Introduction: Prognostic factors can be useful to identify node-negative patients at increased risk of relapse who should receive adjuvant treatment. In the past, oestrogen receptor status and mitotic index have been shown to be significant predictors of prognosis. Different techniques for the measurement of these prognostic factors are available.
Methods: Paraffin-embedded tumour specimens from 441 pre-menopausal patients with node-negative breast cancer who were previously randomized onto a trial comparing peri-operative chemotherapy with no further therapy were studied. Oestrogen receptor status was determined by the classical biochemical assay and by immunohistochemistry (ER-IA). Mitotic index was assessed by counting the number of mitoses and by calculating the percentage of tumour cells positively staining for the antibody Ki-67.
Results: There was a good correlation between ER-IA and the biochemical ER-assay (P<0.01), and the percentage of Ki-67 positive tumour cells and mitotic counts (P<0.01) respectively. However, ER-IA significantly predicted disease-free survival (RR=2.67, 95% CI: 1.60-4.44, P<0.01) whereas the biochemical assay was only borderline significant (RR=1.54, 95% CI: 1.00-2.36, P=0.05). Similarly, Ki-67 was a stronger indicator of prognosis (RR=2.84, 95% CI: 1.80-4.48, P<0.01) than mitotic counts (RR=1.56, 95% CI: 1.22-2. 00, P<0.01).
Conclusions: We conclude that ER-IA performs better in predicting prognosis than the classical biochemical oestrogen receptor assay. Ki-67 is a more accurate marker for tumour cell proliferation and predicts prognosis of patients with breast cancer better than do mitotic counts.
Copyright 1999 W.B. Saunders Company Ltd.