Imprint cytology predicts axillary node status in sentinel lymph node biopsy.

INTRODUCTION

Sentinel lymph node biopsy (SLNB) has gained widespread acceptance for axillary node staging in breast carcinoma. Results from the ALMANAC trial confirmed clear benefits for SLNB in terms of arm function and quality of life measures1.

An accurate and reliable method of assessing sentinel lymph node (SLN) status at the time of primary surgery is desirable. It would avoid the need for a second operation in the significant number of patients who are node-negative at diagnosis.

Imprint cytology is a well-recognised simple technique for preparing a surgical specimen for pathological assessment. The excised SLN is sent fresh to the pathologist who processes it immediately. The cut surfaces are pressed onto a glass slide, which is then fixed and stained.

The aim of this study was to assess the use of imprint cytology as an infra-operative tool for evaluating sentinel lymph nodes in patients with clinically node-negative breast cancer.

MATERIALS AND METHODS

Data was collected prospectively in a specialist breast unit in a district general hospital serving a population of 300,000 and treating approximately 170 new breast cancers each year. Patients meeting the inclusion criteria stipulated by the ‘New Start’ programme2 were studied consecutively.

Three consultant breast surgeons underwent training in SLNB (under the auspices of New Start/ Royal College of Surgeons of England). Each surgeon performed 5 cases proctored by a surgeon experienced in the technique, followed by a further 25 cases without supervision. After SLNB, a formal axillary node clearance (ANC) was performed.

SENTINEL NODE IDENTIFICATION

Sentinel nodes were identified using a combination of patent blue dye (Patent Blue V; Guerbet Laboratories Ltd, Birmingham, UK) and radioactive tracer (99mTechnecium-labelled human albumin nanocolloid particles; Nanocol; Nycomed Amersham PLC, UK). Depending on the scheduled time of surgery, 0.2mls of either 15mBq (same day) or 20mBq (following day) 99mTc nanocolloid was injected intra-dermally into the peri-areolar area of the breast. Uptake of radiocolloid was mapped and a skin mark placed at the level of the sentinel node, as identified by the static detector.

Following induction of general anaesthesia, 2mls of patent blue dye mixed with 2mls 0.9% saline was injected subdermally and gently massaged for 5 minutes. A small incision was made in the axilla at the point of maximum radioactivity as determined by the portable gamma probe (Europrobe, Bright Technologies Ltd, Sheffield, UK). Sentinel nodes were identified by tracing the blue dye (through direct visualisation of the blue-stained lymph channels) and the radio-isotope (using the gamma probe). All blue-stained nodes and/or radioactive foci were excised and sent to the pathology laboratory for immediate processing and analysis. The gamma probe was used to ensure that radioactivity levels had fallen to a point that would be consistent with removal of all sentinel nodes. The axilla was then examined to identify if there were any palpable nodes before it was cleared in the usual manner.

STAINING AND EXAMINATION OF NODES

Each sentinel lymph node was sectioned transversely into 2mm slices. Depending on the size of the node, 2–4 imprints were made from each slice by gently touching the cut surface of the node onto a glass slide. These were air-dried and stained with Rapi-Diff II stain (Triangle Biomedical Sciences Ltd, Lancashire, UK) before being reviewed by 2 or more pathologists using a multi-headed microscope. Analysis was performed during each operation with nodal status consistently determined within 45 minutes of the specimen leaving theatre. However, as these cases were carried out as part of the audit phase of SLNB training, results were not relayed intra-operatively and thus had no impact on the surgery performed.

In some cases, suspicious groups of cells were present on the imprints. However, a positive report was only given if the number and/or the morphological features of the cells were sufficient to give a diagnosis of definite metastasis. (Figure 1) All slices of the sentinel lymph node were formalin-fixed and embedded in paraffin. They were then examined after Haematoxylin and Eosin (H&E) staining. If negative on H&E, nodes underwent immunohistochemical staining with the monoclonal anticytokeratin antibody Clone MNF 116 (Dako, Glostrup, Denmark) using the avidin-biotin-peroxidase complex method. The pathologist who prepared and reported the imprint was also responsible for reporting the sentinel lymph node H&E / immunohistochemistry sections, the axillary nodes and the breast specimen.

Fig 1

positive imprint cytology; Rapi-Diff II stain; x40 magnification

RESULTS

Over an 11 month period, 102 consecutive patients with clinically node-negative disease had SLNB followed by axillary clearance. The mean age was 58.8 (28-89). The median tumour size was 20mm. Most patients (60%) had grade 2 tumours, 20% had grade 3 and the remainder had grade 1. Lymphovascular invasion was present in 33% of patients. (Table 1)

Table 1

Primary Tumour Characteristics

Mean Age (years)	58.8
Mean Tumour Size (mm)	20
Lymphovascular Invasion (% patients)	33
Histopathology (% patients)	Ductal	63
	Lobular	15
	Other	22
Tumour Grade (% patients)	1	20
	2	60
	3	20

An average of 2.3 nodes (1-9) were identified per patient. The identification rate was 100%. Sentinel node metastases were detected in 41 patients. Metastatic deposits >2mm were designated macrometastases (Figure 2) while those ranging between 0.2 and 2mm were considered micrometastases (Figure 3). Histopathological analysis of the axillary clearance nodes revealed metastatic disease in 44 patients, giving SLN biopsy a sensitivity of 93.2%.

Fig 2

lymph node with macrometastasis; MNF 116 stain; x2.5 magnification

Fig 3

lymph node with micrometastasis; MNF 116 stain; x2.5 magnification

Intra-operative imprint cytology identified 33 of the 41 patients with sentinel node positive disease. There were no false positives. There were 8 cases where imprint cytology of the sentinel node was negative but metastases were detected by H&E and immunohistochemistry. Of these false negatives, 4 were macrometastases and 4 were micrometastases. Immunohistochemistry did not detect any further metastases on the H&E negative sections. In total, there were 6 cases of micrometastases. Two of these were positive on imprint cytology, with one case having micrometastatic deposits in 2 out of the 4 detected sentinel nodes.

These figures give imprint cytology in this study an overall sensitivity of 80%, a specificity of 100% and a negative predictive value of 88%.

DISCUSSION

It is well accepted that axillary node status is the most important prognostic indicator in patients with invasive breast cancer. Knowing the nodal status is essential for correct cancer staging and helps determine the need for adjuvant therapies. However axillary lymph node dissection (ALND) is associated with significant morbidity, with up to 60% of women experiencing long-term side-effects. (3). Moreover ALND is unnecessary for women who have node-negative disease and studies have shown this can be as high as 70% in those with Tl and T2 tumours.3

Sentinel nodes have been shown to be representative of the presence or absence of metastases in the remainder of the nodal basin. Sentinel lymph node biopsy is increasingly being used to predict axillary node status in breast cancer on a world-wide basis. It allows directed therapeutic node dissections and confines the morbidity of the procedure to patients who will potentially benefit from removal of involved nodes.4 Data from the randomized controlled ALMANAC trial (Axillary Lymphatic Mapping Against Nodal Axillary Clearance) confirmed clear benefits for clinically node-negative patients undergoing SLNB, rather than conventional axillary treatment, in terms of arm function and quality of life measures.1,5

Various techniques for localization and assessment of the sentinel node have been employed by different centres over the past decade. Whilst there is no current “optimal” protocol6, detection using a combination of radiotracer administration, preoperative nuclear medicine imaging, blue dye injection and intraooperative gamma counting has been advocated as this appears to increase the sentinel node yield and reduce the learning curve6,7.

Once harvested, sentinel nodes undergo a thorough histopathological examination. Multisectioning rather than routine bisectioning is known to decrease the sampling error phenomenon and increase metastatic tumour detection6. Other studies have shown that cytokeratin immunohistochemistry staining also increases metastatic tumour detection when compared with H&E staining6. However immunohistochemical analysis of H&E negative sentinel nodes did not upstage any of the patients in our study.

The frequently reported methods of intraoperative assessment are frozen section histology and imprint, or touch-preparation, cytology. Reports of frozen section examination have described a sensitivity of 44-100% and a specificity close to 100%8. However, the procedure is time-consuming and the process of freezing, then thawing, the specimen can introduce artefacts. Furthermore, there is often significant tissue loss, potentially interfering with subsequent more detailed pathological examination with paraffin sectioning8.

Imprint cytology proved to be a very efficient tool for intraoperative assessment of the sentinel nodes due to clear lines of communication between theatre and pathology staff. Although the time varied depending on the size of the node and the number of nodes requiring analysis, results were usually available within 45 minutes. This is comparable with the experience of other units carrying out intra-operative imprint cytology9. Whilst waiting for the result, surgery to the breast can be performed.

The literature suggests that imprint cytology is comparable in accuracy to frozen sectioning10. In this study the technique had a sensitivity of 80% and a specificity of 100%. It is important to remember that the definitive SLN status assessed by histopathological assessment of the node is the standard with which results of intra-operative evaluation are compared. Hence the detailed sampling carried out may indicate a less favourable intra-operative accuracy than if a limited sampling of the SLN had been performed.

There were only 8 false negative cases and 4 of these were micrometastases. The sensitivity of imprint cytology in detecting micrometastases in this study was 33% (Table 2). This proportion is higher than that found by other centres8 and may be due to the relatively small numbers in the study. It is noteworthy that most subsequent cases of micrometastases in our unit have been imprint negative.

Table 2

Sensitivity of imprint cytology

Sentinel node metastases	Final histopathology (no. of cases)	Imprint cytology (no. of cases)	Sensitivity of imprint cytology (%)
Macrometastases	41	33	80%
Micrometastases	6	2	33%

As SLNB becomes more widely used, detection of micrometastasis in sentinel nodes is increasingly proving a therapeutic dilemma. The prognostic significance and clinical relevance of these previously occult metastases is controversial11. Hansen et al examined the John Wayne Cancer Centre experience with 790 patients who had undergone SLNB11. They observed that, at 8 years, patients with micrometastases in their SLNs had better prognosis than patients with SLN macrometastases and had prognosis equal to those with SLN-negative disease.

In contrast, the International (Ludwig) Breast Cancer Study, one of the largest studies of patients with nodal micrometastases described to date, reported that 83 patients with micrometastases had a worse disease-free and overall survival after 5 years median follow-up than did patients who were node-negative on retrospective analysis and serial sectioning12. A large recently published retrospective review from the Netherlands also demonstrated that micrometastastes were associated with an absolute reduction in 5-year disease-free survival of nearly 10%13.

Further data from larger studies with longer follow-up is thus required before definitive conclusions regarding the relevance and optimal management of micrometastases in sentinel lymph nodes can be made. Notwithstanding, even if micrometastases were regarded as positive nodal disease, in clinical practice 80% of patients in this series could have had their primary tumour and their axilla treated in one operation.

CONCLUSION

Sentinel lymph node biopsy is rapidly becoming the standard of care for patients with breast cancer. The technique can be learned quickly, but SLNB is a multidisciplinary process requiring continuous audit.

Imprint cytology has been shown to be reliable for predicting SLN status. This study has demonstrated that imprint cytology allows a one stage procedure in 80% of patients with node positive disease, and is to be commended as a useful and practical technique for those using SLNB for axillary staging in breast cancer.