“…At present, these are the most sensitive techniques for MRD detection (one leukaemic cell/10 5 −10 6 normal cells), although molecular abnormalities are detectable only in a proportion of ALL patients (Biondi et al , 1992; Campana & Pui, 1995; Gibson et al , 1996; Heid et al , 1996; van Dongen et al , 1999; Foroni et al , 1999; Freeman et al , 1999; Hosler et al , 1999; Martuza et al, 2002). MRD detection by flow cytometry is based on the identification of leukaemia‐associated marker combinations, which are either not expressed or expressed at different levels of intensity by normal BM cells (Syrjälä et al , 1994; Campana & Pui, 1995; Jennings & Foon, 1997; Ciudad et al , 1998a; Campana & Coustan‐Smith, 1999; Griesinger et al , 1999; San Miguel et al , 1999; Garcia Vela et al, 2000; Porwit‐McDonald et al, 2000). A number of studies, using either flow cytometric or molecular approaches, showed that the presence of detectable MRD at any time point during the treatment course can predict relapse in childhood ALL (Biondi et al , 1992; Campana & Pui, 1995; Bear, 1998; Campana & Coustan‐Smith, 1999; van Dongen et al , 1999; Foroni et al , 1999; Griesinger et al , 1999; San Miguel et al , 1999; Coustan‐Smith et al, 2000; Pui & Campana, 2000; Radich, 2000; Sievers & Radich, 2000; Dworzak et al, 2002) and consequently current multicentre protocols have been designed on the basis of MRD monitoring.…”