Sentinel lymph node biopsy technique and multifocal breast cancer – The Aarhus experience

Staging of the axilla in breast cancer is essential because lymph node status is the most important prognostic factor [1]. On patients with unifocal breast cancer the sentinel lymph node biopsy technique (SLNB) is a reliable tool to detect lymph node metastases and perhaps even more importantly to identify patients without metastases [2–6]. Patients without metastases can avoid a total axillary dissection and the related complications [7–12]. The method has rapidly been implemented as standard procedure despite there being no documented evidence of a beneficial effect on overall survival and relapse rate. However, trials with the aim of investigating these important issues are ongoing [13–15].

Compared with the numerous studies of the SLNB on patients with unifocal breast cancer relatively few validation studies have been made on patients having multifocal or multicentric tumors (Table I). Unfortunately, the published works are not comparable. This is mainly because different definitions of multifocal and multicentric breast cancer are used, but also that the inclusion criteria used in the studies are not uniform. Few of the studies have been performed as prospective studies, and with the only aim to investigate the methods reliability. Despite these facts, it is now generally accepted, that the method works among patients with more than one tumor in the same breast [16].

Table

Author	Published	N	MC/MF	Detection rate	False negative rate
Mertz [37]	1999	16	MC + MF	98	0
Schrenk [38]	2001	19	MC	100	0
Kim [39]	2002	5	MC	100	0
Fernandez [40]	2002	53	MC + MF	98	0
Ozmen [41]	2002	21	MF	86	11
Kumar [42]	2003	59	MC + MF	93	0
Tousimis [32]	2003	70	MC + MF	96	8
Kumar [43]	2004	10	MC + MF	100	0
Goyal [33]	2004	75	MF	95	9
Bergkvist [34]	2005	56	MF + MC	95	21

Summary of published validation studies on multifocal and multicentric breast cancer.

MF: multifocal breast cancer; MC: multicentric breast cancer.

The aim of the present study was to validate the use of SLNB in patients, who preoperatively had been diagnosed with more than one tumor on either mammography and/or ultrasonography.

Materials and methods:

In the present study we defined a patient as having more than one tumor in the breast if there was more than 20 mm of normal breast tissue between two malignant tumors on the mammogram or ultrasound. This was the definition used by the Danish Breast Cancer Cooperative Group (DBCG) until 2006 [17], and these patients are defined as having multifocal breast cancer in the text to follow.

The inclusion period ranged from August 2001 until October 2004. All patients with multifocal breast cancer according to DBCG′s definition were included in the study.

Excluded were:

1. Patients with preoperatively detected metastases to axillary lymph nodes;

2. Patients with a dominant tumor size larger than 4 cm;

3. Patients who had other contraindications to SLNB (Table II).

In Table II overview of the total study population and excluded patients are shown. In Table III, basic characteristics of the patients who underwent SLNB and are shown.

Table

Total number of patients with multifocal, early breast cancer	59
Excluded patients
Preoperatively diagnosis of lymph node metastases	14
Dominant tumor size larger than 40 mm	10
Other reasons*	6
Declined participation	4
Not asked for participation	1
Included patients	24

Summary of all patients with multifocal, early breast cancer during the inclusion period.

*Prior operation in the axilla, patients with malignant micro calcifications, earlier diagnosed lymphomas.

Table

Included patients	24
Age
Median	53 years (30–78)
Localisation of dominant tumor size
Upper lateral quadrant	16
Lower lateral quadrant	1
Upper medial quadrant	4
Lower medial quadrant	0
Central	3
Lymph node status
No lymph node metastases	8
Lymph node metastases	16
Distribution on T-stage
T1b	1
T1c	10
T2	13
Total tumor burden
Median number of tumors	2 (max. 3)
Median total tumor size	28 mm (range 10–70)
Histological types
Ductal	21
Lobular	2
Mucinous	1
Malignancy grade
Grade 1	11
Grade 2	10
Grade 3	2
Not graded	1
Estrogen receptor status
Positive	21
Negative	3

Demographic characteristics of the included patients.

All patients had an ultrasound of the axilla performed as part of the normal diagnostic workup. If the radiologist suspected that malignant lymph nodes were present fine needle aspiration biopsies (FNAB) were performed on the suspected nodes to confirm the findings. If malignant cells were found from lymph nodes the patient was excluded from the study. Two to 4 h before the operation a Technetium-99 labelled colloid (ALBU-RES, Sorin Biomedica Diagnostics, Milan, Italy or NanoColl), 15 Mbq, was injected peritumorally around the largest tumor. Immediately before the operation (15–20 min) Patent Blue V (Laboratoire Guerbert, Aulney-sous-Bois, France) was also injected peritumorally around the largest tumor. A gammaprobe (C-track or a Neoprobe 2000) was used to detect the sentinel lymph nodes. Lymphoscintigraphies were not performed.

Sentinel lymph nodes were defined as any blue and/or hot lymph nodes with a specimen count above 10 Bq. After removal of the sentinel lymph nodes a level I and II axillary dissection was performed. All surgical procedures were performed by either consultants or senior trainees and the pathological examination were performed by an experienced pathologist (consultant). Serial sectioning of the sentinel lymph nodes was performed for each 250 m and stained with haematoxylin-eosin staining (HE) and with cytokeratin. Non-sentinel lymph nodes were divided and stained with HE. All breast specimens were reevaluated by the pathologist who confirmed the multifocality of the tumors.

Detection rate, false negative rate, accuracy and sensitivity were calculated using the formulas below:

Ethics

The study was approved by the Scientific Committee in Aarhus County and was in accordance with the Helsinki Declaration.

Results

In the study group (Table IV) we had a successful detection rate of 92% (95% CI: 74–98%. The two patients where no sentinel nodes could be detected had both metastases to the lymph nodes. These two patients are not included in the calculations of the false negative rate and sensitivity. The median number of sentinel lymph nodes removed was 1 (range 1–5), and the mean number was 1.8 sentinel lymph nodes. Overall, the median number of removed lymph nodes was 13. Among the patients with successful sentinel lymph node detection 14 patients had lymph nodes metastases. The false negative rate was 29% (95% CI: 7–39%) and the sensitivity 75% (95% CI: 51–90%). The accuracy rate was 82% (95% CI: 61–93%).

Table

Results of the study	Present study	Lauridsen et al. [6]
N	24	124
Successful SLNB	92% (22/24)	98%
No. of removed SN (median)	1 (min. 1, max. 5)	2 (min. 1–max. 7)
False negative rate n (%)	29% (n = 4) ((14–10)/14)	1.5% (n = 2)
Accuracy	82% ((10 + 8)/22)	98%
Sensitivity	75% (12/16)	100%

For comparison reasons, the validation study from our department on patients with unifocal, early breast cancer are shown. 95% CI are mentioned in the text.

We had four false negative procedures and data on these patients are shown in Table V. All false negative patients had two tumors on their mammograms and the dominant tumors were in all cases located in the upper lateral quadrant. In three cases the secondary tumor was located on the border between the upper lateral and upper medial quadrants. In one case both tumors were in the same quadrant. The dissected sentinel lymph nodes were both radioactive and stained blue and when the sentinel lymph nodes were removed during the operation no significant radioactivity remained in the axilla. The dominant tumor size was in all cases below 40 mm with the largest tumor being 33 mm. Histology showed three ductal cancers and one lobular carcinoma.

Table

Age	51	50	51	68
Localization of dominant tumor	U.L.Q^α	U.L.Q	U.L.Q	U.L.Q
Distance between tumors	24	33	21	60
Dominant tumor size	30	19	13	19
Tumor type	Ductal	Ductal	Ductal	Lobular
No. of sentinel lymph nodes	1	2	1	1
Radioactive and blue lymph nodes	1	2	1	1
Radioactive or blue lymph nodes	0	0	0	0
No. of positive lymph nodes	1/13	1/15	1/12	2/16

Discussion

The largest tumor diameter is a validated parameter to classify patients with multifocal breast cancer according to previous versions of The American Joint Committee on breast cancer TNM staging system [18]. Studies of some classic prognostic tumor markers have shown, that apparently there are no differences between patients with unifocal tumors and multifocal tumors suggesting that patients with multifocal breast cancer and especially with intraductal carcinomas have tumors originating from the same duct in the breast [19]. In contrast, using a calculation of the total tumor volume it has been shown that multifocal breast cancers disseminate at lower volumes compared with unifocal tumors, suggesting that factors other than tumor burden alone play a role in the spread to lymph nodes [20]. In concordance with this, the frequency of lymph node involvement using the diameter of the largest tumor in patients with multifocal breast cancer compared with patients with unifocal breast carcinomas of similar diameter have shown a higher frequency of lymph node metastases among patients with multifocal breast cancer [21].

Traditionally the sentinel lymph node biopsy technique is limited to patients with stage I and II disease i.e. patients with tumors less than 5 cm in diameter, because the risk of metastases increases with increasing tumor size [22]. Theoretically, the afferent lymphatic vessel to a metastatic lymph node can be obstructed by tumor deposits. If this is the situation for the afferent lymphatic vessel to a sentinel lymph node, this will increase the risk of a false negative SLNB procedure. Furthermore, this obstruction can cause alternative lymphatic drainage patterns, for example to internal mammary nodes. This can lower the detection rate and eventually increase the axillary false negative rate [23]. Indeed, in our study three of the four false-negative patients had only one axillary sentinel lymph node detected. We did not use lymphoscintigraphy and detection of sentinel lymph nodes in the internal mammary chain did not occur. Using lymphoscintigraphy could possibly help in revealing sentinel lymph nodes outside the axilla. If this has any impact on the false negative rate remains to be investigated.

In the present study we used peritumoral injection of the radioactive colloid and blue dye. At least in patients with unifocal breast cancer the peritumoral injection technique has been shown to be as good as more superficial injection techniques [24–26]. The accumulating knowledge and insight into the lymphatic drainage pattern has demonstrated that the breast lymphatic drainage goes through a plexus located in the subareolar region and via the same few superficial lymphatic vessels to just a few axillary nodes. This suggests that superficial injection techniques could have advantages compared with peritumoral injection technique, if the sentinel lymph node technique should be standard treatment to patients with multifocal breast cancer [27–29]. However, internal mammary lymph nodes will not to the same extent as with peritumoral injection be visualised using the superficial injection techniques [27], [28]. The understanding of the lymphatic patterns from the breast has by some lead to the recommendation that SLNB can be safely used in patients with multifocal breast cancer despite the fact, that there are only few published papers with only small number of patients [30], [31].

In this study we used the same diagnostic setup as in our first validation study in order to evaluate the method among patients with more than one tumor in the breast. It is important to emphasize, that validation studies should use the same techniques, indications and contraindications as when the technique would be used in routine clinical practice. To our knowledge none of the earlier published papers on patients with multifocal breast cancers have applied that principle. One of the studies included patients with palpable lymph node metastases, although there is no benefit in this group of performing SLNB [32]. Two of the studies included patients where additional tumors in the breast were found on the pathological examination and therefore were not known preoperatively [33], [34]. This makes these studies difficult to compare with results on patients known preoperatively [35], [36] having multifocal or multicentric breast cancer. The results from these two studies are applicable to the clinical scenarios they report. However, they do not answer the question, whether SLNB is a safe procedure among patients with preoperatively known multifocal breast cancer.

Our results further show that not many patients with multifocality benefit from SLNB. With the use of ultrasound in the diagnostic scenario of breast cancer more patients will have metastases detected preoperatively. Of 59 patients just eight patients were without metastases. This suggests that patients with early, multifocal breast cancer represent a much selected population. However, our study has a limited number of patients and the estimates have large confidence intervals, so the conclusions are not that strong.

We therefore conclude from this study that more and larger validation studies are needed, before it is acceptable to use the sentinel lymph node biopsy technique, when treating patients with preoperatively diagnosed multifocal breast cancer.

Acknowledgements

This research was funded by a grant from the Danish Cancer Society. No:98 100 29