Abstract The most promising techniques for detecting minimal residual disease (MRD) in canine lymphoma are flow cytometry (FC) and polymerase chain reaction amplification of antigen receptor genes (PARR). However, the agreement between these methods has not been established. MRD was monitored by FC and PARR following treatment of dogs affected with diffuse large B-cell lymphoma (DLBCL), comparing results in lymph node (LN), peripheral blood (PB) and bone marrow (BM) samples. The prognostic impact of MRD on time to relapse (TTR) and lymphoma-specific survival (LSS) was also assessed. Fourteen dogs with previously untreated DLBCL were enrolled into the study; 10 dogs eventually relapsed, while four dogs with undetectable MRD were still in remission at the end of the study. At diagnosis, the concordance rate between FC and PARR was 100%, 78.6%, and 64.3% for LN, PB and BM, respectively. At the end of treatment, the agreement rates were 35.7%, 50%, and 57.1% for LN, PB and BM, respectively. At least one of the follow-up samples from dogs experiencing relapse was PARR+; conversely, FC was not able to detect MRD in seven of the dogs that relapsed. PARR was more sensitive than FC in predicting TTR, whereas the combination of PARR and FC was more sensitive than either technique alone in predicting LSS using PB samples. The results suggest that immunological and molecular techniques should be used in combination when monitoring for MRD in canine DLBCL.