A biological characterisation carried out on 14,007 primary breast cancers provided evidence in favour of a relation between advanced age and favourable features (positivity for oestrogen/progesterone receptors, low proliferative rate, absence of p53 accumulation, bcl-2 overexpression, diploid DNA content), showed a similar pattern of association between patho-biological variables regardless of patient age, and demonstrated a relation between biological variables and disease outcome in the elderly, comparable to that already reported for younger patients. In fact, oestrogen receptor and proliferative activity provided independent prognostic information either in node-negative or in node-positive tumours treated with radical or conservative surgery plus radiotherapy, alone or followed by adjuvant hormonal therapy. It would be thus reasonable to use biomarkers as a complement to clinico-pathological features in a 'risk-factor profile system' even for elderly patients, upon their validation in prospective studies and after assessing the cost-benefit of treatments planned on the basis of biological information.