Microvessels that predict axillary lymph node metastases in patients with breast cancer

Hypothesis: The density of vasoactive endothelial growth factor receptor 3- immunostained microvessels in primary breast cancers correlates with the in cidence of axillary lymph node metastasis. Design: Breast cancer microvessel clusters ("hot spots") were sequentially immunostained for factor VIII, type IV collagen, and vasoactive endothelial growth factor receptor 3. Microvessels were counted under light microscopy at a magnification of x200. Axillary lymph nodes were evaluated for metast ases by light microscopy. Setting: A multidisciplinary breast cancer clinic and laboratory. Patients: Sixty patients with T2 breast cancers treated by lumpectomy (or m astectomy) and axillary lymphadenectomy. Main Outcome Measures: Putative lymphatic microvessel density compared with axillary metastases. Results: There were 16% (SE, 1.4%) vs 4% (SE, 0.8%) vasoactive endothelial growth factor receptor 3-immunostained microvessels (P<.001), 38% (SE, 3.9% ) vs 65% (SE, 3.1%) type IV collagen-immunostained microvessels (P<.001), a nd 46% (SE, 4.1%) vs 31% (SE, 3.2%) unstained microvessels (P = .004) in no de-positive vs node-negative patients, respectively. A fitted logistic mode l based on the relative percentage of putative lymphatic microvessels to bl ood microvessels correctly predicted that 23 (96%) of 24 patients would hav e a low risk and that 26 (96%) of 27 patients would have a high risk of lym ph node metastases. Six (67%) of 9 patients predicted to have an intermedia te risk had lymph node metastases. Conclusion: The odds of a patient with breast cancer having axillary lymph node metastasis increased substantially as the proportion of putative lymph atic microvessels increased and the relative proportion of blood microvesse ls in angiogenic hot spots decreased (log likelihood = 14.6; chi(2) = 53.4; P<.001; area under the receiver operation characteristic curve = 0.97).