Background: The Digital PCR (dPCR) technique has the potential to monitor minimal residual disease in patients with BCR-ABL positive leukemias, since it allows absolute quantification of the target sequence. A concern re dPCR is possible non-linearity at high copy numbers (CN). We assessed linearity for BCR-ABL and ABL quantitation by dPCR and compared the rates of molecular responses (MR), on the EUTOS scale, in patients (pts) with chronic myeloid leukemia (CML) who had received 18 months of nilotinib treatment in the ENEST1st trial (NCT01061177) as determined by dPCR or qPCR. Methods: 230 cDNA samples with e13 or e14/a2 BCR-ABL fusion genes were analysed in Leipzig (L, n = 75) or Mannheim (M, n = 155) by qPCR between 2012 and 2013. The cDNA samples were re-analysed in L by dPCR using a Droplet Digital PCR System (BIO-RAD). To assess the linearity of the assay, the ERM standard was diluted to BCR-ABL and ABL CN from 150000 to 1 and measured in replicate. Results by dPCR were about 20% lower than the calculated dilution. Linearity could be shown for CN as high as 120000 for ABL and BCR-ABL with a fraction of 81-165 negative droplets at the highest concentration. Results: The median CN of BCR-ABL and ABL was 12 (range 0-2050) and 59350 (range 7690-176000) by dPCR and 10 (range 0- 1529) and 53537 (range 4013 – 250800) by qPCR, respectively. Both methods detected similar numbers of BCR-ABL+ samples (dPCR 186, qPCR 189) with a median % BCR-ABL of 0.022 by dPCR compared to 0.019 by qPCR before, and 0.014 after, conversion to IS.he cumulative rates of MMR, MR4 and MR4.5 or better at 18 months of treatment were 83, 43 and 29% with qPCR. The distribution in MR classes was significantly different between the methods (p < 0.001), with dPCR indicating decreased cumulative rates of MR3, MR4 and ≥ MR4.5. Significantly fewer pts achieved ≥ MR4 when analyzed by dPCR (n = 77 vs. 100, p < 0.05). Of the 191 samples scored ≥ MMR by qPCR, 71 (37%) were one or more MR classes higher by dPCR. In contrast, only 13% were scored into a deeper MR class. Conclusions: dPCR is linear, even at high CN, and compared to qPCR, generates differing rates of MR in patients with CML .