Sentinel node micrometastasis in breast cancer
E. Rutgers
DOI: https://doi.org/10.1002/bjs.4800
2004-10-01
Abstract:Early knowledge of nodal involvement in invasive breast cancer is thought to be valuable. Certainly, early treatment of regional lymph node metastasis improves regional control and may improve survival. Such knowledge is also of prognostic worth and it may influence the choice of adjuvant systemic treatment. The sentinel node procedure is one of the most reliable methods of assessing axillary lymph node involvement in breast cancer1, but what can be expected of this diagnostic tool? First, does it provide further prognostic information to determine the choice of adjuvant systemic treatment? Second, does it determine whether or not the axilla should be treated electively? Sentinel nodes are now examined with great thoroughness, and many protocols for this exist throughout Europe2. The result is the frequent finding of tiny metastases. Micrometastasis is defined as a metastasis of less than 2 mm in diameter, and submicrometastasis as clusters of isolated tumour cells smaller than 0·2 mm in size3. Micrometastases are found in about 30 per cent of all tumourpositive sentinel nodes4. To what extent does such thorough examination of sentinel nodes serve a diagnostic purpose? For example, consider 100 patients with operable T1–2 N0 breast cancer of up to 3 cm in size. Assume that 38 of them have positive axillary nodes. With a falsenegative rate of 0–5 per cent, about 97 per cent of the positive nodes should be identified by the sentinel node technique using serial sectioning and immunohistochemical staining; that is to say, about 37 of the 100 patients should be correctly identified and one would be missed. Of these 37, one-third would have micrometastasis of less than 2 mm – about 12 patients. If these 12 patients had axillary lymph node dissection, two of them (17 per cent) would have further positive nodes removed. Of patients with submicrometastasis, about 9 per cent (one patient) would have further positive nodes identified. All of these figures are to be found in the excellent review paper by Cserni et al.4 in this issue of BJS. Consider now the two or three patients who have further tumourinvolved nodes removed at axillary dissection. About half will have macrometastases and about half will have micrometastases in the nonsentinel nodes. From this it can be appreciated that the omission of multiple slicing of sentinel nodes with immunohistochemical staining would result in one patient with macrometastasis being missed from the original group of 100. This figure is lower than the false-negative rate of the sentinel node procedure itself. With complete axillary lymph node dissection (ALND) staging, all of the nodes with micrometastasis would be removed. However, routine examination of the nodes from an ALND specimen would still miss micrometastases, as it is known that more thorough examination of apparently negative nodes reveals that 10–20 per cent of them are actually positive, a few even with macrometastases5,6. So, even with standard ALND and standard pathological examination, prognostic information would be lost, just as for the sentinel node technique. Does the finding of micrometastasis in lymph nodes yield any worthwhile additional prognostic information? A number of studies suggest that such micrometastasis does not independently influence survival prognosis, when compared to tumour size and grade6. So, missing one micrometastasis by omitting multiple slicing and immunohistochemistry would have hardly any influence in terms of prognostic information. Furthermore, many patients would have adjuvant systemic treatment based on primary tumour characteristics, which may be of much greater prognostic value than nodal status7. It may be argued that ALND results in improved tumour control in the axilla, with a long-term axillary recurrence rate of less than 1 per cent. The analogous rate after a negative sentinel node procedure is not yet available, but short-term follow-up is reassuring, with axillary relapse rates of between 0 and 1·4 per cent8. Examination of sentinel nodes in invasive breast cancer must serve a diagnostic and prognostic purpose. In routine practice, surgeons are not interested in extra information of unknown prognostic value, and it is unreasonable to tax pathologists with a mountain of work of uncertain clinical relevance. Protocols for investigating sentinel (and other) lymph nodes should aim at detecting macrometastasis, which has a known application in day-to-day clinical practice9; omitting immunohistochemistry is not a ‘capital offence’. Still, whatever protocol is used, once micrometastasis is