Complete Remission with Incomplete Count Recovery (CRi) Prior to Allogeneic Hematopoietic Cell Transplantation for Acute Myeloid Leukemia Is Associated with a High Non-Relapse Mortality without Increased Relapse Risk

Background:

For the majority of patients with acute myeloid leukemia (AML), allogeneic hematopoietic cell transplantation (HCT) offers the best prospect of long-term survival. However both the risk of relapse and non-relapse mortality remain stubborn barriers to successful outcome. Remission status is one of the strongest predictors of outcome, and while conventional risk stratification models account for this, complete remission with incomplete count recovery (CRi) is typically ascribed the same predictive value as complete remission (CR).

Methods:

In this single center retrospective study we assessed the impact of incomplete count recovery on outcome of allogeneic HCT for AML. Patients' disease status was classified immediately prior to HCT as complete remission (CR), complete remission with incomplete count recovery (CRi) or no complete remission (no CR/CRi). The conventional definition of complete remission (<5% marrow blasts by morphology) was employed and incomplete count recovery was defined as platelet count <100 x 10⁹/L and/or absolute neutrophil count < 1 x 10⁹/L. Measurable residual disease (MRD) was defined by detectable disease by a contemporaneously accepted standard methodology at the time of transplant (e.g. PCR or immunophenotyping). The primary outcome measure was overall survival, and secondary endpoint of non-relapse mortality and relapse were also assessed.

Results:

Between January 2005 and December 2017, 155 patients underwent allogeneic HCT for AML in our institution. Disease status was defined as CR in 82 (53%), CRi in 54 (35%) and no CR/CRi in 19 (12%). No significant demographic or transplant characteristics were observed between the groups. With a median follow-up of 3.4 [0.1 - 12.8] years, survival at 3 years was 42.5% for the whole cohort with significant differences between the three groups: 54.3% for CR, 34.8% for CRi and 13.4% for no CR/CRi (p<0.001; fig 1A; HR 1.94 for CRi, and 3.4 for no CR). These differences remained after corrections for standard transplant risk factors. The significant differences in overall survival between CR and CRi groups were accounted for by increased non-relapse mortality (7.5% vs. 24.2% at 100 days and 28.7 vs 44.9% at 3 years (p=0.003), fig 1B). The overall cumulative incidence (CI) of relapse at 3 years was 32.4%, and while those patients transplanted not in CR/CRi has a significantly increased CI of relapse (55.3%, p=0.006), there were no significant differences between the CR and CRi groups (25.1 and 36.7% respectively (p=0.441; fig 1C). Patients without MRD (n=107) had a lower incidence of relapse compared to those with MRD (n=48, 21.9% vs. 48.1%, p=0.009), however MRD negative patients with incomplete recovery still had a significantly worse 3-year survival than those with complete recovery (37.5 vs 56.7, p=0.006) resulting from higher NRM (46.4% vs 30.4%%, p=0.003, fig 1D).

Discussion:

CRi patients have a poorer survival and higher NRM than CR patients with full marrow recovery, without evidence of increased relapse. While death from NRM is a competing risk for relapse, the increased incidence of relapse remains overtly demonstrable in those patients transplanted not in CR/CRi, despite an equally high NRM. Additionally, the persistently increased NRM in CRi patients without MRD, suggests that the poorer survival may result from poor fitness for transplant and increased susceptibility to transplant related complications rather than an imminent relapse. With the advent of increasingly sensitive MRD technologies, these data may suggest that further pre-transplant optimization to promote marrow recovery may be permissible in order to minimize NRM without risking increased risk of relapse in CRi patients.

Figure 1.

View largeDownload slide

Figure 1.

View largeDownload slide

Close modal