Currently, breast cancer chemotherapy response can be predicted based on various parameters, with common reporting of tumour grade and Ki67 proliferation index. We analysed their association with pathological complete response (pCR) in a multivariate approach.
The study was carried out in a group of 353 patients, treated by preoperative chemotherapy and prospectively observed. In selected patients, parallel to routing core needle biopsy assessment, gene expression profile of tumour was analysed by oligonucleotide microarrays.
Tumour parameters associated with pCR in univariate analysis were: tumour grade, nuclear grade, mitotic index, Ki67, oestrogen and progesterone receptor (all p < 0.0001), and triple-negative status (p = 0.0032). The highest increase of pCR chance was observed in patients with high-grade tumours and with Ki67 ≥ 20%. In multivariate analysis, only tumour grade and oestrogen receptor status were predictive for pCR independently of other variables, with high grade increasing the odds of pCR 2.42 fold, and high ER decreasing the chance of pCR 0.41 fold. Tumour grading reflects important biological features of breast cancer and is not inferior to proliferation markers, including Ki67. It should be taken into account in decision-making for preoperative chemotherapy in parallel to breast cancer biologic subtypes, because grade 3 tumours exhibit a higher proportion of pCR.