We analysed 459 samples from 206 chronic myeloid leukaemia (CML) patients at diagnosis and during or after treatment with interferon-alpha (IFN-alpha) by quantitative Southern blot analysis for BCR rearrangement. In a minority (2%) of Ph-positive patients, no BCR rearrangement was detectable due to breakpoints outside the major breakpoint cluster region (M-bcr) or possibly due to M-bcr deletions. Results from 235 samples were compared with the proportion of Ph-positive metaphases found in contemporaneous bone marrow specimens analysed by conventional cytogenetics. The rank correlation between both methods was 0.82 (P < 0.001). The proportion of CML cells in samples determined by Southern blot analysis (BCR ratio) was significantly different between cytogenetically-defined minor, partial, and complete response groups (P < 0.001). Empirically-derived cut-off points in the BCR ratio were introduced in order to define molecular response groups for comparison to standard cytogenetic response groups: a BCR ratio of 0% was defined as complete molecular response and ratios of 1-24%, 25-50%, and > 50% were defined as partial, minor, and no molecular response, respectively. Using these cut-off points the concordance between both methods was 67% (P < 0.0001), a major cytogenetic response could be predicted or excluded in more than 90% of cases (P < 0.0001). Our findings demonstrated that quantitative Southern blot was as sensitive as cytogenetics and as peripheral blood samples are suitable for this technique it should be considered as the method of choice for routine monitoring IFN-alpha therapy in CML patients.