Does triple-negative phenotype accurately identify basal-like tumour? An immunohistochemical analysis based on 143 ‘triple-negative’ breast cancers

Abstract

Basal-like cancer has been identified as a breast cancer molecular subclass by genomic analysis [1.Perou C.M. Sorlie T. Elsen M.B. et al.Molecular portraits of human breast tumours.Nature. 2000; 406: 747-752Crossref PubMed Scopus (11740) Google Scholar]. Immunohistochemically, these tumours are characterized by the absence of estrogen receptor (ER), progesterone receptor (PR) and HER2 expression, together with expression of EGF-R and/or CK 5/6 [2.Livasy C.A. Karaca G. Nanda R. et al.Phenotypic evaluation of the basal-like subtype of invasive breast carcinoma.Mod Pathol. 2006; 19: 264-271Crossref PubMed Scopus (859) Google Scholar], whereas Bcl2 and CK 8/18 expression characterizes a luminal phenotype. This molecular subclass bears a high rate of p53 mutations and is common among BRCA1 germline mutation carriers [3.Brenton J.D. Carey L.A. Ahmed A.A. et al.Molecular classification and molecular forecasting of breast cancer: ready for clinical application?.J Clin Oncol. 2005; 23: 7350-7360Crossref PubMed Scopus (748) Google Scholar]. In daily practice, tumours exhibiting a triple negativity for PR, ER and HER2 are commonly referred to as basal-like tumours, although the level of evidence for such a definition is low. In the present study, we analysed by immunohistochemistry (IHC) whether the triple-negative phenotype accurately identified basal-like tumours. Out of 823 tumours from localized breast cancer patients included in two prospective randomized trials included in a tissue microarray [4.Conforti R, Boulet T, Tomasic G, et al. Breast cancer molecular subclassification and estrogen receptor expression to predict efficacy of adjuvant anthracyclines-based chemotherapy: a biomarker study from two randomized trials. Ann Oncol 2007; doi:10.1093/annonc/mdm209Google Scholar], 150 were triple negative for ER, PR and Her2 expression. EGFR, CK5/6, CK8/18, p53 and Bcl2 expression were assessed on 143 of these 150 samples by IHC (Table 1). In addition, ER and PR positivity evaluated by a ligand-binding assay were determined in the past for most of the patients. The cut-off for positivity was 10% tumour cells for ER, PR, Her2 (strong and continuous staining), p53, CK 8/18 and Bcl2, while EGF-R and CK5/6 staining were considered positive when at least one tumour cell was stained. A basal-like phenotype was defined as a Her2–/ER–/EGFR+ and/or CK5/6+ expression [2.Livasy C.A. Karaca G. Nanda R. et al.Phenotypic evaluation of the basal-like subtype of invasive breast carcinoma.Mod Pathol. 2006; 19: 264-271Crossref PubMed Scopus (859) Google Scholar]. Ninety-five (66%) out of the 143 triple-negative assessable tumours presented a basal-like phenotype, while 48 did not (34%). Table 1 reports patient and tumour characteristics of these two entities. The ‘true’ basal-like tumours exhibited a significantly higher tumour grade. Interestingly, a high proportion of ER–/PR–/Her2–/non-basal tumours expressed luminal-associated biomarkers as ER expression by ligand-binding assay (82%), PR expression by ligand-binding assay (61%), and CK18 expression (76%), and could actually be considered as luminal-B tumours. Basal-like cancers exhibited features of apoptosis deregulation with lower Bcl2 and higher p53 expressions when compared with ER–/PR–/Her2–/non-basal cancers. Finally, there was no difference in disease-free survival nor overall survival between the 95 basal-like and the 48 Her2–/ER–/PR–/non-basal tumours [hazard ratio for relapse or death adjusted for tumour grade and stage, and age of the patient: : HR = 1.17, IC95% = (0.62 – 2.20), adjusted HR for death = 1.38, IC95% = (0.68 – 2.78)].Table 1Characteristics of basal-like and non-basal-like cancers in the triple-negative groupCharacteristicsBasal-like n = 95Non-basal-like n = 48P-valueN%N%Age <4632348170.36 46–5199715 51–5619201225 56–6118191225 61–6677510 66101148Grade I0048<0.001 II19202961 III76801531 Unknown––––ER (ligand-binding assay) Negative5567718<0.001 Positive27333182 Unknown13–10–PR (ligand-binding assay) Negative546615390.01 Positive28342361 Unknown13–10–CK 18 (IHC) Negative66711124<0.001 Positive27293576 Unknown2–2–P53 (IHC) Negative45473981<0.001 Positive5053919 Unknown0–0–Bcl2 (IHC) Negative889333790.03 Positive77921 Unknown0–6– Open table in a new tab In conclusion, this study suggests that the triple-negative phenotype, when defined by IHC using a 10% cut-off for the three biomarkers, is not optimal to define basal-like breast cancer. This phenotype indeed includes a mixture of basal-like tumours together with some luminal-B tumours characterized by a decreased ER expression. The triple-negative phenotype should therefore be used with caution to select patients for clinical trials.