Tumour sizes varied across the trials. Three trials,8,10-12,27 including the Australian SNAC I trial,8,27 included patients with tumours equal to or less than three centimetres in diameter. The Milan trial4,5 excluded patients with tumours more than two centimetres in diameter. Two trials6,20 did not exclude patients based on tumour size. However, up to 80% of patients in these two trials had patients with tumours equal to or less than two centimetres in diameter. Only 2% of patients in the NSABP B-32 trial20 had tumours more than four centimetres in diameter and 2% of patients in the ALMANAC trial6 had tumours more than five centimetres in diameter.
While the NBCC* definition of early (operable) breast cancer refers to tumours not more than five centimetres in diameter1, there are limited trial results to support sentinel node biopsy in patients with tumours greater than three centimetres in diameter.
* In February 2008, National Breast Cancer Centre (NBCC), incorporating the Ovarian Cancer Program, changed its name to National Breast and Ovarian Cancer Centre (NBOCC). In July 2011, NBOCC amalgamated with Cancer Australia to form a single national agency, Cancer Australia, to provide leadership in cancer control and improve outcomes for Australians affected by cancer.
Three of the trials reported on overall and disease-free survival, with equivalent rates in both the sentinel node biopsy and the axillary dissection groups. Although reporting at five-year follow-up, the Milan5 and GIVOM11 trials are small trials and the ALMANAC trial6 has only reported a median follow-up of 12 months, so it is difficult to draw firm conclusions based on these results. Further results from longer follow-up are needed to determine if sentinel node biopsy provides an overall survival and disease-free survival effect compared to axillary lymph node dissection.
The current reported rates of axillary recurrence in the trials are low, with a relatively short (five year) follow-up. Further results from longer follow-up are needed to determine the effect of sentinel node biopsy on axillary recurrence rates compared to axillary lymph node dissection.
The trials6,8,9,11 reported high success rates for localising the sentinel node for both the sentinel node biopsy and axillary dissection groups. Rates reported ranged from 93% to 98%, with slightly higher rates observed in the sentinel node biopsy arms. The accuracy (up to 97%) of sentinel node biopsy in the detection of the sentinel node could only be reported in trials that performed sentinel node biopsy followed by axillary dissection in the control arm.4,9,11
Sentinel node biopsy is associated with lower morbidity compared to axillary lymph node dissection, including lower rates of lymphoedema, arm morbidity and sensory deficit. The trials reported on a variety of physical morbidity outcomes, which were often measured differently. Lymphoedema, shoulder and arm functioning, numbness and pain were reported less often in the sentinel node biopsy groups than the axillary dissection groups.4,6-8,10,12
Three of the trials, ALMANAC,6 SNAC I22 and NSABP B32 9,21 reported that prior to participation in the trial, surgeons must have been trained and accredited in the sentinel node biopsy technique to ensure optimal performance. Training was measured differently, however, generally each surgeon had to perform a certain number of sentinel node biopsies and meet set criteria (for example localisation rates more than 90% and false negative rates less than 5%). The training requirements to achieve these criteria differed between the trials (from 1–40 completed procedures). The Milan5 trial reported their surgeons were very experienced and expert pathologists were used. In the Australian setting, the SNAC I22 trial protocol states: the surgeon should have completed sentinel node biopsy in at least 20 consecutive cases with a greater than 90% success rate in locating the sentinel node. The ongoing SNAC II28 trial protocol states the surgeon should meet the criteria for SNAC I22 or have equivalent evidence of their experience with the technique.
Most sentinel nodes will be located in the axilla. Results of the trials are based on the excision of all detected sentinel nodes (i.e. hot and blue nodes, or hot nodes where the isotope count is greater than ten times the background count, or blue nodes). There was a higher rate of detection of non-axillary sentinel nodes using a peritumoral injection site.16 There are limited data from the trials to determine the utility of excision of non-axillary sentinel nodes. Identification of non-axillary sentinel nodes only occurred in up to 13.4% of cases.16
Sentinel node biopsy can produce false negative results. Although not explicitly tested, comparison of trial results suggests that experience of the surgeon performing the procedure may influence false negative rates. 4,8,9,11 The Milan4 trial reported a false negative rate of 8.8%, similarly the SNAC I8 and NSABP B-329 trials reported a false negative rate of 8.2% and 9.7% respectively. The GIVOM11 trial, which did not require formal training for surgeons participating in the study, reported a higher false negative rate of 16.7%.
The pathological methods used to examine the sentinel node differed for each trial. The Milan4,5 and GIVOM11 trials used frozen section intraoperative evaluation, and GIVOM11 followed this by staining with haematoxylin-eosin (H&E) and cytokeratin antibody (immunohistochemistry) of serial sections for definitive histology. NSABP B329 undertook intraoperative cytology and H&E staining of serial sections. ALMANAC6 examined the lymph nodes by H&E staining of serial sections only. The SNAC I22 and Cambridge10 trials used H&E staining of serial sections, followed by immunohistochemistry to detect smaller metastases.
A positive result on intraoperative assessment (cytological methods or frozen section) of the sentinel node may negate the need for second surgery, as axillary dissection may be performed immediately. There are two methods of intraoperative assessment for sentinel nodes; cytology and frozen section, both of which provide excellent specificity and acceptable rates of sensitivity. A moderate rate of false negative is inevitable with both techniques.9, 11
A potential disadvantage of the frozen section technique is that tissue is discarded during the preparation of the frozen section, and all care should be taken to keep this to a minimum.23 Each pathology laboratory, in consultation with the surgical teams, should determine which procedure is most appropriate to be implemented depending on the resources and expertise available in the laboratory. Definitive histological assessment is needed to confirm intraoperative results.
Associated risks of radioisotope use in sentinel node biopsy are minimal and within acceptable limits for patients and staff.24 Data relating to the use of radioisotope are available in trial protocols.
Most trials excluded patients with known allergies to blue dye or isotope. The NSABP B-329 trial is the only trial to report on adverse allergic reactions. Allergic reactions occurred in 0.7% of patients (0.2% grade 3 or 4 reactions). A review by the Australian Medical Services Advisory Committee (MSAC)24 found case series reported small percentages of allergic reactions (0–1.6%) to blue dye. Up to August 2002, the Australian Drug Reactions Advisory Committee (ADRAC)26 had received 42 reports of reactions to Patent Blue V. Five cases of anaphylaxis have been reported between October 2000 and August 2002 in women undergoing breast surgery, four of which were considered severe. The Australian Therapeutic Goods Administration (TGA) has stated that surgeons and anaesthetists should be aware of the potential for severe allergic reactions to Patent Blue V and that testing for hypersensitivity is recommended by the Product Information.26
Sentinel node biopsy appears to offer equivalent or improved quality of life for patients compared with axillary dissection, such as reduced arm morbidity (for example, reduced risk of lymphoedema).4-8,10,12 Quality of life was assessed by various validated scales and questionnaires by four randomised control trials.6-8,10,12 Significantly higher quality of life was reported by patients receiving sentinel node biopsy than those receiving axillary lymph node dissection in three trials,6-8,10 however, one of the trials10 reported no significant difference on some measures. One trial12 reported no significant difference between patients receiving sentinel node biopsy or axillary lymph node dissection; however, quality of life information was available for only small number of patients in this trial. Most trials used validated questionnaires, such as the FACT or SF-36 (except Milan4,5), and the SNAC I8 trial created its own validated scale. The ALMANAC6,7 trial used a breast cancer specific questionnaire, FACT-B, which showed improved quality of life in the sentinel node biopsy arm compared to axillary dissection. Where trials used broad, general scales, differences in quality of life between patients receiving sentinel node biopsy and patients receiving axillary dissection were not reported.
Length of hospital stay
Patients receiving sentinel node biopsy reported shorter stays in hospital compared to patients receiving axillary dissection.4,6
Three trials investigated differences in detection agents used to identify the sentinel node.13-15 Comparisons were made between blue dye alone and a combination of blue dye and radioisotope. Two trials14,15 reported that higher sentinel node detection rates were associated with combination methods for lymphatic mapping. Overall, accuracy in these trials was high and no difference was reported between combination methods and blue dye alone. One other trial13 reported a possible association between combination methods and improved accuracy; however, accuracy rates in this trial were lower for both combination and blue dye alone groups than the two previous trials. Sensitivity, specificity and false negative rates appear to be similar between blue dye alone and combination methods, however each of these trials contained small numbers of patients. Further research is needed to determine optimal detection agent(s) for sentinel node biopsy. Please refer to Table 1 for the detection methods used in the main trials and Table 2 for details of trials investigating technique of sentinel node biopsy.
Two trials16,17 reported on differences in the injection site used for sentinel node biopsy (i.e. peritumoral, periareolar, intradermal) and each investigated different routes. Further research is needed to determine the optimal injection site of blue dye/radioisotope for sentinel node biopsy. Non-axillary sentinel node sites were identified more commonly with peritumoral injection. For example, more internal mammary sentinel nodes were identified with peritumoral injection.16
One randomised trial18 reported on treatment of patients with one or two positive sentinel nodes. Patients receiving sentinel node biopsy and axillary lymph node dissection reported more adverse surgical effects than patients receiving sentinel node biopsy alone. Two large ongoing randomised trials are investigating how to treat patients with micrometastases in the sentinel node.29,30