CD34+/TNC ratio predicts relapse in AML/MDS patients receiving peripheral blood matched allogeneic stem cell transplantation with post-transplant cyclophosphamide

Abstract Introduction. Allogeneic hematopoietic stem cell transplant (SCT) is an effective treatment for hematologic malignancies. The engrafted cells provide hematopoietic recovery and immune reconstitution and mediate the graft-vs-malignancy effect. We examined the cellular composition of the graft to determine how it impacts outcomes post-SCT in patients with HLA-matched donors and post-transplant cyclophosphamide (PTCy) based graft-versus-host (GvHD) prophylaxis. Methods. Patients who underwent their first matched unrelated (MUD) or related (MRD) SCT, with cryopreserved peripheral blood stem cells, between 2018-2023 at our institution and received PTCy-based GvHD prophylaxis were eligible for this study. The impact of CD34+ cells (x106/kg), TNC (x108/kg), and the ratio (x102) of CD34+ cells/kg/TNC/kg on post-SCT outcomes were evaluated by quartile analyses. Based on the distribution of each marker, we used the respective cut-off values that demarcate the 25th, 50th, and 75th percentiles to define 4 groups (G1-G4), and we compared outcomes across the 4 groups defined for each marker. The primary analysis was performed in the MUD cohort and the findings were then tested in the MRD cohort for validation. All analyses were performed accounting for competing risks as applicable. Results. A total of 805 consecutive patients underwent their first MUD (N=446) or MRD (N=359) SCT with PTCy during the study period. Of these, 344 were AML/MDS patients (including 215 MUD and 129 MRD SCT recipients) who received PTCy/tacrolimus/mycophenolate mofetil GvHD prophylaxis. We used the distribution of CD34+ cells, TNC, and the CD34+/TNC ratio in the 446-MUD cohort to define 4 groups (G1-G4) for each marker. The association between each marker and outcomes was then evaluated in the AML/MDS subset. The median (range) [25th and 75th percentile] CD34+ cells was 6.98 (0.6-25) [5.3, 8.3], median TNC was 7 (1.5-21) [5.3, 9.1], and median CD34+/TNC ratio was 0.97 (0.15-3.35) [0.68, 1.29] in the 446-MUD cohort. In the 344-patients AML/MDS subset, median (range) recipient age in the MUD and MRD cohorts was 65 (18-77) and 62 (20-78) years, respectively. Median Hematopoietic Cell Transplantation-specific Comorbidity Index (HCT-CI) was 3 (0-10) in both cohorts, 47% and 54% of patients had high-risk (defined by ELN17 for AML and IPSS-R for MDS) disease, and 59% and 62% received myeloablative conditioning, respectively. The median follow-up was 29 (range 3-79) months, overall. The 3-year progression-free survival (PFS) rates were 52% (95% CI: 45-59) and 58% (49-67) for the MUD and MRD cohorts, respectively, and disease progression rates were 30% (24-37) and 24% (17-33). In the MUD AML/MDS cohort (N=215), univariate analysis showed that the CD34+/TNC ratio was statistically a stronger predictor of outcomes compared with CD34+/kg or TNC/kg each considered separately. G1-G4 of the CD34+/TNC ratio were respectively defined as CD34+/TNC ≤0.68, >0.68-0.97, >0.97-1.29, and >1.29. One patient (in G1) had primary graft failure. The time to neutrophil and platelet engraftment decreased with increasing CD34+/TNC ratio. The median (range) time to neutrophil engraftement was 18 (12-33) vs 15 (12-24) days, and time to platelet engraftment was 30 (12-164) vs 17 (9-42) days in G1 vs G4. In multivariate analysis, the lowest ratio group (G1, N=56 vs G2-4, N=159) was associated with a lower PFS (30% vs 60%, HR=2.1, P<0.001) and a higher progression rate (HR=2.0, P<0.007). PFS and progression rates did not differ among G2-G4. Inferior outcomes in the lowest ratio group (G1) persisted irrespective of donor age (≤ vs > 35 years), gender, or CMV serostatus. Patient (age, gender, HCT-CI) and disease characteristics did not differ between G1 vs G2-4. There was no association between the CD34+/TNC ratio and non-relapse mortality or grade III-IV acute GvHD. Compared with G1-3, the highest ratio group (G4, N=48) was associated with higher (HR=2.9, P=0.001) rate of chronic GvHD. Consistent results were observed for disease progression, PFS, and chronic GvHD in the MRD cohort. Conclusions. These data suggest that the graft composition affects SCT outcomes. We identified a group of AML/MDS patients who received grafts with a low CD34+/TNC ratio and were at high-risk of relapse and inferior PFS. These patients should be considered for novel post-SCT treatments designed to reduce the risk of relapse, and the mechanisms of the graft composition effects on SCT outcome need to be delineated.