Gene expression profiling can predict pathological complete response (pCR) to anthracycline-monotherapy in estrogen-receptor (ER) negative breast cancer (BC) patients.

Desmedt C, Azambuja E, Di Leo A, Larsimont D, Durbecq V, Antoine D, Lallemand F, Haibe-Kains B, Cardoso F, Nogaret JM, Fontaine D, Liebens F, Duhem C, Kains JP, Maerevoet M, Richard V, Vindevoghel A, Delaloge S, Buyse M, D’Hondt V, Piccart M, Sotiriou C.

Purpose: The TOP trial is an international study that aims to identify biological markers associated with pCR to neoadjuvant anthracycline chemotherapy (CT). Its unique characteristics are: 1/Determination of the predictive factors of response to single agent epirubicin; 2/Evaluation of response in ER-negative pts only, eliminating the confounding effect of indirect ovarian suppression in ER+ BC. 3/Sample size calculation based on biological hypothesis.

Patients and methods: In an interim analysis, 81 pts were evaluated (64 early BC pts: epirubicin 100mg/m2 q3wks x4 / 17 inflammatory or locally-advanced pts: 100mg/m2 q2wks x6). pCR was defined as disappearance of all invasive cancer in the breast and lymph nodes after epirubicin. After surgery, all pts received docetaxel. We evaluated protein expression of ER, PgR, HER2, Ki67, CK5/6 and topoII, and gene amplification of HER2 and topoII. Gene expression profiles were generated from pre-treatment biopsies using Affymetrix U133 Plus 2.0 chips.

Results: The median age and tumor size were 46 yrs and 3.0 cm. Of the 69 evaluable pts, 13% showed a pCR and 60% a clinical response. No association was observed between pCR and age, size and the evaluated markers, except that all pCR pts had high Ki67 expression (>25%). Microarray data was available for 62 pts and a student t-test analysis identified 568 genes that were significantly associated with pCR (p<.001). The probability of selecting these genes by chance was estimated to be p=0.005 after 1000 random permutations. These genes were mainly involved in cell growth, death, morphology and signaling. We recently developed gene expression indices for several hallmarks of BC (Sotiriou et al., ASCO 2006) and found that pCR’s were associated with the expression of stroma and topoII index genes.

Conclusions: These results suggest that a group of genes can identify ER-negative BC pts likely to respond to epirubicin. Since 400 pts will be enrolled, this gene set will soon be tested on a larger cohort of pts and a predictor will be developed. We are also testing whether these genes predict response to chemotherapy in general or only to anthracyclines.