact, is based on a consensus among Dutch physicians that the benefit of treatment should ideally involve at least 5% of patients.5 However, a 5% risk assessment may be too conservative, and something closer to 10% might be more reasonable for some patients. With that in mind, the data presented in this paper suggest that patients with tumors <1 cm, and patients with 1.1–3.0 cm tumors, no lymph and/or blood vessel invasion, and only isolated tumor cells in the SN could be considered for the omission of completion ALND, since the predicted risk of non-SN involvement is 9.7% and the observed risk was 7.1%. However, as the confidence interval is still a little wide (4%–23%), due to the small sample size, further data are eagerly awaited before definite recommendations can be made. Obviously, the proof of principle is made on axillary relapse rates and on overall survival values. In breast cancer it may require considerable time before small metastases left behind become clinically manifest as regional recurrences or the source of distant metastases. A study of SNs in 243 melanoma patients had previously shown that early nodal recurrence after negative SN findings could be explained by micro-metastases being overlooked at first analysis.14
An ongoing Milanese trial (IBCSG 2301) randomly allocates breast cancer patients with micro-metastatic SNs to completion ALND or surveillance. An American study (NSABP-32) compares SN resection with conventional ALND in clinically node-negative breast cancer patients. These trials may provide greater clinical evidence for the formulation of policies on axillary sparing after a positive SN biopsy.
To this end, we conclude that completion ALND should still be recommended to all patients with a "positive" SN, i.e., with isolated tumor cells, micro-metastases or macro-metastases, as we were not able to identify a specific group of patients with favorable primary tumor and SN characteristics in whom the incidence of positive non-SNs would reliably be predicted to be less than 5%.
|
ACKNOWLEDGMENTS |
|---|
Received for publication June 29, 2006. Accepted for publication July 3, 2006.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  C. H. M. van Deurzen, M. de Boer, E. M. Monninkhof, P. Bult, E. van der Wall, V. C. G. Tjan-Heijnen, and P. J. van Diest Non-Sentinel Lymph Node Metastases Associated With Isolated Breast Cancer Cells in the Sentinel Node J Natl Cancer Inst, November 19, 2008; 100(22): 1574 - 1580. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. G. Van den Eynden, M. K. Vandenberghe, P.-J. H. van Dam, C. G. Colpaert, P. van Dam, L. Y. Dirix, P. B. Vermeulen, and E. A. Van Marck Increased Sentinel Lymph Node Lymphangiogenesis is Associated with Nonsentinel Axillary Lymph Node Involvement in Breast Cancer Patients with a Positive Sentinel Node Clin. Cancer Res., September 15, 2007; 13(18): 5391 - 5397. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
