In the present study, the prevalence of positive staining for P-glycoprotein using C219 monoclonal antibody was assessed in 58 tissue samples of high-grade lymphoma from dogs before initiation of chemotherapy. Samples were also evaluated at relapse in 22 dogs, at necropsy in 34 dogs, and at all 3 times in 15 dogs. The frequency of positive staining was significantly higher than that found prior to the initiation of chemotherapy at the following times: relapse (P = .0001), necropsy (P < .0001), and both relapse and necropsy (P < .001, sequential data). The frequency of positive staining prior to the initiation of chemotherapy was significantly inversely related to remission (P < .001) and survival times (P = .0012). Similarly, when populations below and above the median initial C219 score were compared with respect to remission and survival times, the population with scores greater than the median had significantly lower remission (P < .001) and survival (P = .008) times, respectively. The frequency of positive staining determined at relapse was significantly inversely related to the time from relapse to death (P = .0102). Similarly, when populations below and above the median relapse C219 score were compared with respect to the time from relapse to death, the population with C219 scores greater than the median had a significantly lower time from relapse to death (P = .006). It appears that this immunohistochemical methodology may be used as a predictor of remission time, survival time, and the time from relapse to death. Additional studies are required to confirm the usefulness of C219 as a true marker of P-glycoprotein and to evaluate P-glycoprotein as a useful prognostic factor in dogs with lymphoma.