Anatomic Location of Breast Cancer Micrometastasis in Sentinel Lymph Node Predicts Axillary Status

To the Editor:

We read with interest the paper by Viale et al,¹ and we would like to comment on the topic of sentinel node prediction of axillary status, based on data from a series we collected at our institution.

The assumption is that the great majority of patients with positive sentinel lymph node (SLN) are not expected to harbor additional metastases, after complete axillary lymph nodes dissection. Accordingly, Viale et al¹ tried to identify pathologic features able to predict the risk of additional non-SLN metastases, to avoid unnecessary surgery. They suggested that 3 features are strongly associated with non-SLN metastases, namely, the size of metastatic deposits, the number of SLN involved, and the peri-tumoral vascular invasion of primary tumor, the former being the most powerful predictor.

From January 2000 to December 2004, 540 patients with clinically negative lymph nodes and “small” (<3 cm) breast cancer underwent SLN biopsy at the Istituto Clinico Humanitas (Milan, Italy), according to previously described protocols.² Nodal metastases were found in 162 (30%) patients who underwent complete axillary dissection. SLN and non-SLN were examined on formalin-fixed, paraffin-embedded sections, as also reported by Viale et al,¹ without the aid of cytokeratin immunostaining. Clinicopathologic features of these patients were recorded and data reported as number and percentage, or mean and standard deviation, when appropriate; categorical data were compared with Fisher exact test.

Of 162 patients, SLN micrometastasis and metastasis were documented in 62 (38%) and 100 (61%) cases, respectively. The overall prevalence of additional non-SLN metastases was 32.7% (53 of 162 patients); patients with SLN metastasis had a significantly (P < 0.0001) higher proportion of additional metastases (43 of 100, 43%), as compared with patients with SLN micrometastasis (10 of 62, 16.1%). Interestingly, additional metastases from SLN micrometastasis were associated not only to the size of metastatic deposits, as shown by Viale et al,^1,3 but also to the anatomic site. Indeed, when SLN micrometastases were stratified according to their nodal site, the prevalence of additional metastasis was 3% for sinusal (1 of 31) and 29% (9 of 31) for intranodal location (P = 0.026); when SLN micrometastasis were stratified according to their size (up to 1 mm versus 1–2 mm), the prevalence of additional metastasis was 8% for those up to 1 mm (3 of 37) and 28% for those of 1 to 2 mm (7 of 25) (P = 0.045).

The association between the intranodal site of SLN micrometastasis and the axillary involvement has not been documented previously. According to our data, additional non-SLN metastases are significantly more frequent in patients with intranodal SLN micrometastasis as opposed to sinusal micrometastasis, being a new predictor for axillary status. Notably, analogous features have been reported in malignant melanoma: a worse outcome was signaled by Starz et al⁴ in patients with intranodal SLN metastasis as opposed to capsular/subcapsular location. Common metastatic targets seem to be shared by malignant melanoma and breast carcinoma, and increasing evidence indicates that certain chemokines may play a specific role. Indeed, both tumors express the receptors CXCR4 and CCR7, which bind to cognate ligands CXCL12 and CCL21;⁵ interestingly enough, these ligands are overexpressed in lymph nodes, lung, liver, and bone marrow which, in turn, are the commonest sites of breast cancer and malignant melanoma metastasis. “In vitro” experiments showed ligand–receptor interactions with generation of pseudopodia and invasive properties while neutralizing antibodies against ligand/receptor abrogated lymph node homing.⁶ It is tempting to speculate that “in vivo” neoplastic cells, entering the subcapsular sinuses through afferent lymphatics, adhere to nodal parenchyma if specific ligands are expressed; if not, they can skip lymph node drainage and redirect to other targets. Studies on the chemokine profile of micrometastasis and clinical follow-up data are required to confirm this hypothesis.

Our results support the hypothesis that the size of micrometastasis is useful to predict the axillary status. This feature was previously investigated³ and recently emphasized in the paper by Viale et al¹ where isolated tumor cells were also considered. Viale et al¹ suggested that patients with metastatic deposits to SLN can be stratified in 3 different risk groups, according to the size of metastatic deposits. We found that the risk of detecting additional metastasis in non-SLN is low but not negligible in micrometastasis up to 1 mm (8%), significantly greater for > 1 mm micrometastasis (28%), and fairly high in patients with SLN metastasis (43%). Our results are therefore in keeping with the suggestion that patients with positive SLN can be stratified into 3 different risk groups, namely, low (up to 1 mm), intermediate (1–2 mm), and high (>2 mm).

In conclusion, we not only confirmed that the size of micrometastasis in SLN breast cancer is a powerful predictor axillary status, but we also propose that the intranodal location of metastatic deposits should be taken into account as a new predictor of axillary status; whether the intranodal/sinusal location of SLN node micrometastasis actually reflects a chemokine-pattern driving phenomenon remains to be established.

ACKNOWLEDGMENTS

The authors thank Miss Margherita Sanna for her precious help with data managing.

Luca Di Tommaso, MD*
Carmelo Arizzi, MD*
Daoud Rahal, MD*
Annarita Destro, PhD, BS*
Massimo Roncalli, PhD, MD*
Marco Alloisio, MD†
Sergio Orefice, MD†
Arianna Rubino, MD†
Emanuela Morenghi, PhD‡
Giovanna Masci, MD‡
Armando Santoro, MD‡
Ivan Del Prato, MD§
Roberto Sacco, MD∥
Department of *Pathology School of Medicine University of Milan Milan, Italy
Departments of †Surgery and ‡Oncology
Humanitas Clinical Institute Rozzano, Italy
Departments of §Senology and ∥Surgery Humanitas-Gavazzeni Clinical Institute Bergamo, Italy luca.di_tommaso@humanitas.it