“…It is generally agreed that tumour markers in breast cancer patients are not a tool for primary diagnosis, because of their low sensitivity and specificity (Tondini et al, 1989;Fateh-Moghadam and Stieber, 1993;Lamerz et al, 1993). Their use for early detection of metastases seems to be promising (Lamerz et al, 1991;Stieber et al, 1992;Vizcarra et al, 1994;O'Hanlon et al, 1995) and their use for measuring therapeutic response in metastatic disease is widely accepted (Tondini et al, 1988;Dnistrian et al, 1991;Robertson et al, 1991;Safi et al, 1991). Many studies tried to assess the prognostic role of these tumour markers (some analysed in serum, some in tissue), but most of them had low patient numbers or short follow-up periods, and used only univariate analyses (Myers Tormey and Waalkes, 1978;Bezwoda et al, 1981;Mansour et al, 1983;Kallioniemi et al, 1988;Hammer et al, 1992;O'Hanlon et al, 1995).…”