PCR analysis has been demonstrated as a valuable tool for detection of minimal residual disease (MRD) in lymphoid malignancies. However, the finding that patients with evidence of MRD sometimes remain in long-lasting remission directs further investigations toward biology of residual disease and/or quantification of MRD level. The study included 40 B-NHL patients--13/40 patients with high- (HG) and 27/40 with low-grade (LG) lymphoma. Seven patients achieved partial clinical remission (PR) and 33 patients achieved complete clinical remission (CCR) after chemotherapy. Peripheral blood samples were analyzed for MRD at up to ten follow-up points while samples of MRD+ patients and patients who achieved partial clinical response after therapy were further analyzed for the presence of t(14;18) and P53 and RAS genes mutations. The level of MRD was quantified in eight patients by PCR-limiting dilution method.MRD was found in 13/33 patients (12 LG and 1 HG) who achieved CCR. The incidence of relapse was significantly higher in MRD+ vs. MRD- B-NHL patients (Fisher's exact test, p = 0.0083). In the LG group the incidence of relapse between MRD+ and MRD-patients was not significantly different. In the HG group MRD was detected in only one patient who subsequently relapsed. Significant difference in DFI between MRD+ and MRD- patients was not observed. Concerning MRD+ patients in CCR and patients who achieved PR, t(14;18) was found in six patients (4 relapsed). In the same group of patients P53, K- and N-RAS mutations were not found. H-RAS mutations were found in six patients--3 relapsed and 3 remain in CCR. The calculated number of IGH copies ranged from 4800 to 44,000. Our results demonstrated positive correlation between MRD-positivity and incidence of relapse in B-NHL patients, but could not indicate significance of P53 and RAS mutations for evaluation of residual clone malignancy. The study implies that MRD level, measured at one follow-up point, does not correlate with clinical outcome. The measurement of MRD level sequentially, at different follow-up points, seems to be a better parameter for the prediction of disease course.