Immunophenotypic analysis of inflammatory breast cancers: identification of an 'inflammatory signature'.

Inflammatory breast cancer (IBC) is a rare but very aggressive form of breast cancer. Its definition is based on clinical criteria, but a molecular definition could be useful when data are incomplete or features are missing. Recently, the identification of overexpression of E-cadherin in IBC has improved understanding of the molecular basis of this disease. Consequently, the aim of this study was to try to determine an immunophenotypic 'signature' of IBC. A series of 80 cases of IBC were compared with 552 non-IBC control cases and a model was elaborated to evaluate the probability of an inflammatory carcinoma being present in any clinical situation. Tissue microarrays (TMAs) were used to determine the immunohistochemical profile of eight proteins including E-cadherin, EGFR, oestrogen and progesterone receptor (ER and PR), MIB1, ERBB2, MUC1, and P53. All the parameters tested were differentially expressed between IBC and control cases in univariate analysis (p < 0.001). The five variables that were significantly associated with IBC in multivariate analysis were E-cadherin > or = 300 [HR = 5.64 (2.92-10.87)], ER negative [HR = 3.00 (1.67-5.51)], MIB1 > 20 [HR = 3.54 (1.87-6.71)], MUC1 cytoplasmic staining [HR = 2.72 (1.49-4.96)], and ERBB2 positive 2+ or 3+ [HR = 2.46 (1.26-4.78)]. The probability that a breast cancer with this full phenotype at diagnosis was an IBC was 90.5%. If any one of the five parameters was missing, this probability dropped to 75% and was less than 50% when one, two, or three parameters were present. The 5-year overall survival (OS) and 5-year disease-free survival (DFS) of patients with IBC were not significantly different from those of the non-IBC control group that expressed four or five parameters (nIBC-1), but this nIBC-1 control group had a significantly worse outcome than the non-IBC control group (nIBC-2) with only 0-3 parameters (p = 0.0049 for OS and p < 0.0001 for DFS). In conclusion, an immunophenotypic signature was suggested for IBC. This could help to determine the worst cases, independent of clinical criteria. Copyright 2004 Pathological Society of Great Britain and Ireland. Published by John Wiley & Sons, Ltd.