Clinical evidence has shown the limitation of pathological staging to give accurate prognostic information and its inability to predict clinical response to specific treatments. It therefore has been considered of utmost importance to integrate pathological stage with biological variables of consolidated predictive relevance to improve prognostic accuracy and possibly to identify indicators of response to different treatments.With regard to the first point, successive generations of prognostic factors have been proposed (Table 1): a first generation of pathological factors including tumour size, stage and morphology; a second generation including biochemical markers, steroid and growth factor receptors, cell kinetics and ploidy; and a third generation including oncogenes and oncosuppressor genes, proteases, extra-cellular matrix-related antigens and angiogenesis markers. For some tumour types, biological markers have been increasingly used to complement clinicopathological findings for a more accurate definition of individual patient prognosis and for helping clinicians in treatment decision making.
PROGNOSTIC RELEVANCEThere has been considerable interest in proliferative activity for several years. Preliminary results showing the relevance of proliferative status as an indicator of biological and clinical aggressiveness have led to the proposal of several approaches based on dilferent rationales, with the main purpose to have efficient, inexpensive, highly feasible techniques without compromising the reliability of the results (Quinn & Wright 1990). The two most intensively investigated indicators first used to obtain information on cell proliferation are the fraction of cells in the S-phase, quantified on the basis of nucleic acid precursor incorporation ( [3H]-thymidine labelling index, [3H]dT LI, bromodeoxyuridine labelling index, BrdUrd LI) or of DNA-synthetic content (FCM-S), and more recently, the entire fraction of proliferating cells, hypothetically quantified by the expression of a nuclear antigen detected by Ki-67 monoclonal antibody. Available data on the prognostic relevance of these cell proliferation markers, generally determined at a single point during the clinical life of a tumour, were derived from a large series of patients with solid and systemic diseases. The clinical endpoints considered were relapse-free and overall survival for the former (possibly subjected to local-regional therapy), and overall survival for the latter diseases (subjected to systemic treatments). However, basic and clinical results do not support the Correspondence: Professor R. Silvestrini, Oncologia Sperimentale C, Istituto Nazionale Tumori,