Relation between measurable residual disease and morphologic relapse in AML.

Abstract

7535 Background: Measurable residual disease (MRD) in AML portends a poor prognosis. The outcomes and treatments of MRD after an initial MRD negative complete remission (MRD- CR) are unclear. Methods: We retrospectively identified 432 patients ≥ 18 years of age treated for AML or high grade myeloid neoplasm (10-19% blasts in blood and/or marrow) at University of Washington/Seattle Cancer Care Alliance from 2008-2017 who achieved MRD- CR after initial treatment. Next disease recurrence was recorded, with patients either developing MRD ( < 5% blasts via multiparameter flow cytometry; n = 44) or developing morphologic relapse (≥5% blasts; n = 100). The remaining patients remained in MRD- CR (n = 288, median follow up time 3.7 years). A landmark analysis at one year was performed to compare overall survival (OS). Results: Patients who developed MRD tended to be older (p = 0.009), but baseline characteristics were otherwise similar. Therapies for MRD included allogeneic transplant, low intensity chemotherapy, and high intensity chemotherapy; no significant associations were found between type MRD directed therapy and survival. Landmark OS at 1 year after MRD- CR was significantly different for patients without relapse at 1 year compared to those with MRD or morphologic relapse (median OS was 8.5 years for no relapse, 2.2 years for MRD, and 1.0 years for morphologic relapse). A multivariable Cox regression model among patients alive at 1 year showed patients without relapse had significantly improved OS compared to those with morphologic relapse [HR 0.18 (95% CI 0.1-0.31)]. Although there was a trend towards improved OS, we did not identify a significant difference between patients with MRD relapse compared to morphologic relapse [HR 0.54 (95% CI 0.27-1.11)]. There were no significant differences in patient characteristics for patients with MRD vs morphologic relapse matched by time of relapse (days 30-90, 90-150, 150-210). Conclusions: Following MRD- CR, development of either MRD or morphologic relapse were both associated with decreased OS. Notably, no significant differences in terms of survival were seen between patients who presented with MRD as opposed to morphologic relapse. No clear predictors were identified for MRD vs morphologic relapse. In this cohort, development of MRD carried a poor OS prognosis, similar to the prognosis of developing morphologic relapse. Future clinical trials should focus on MRD directed therapies, as no consensus exists about optimal treatment of MRD.