ORIGINAL ARTICLEA decision aid for predicting non-sentinel node involvement in women with breast cancer and at least one positive sentinel node
Introduction
Several large validation studies have confirmed that breast cancer patients whose sentinel nodes (SN) are free of metastases have low rates of non-sentinel node (NSN) involvement of 5–10% in experienced hands. The high predictive value of a histologically negative SN has led some groups to abandon routine axillary clearance (AC) for this subset of patients.1, 2 By contrast, when the SN is involved by metastasis, only 30–50% of patients have further metastases to NSNs.3, 4, 5 This observation has prompted the question as to whether there are subsets among SN positive breast cancer patients with low risks of NSN involvement so that AC and its associated morbidity may be avoided.
A small number of studies have attempted to predict the status of the remaining axillary nodes, mostly by univariate analysis. The size of the metastasis in the SN, using the arbitrary figure of 2 mm, by convention defining nodal micrometastases, has received most attention.3, 4, 6 Some groups have suggested that patients with micrometastases may avoid AC, but since approximately 25% of patients with SN metastases ⩽2 mm5, 6 and 15.6% of those with ⩽1 mm SN metastases6 have involved NSNs, this simple heuristic alone is an unsatisfactory basis for avoidance of AC.
By analysis of clinical and pathological characteristics of a series of breast cancers with metastatic involvement to at least one SN, the purpose of this study is to identify independent predictors of the status of the remaining axillary nodes.
Section snippets
Materials and methods
All patients with primary breast cancer treated by SN biopsy and AC at the Royal Adelaide Hospital Breast Unit between June 1995 and September 2000 who had at least one positive SN were included in this study. The SNs were mapped by a combination of blue dye and lymphocintigraphy after peritumoural injection of technetium TC99 and blue dye. The details of the clinical and pathologic features of a subset of these tumours are reported elsewhere.7
The primary tumours were processed routinely. The
Statistical methods
Statistical analysis was done using the S-plus statistical software package.11 Univariate analysis assessed clinical and pathological variables for their association with the presence of any NSN metastases. The test was used for categorical variables. The Wilcoxon rank sum test was used for continuous variables.
For multivariate analysis a logistic regression was used. Using step-wise search procedures, a variety of models were tested for their ability to predict the likelihood of NSN
Results
Eighty-two patients with breast cancer metastatic to at least one SN were included in the study. The patients’ mean age was 54.9 years (S.D. 11.7 years). The mean tumour diameter was 25.6 mm (S.D. 13.8 mm). The mean number of SNs removed was 2.2 nodes (S.D. 1.3). The mean number of NSNs removed was 11.8 nodes (S.D. 5.2). NSN metastases were found in 38 cases (46.3%).
Significant patient and tumour characteristics
The results of univariate analyses for the parameters assessed are demonstrated in Table 1. A strong trend was found in younger women being more likely to have positive NSNs (). The effect of age was gradual without a threshold age associated with a dramatic change in risk (Table 2).
Screen-detected tumours with positive SN were less likely to have NSN metastases. While 23.53% of tumours detected by screening mammography had NSN metastases, 51.61% of patients with symptomatic tumours had
Multivariate analysis
Among all of the features assessed, the logistic regression model that predicted the status of the NSNs most successfully in this study selected patient age, proportion of SN replaced by metastasis and the total number of SNs submitted. Other variables did not improve the fit of the final model significantly and were not retained by the model.
Discussion
In experienced hands SNs can be identified in over 90% of breast cancer patients and histological examination of these nodes provides accurate staging information, particularly when the SNs are negative.13 International studies comparing SN biopsy to AC in women with early breast cancer are nearing completion. These include the NSABP32 trial in North America, the ALMANAC trial in the United Kingdom and the SNAC in Australia and New Zealand. It is likely that the results of these studies will
Conclusions
The international trials evaluating SN biopsy in breast cancer are nearing completion. They are expected to demonstrate that the majority of breast cancer patients may be staged effectively by SN biopsy. Since 50–70% of women with involved SNs do not have further axillary disease, the management of the axilla after SN biopsy in such women will then have to be assessed for each patient. Patients’ preferences are an important consideration in this discussion. Given the morbidity associated with
References (24)
- et al.
Sentinel lymph node biopsy as an indicator for axillary dissection in early breast cancer
Eur J Cancer
(2001) - et al.
Detection by screening mammography is a powerful independent predictor of survival in women diagnosed with breast cancer
Breast
(2004) - et al.
Sentinel-node biopsy to avoid axillary dissection in breast cancer with clinically negative lymph-nodes
Lancet
(1997) Sentinel lymph-node biopsy-based prediction of further breast cancer metastases in the axilla
Eur J Surg Oncol
(2001)- et al.
Predictors of nonsentinel lymph node metastasis in breast cancer patients
Am J Surg
(2002) - et al.
Predictors of positive axillary lymph nodes after sentinel lymph node biopsy in breast cancer
Am J Surg
(2001) - et al.
Prospective observational study of sentinel lymphadenectomy without further axillary dissection in patients with sentinel node-negative breast cancer
J Clin Oncol
(2000) - et al.
Do all patients with sentinel node metastasis from breast carcinoma need complete axillary node dissection?
Ann Surg
(1999) - et al.
Predictors on non-sentinel lymph node metastasis in breast carcinoma
Mod Pathol
(2001) - et al.
Pathologic features associated with nonsentinel lymph node metastases in patients with metastatic breast carcinoma in a sentinel lymph node
Cancer
(2000)
The size of micrometasases in axillary sentinel lymph nodes of breast cancer patients correlates with the occurrence of additional axillary metastases
Mod Pathol
Reliability of sentinel node status in predicting axillary lymph node involvement in breast cancer
Med J Aust
Cited by (47)
Completion axillary dissection can safely be omitted in screen detected breast cancer patients with micrometastases. A decade's experience from a single institution
2013, European Journal of Surgical OncologyCitation Excerpt :Unfortunately women with micrometastases were excluded from their study. Another study12 included 82 SN-positive patients and reported that the screen-detected patients had significantly fewer metastases in the ALND specimen (24%) compared with women with symptomatic tumours (52%) (p = 0.035). The authors noted that the method of detection was correlated with age because their national breast cancer screening programme primarily targeted women over 50 years of age, and did not therefore include the method of detection in the regression analysis.
A genetic algorithm model based on artificial neural network for prediction of the axillary lymph node status in breastcancer
2013, Engineering Applications of Artificial IntelligenceAssessing second echelon lymph nodes during sentinel lymph node biopsy: Can we have more accurate axillary treatment for breast cancer patients?
2011, Medical HypothesesCitation Excerpt :ALND, performed in patients with positive SLNs, significantly increased the risk of complications such as lymphedema, paresthesia or upper extremity dysfunction, due to the anatomic disruption. Actually, approximately 40–60% of patients with positive SLNs were clear of nodal involvement when ALND was performed [4–7]. These patients had undergone unnecessary ALND that lack of any therapeutic benefits.
Meta-analysis of predictive factors for non-sentinel lymph node metastases in breast cancer patients with a positive SLN
2011, European Journal of Surgical OncologyCitation Excerpt :This risk increased to 51% if the metastases size was >2 mm. The reported incidence of NSN tumour involvement varies from about 40–60%.9–40 Small study populations, the size of the tumours investigated and differences in the extent of histological work-up of SLN and NSN partly explain the range in prevalence of NSN involvement.14