Bone Marrow Cellularity, but Not Dysplasia, Is An Additional Prognostic Factor for Patients with Acute Myeloid Leukemia After Allogeneic Stem Cell Transplantation

Relapse incidence (RI) and non-relapse mortality (NRM) are competing risks limiting overall survival (OS) after allogeneic stem cell transplantation (SCT) in acute myeloid leukemia (AML). Disease and transplant specific factors predicting relapse like measurement of minimal residual disease (MRD) and chimerism analysis are widely used to aid prophylactic and preemptive treatment decisions. Prediction of NRM mostly relies on pretransplant features. Although most transplant centers routinely perform bone marrow (BM) cytomorphology after SCT for AML, the impact of factors beyond blast count is not well studied.

We analyzed frequencies and prognostic impact of dysplasia and cellularity upon BM cytomorphology of 112 patients (60 m/52 f, median age 53 [range 17–72] years) with AML at 1^st manifestation/ relapse at day 30 (d30) and day 100 (d100) after SCT. Using peripheral blood as main graft source (n=106), donors were unrelated in 87 cases, related in 25. Conditioning was reduced (RIC, n=72) or myeloablative (MAC, n=40). All patients received G-CSF from day 5 until stable engraftment was achieved. Dysplasia was assessed following WHO criteria with different thresholds (10%, 20%, 50%) to define a hematopoietic lineage as dysplastic. We performed a correlation of dysplasia and age-adapted cellularity with outcome measures, calculating RI and NRM as competing risks. Only patients who achieved blast clearance on d30 after SCT were included in the study. Patients who developed hematological relapse between d30 and d100 were only evaluated for d30. At d30 (d100), BM aspirates from 75 (65) patients were available for morphologic evaluation.

Dysplasia was a frequent event both at d30 and d100, with ≥10% dysplastic features in granulopoiesis in 25.0% of cases at d30 (31% d100), in erythropoiesis in 34.6% of cases at d30 (43.6% d100) and in megakaryopoiesis in 47.7% of cases at d30 (63.5% d100). Overall, cellularity at d30 was increased in 17.3% (d 100: 6.5%), reduced in 37.3% (d100: 38.7), and normal in 45.3% (d100: 54.8%). No significant correlation with CMV reactivation or with the type of immunosuppression (cyclosporine/ methotrexate versus cyclosporine/ mycophenolic acid) was noted. Cumulative incidences of 2-year-RI and 2-year-NRM were 34% (95% CI, 24%-44%) and 17% (95% CI, 9%-25%). Dysplasia both at d30 and d100 did not correlate with OS or RI. Yet, a statistically significant correlation of normal overall cellularity at d30 with less relapses (RI 20.6%) when compared with reduced overall cellularity (RI 32.1%) or increased overall cellularity (RI 76.9%; p=0.001) was observed. Estimated 2-year-OS was 59% in pts with normal overall cellularity versus 31.4% (reduced) and 44.0% (increased), respectively (p=0.009). The same results, favoring normal cellularity, were observed for each lineage (granulopoiesis, erythropoiesis, megakaryopoiesis). Conversely, increased overall cellularity at d30 correlated with lower NRM (8.3%) when compared to normal (NRM 23.7%) and reduced overall cellularity (NRM 39.6%, p=0.031). Thus, whereas reduced overall cellularity at d30 correlated both with higher RI and higher NRM, the impact of increased cellularity on survival was less clear. The analysis of subdistributive hazards in the competing risk factor model revealed a cumulative RI of 62% (95%CI 35%-89%, HR 6.68, p=0.00014) for increased cellularity, making it the most potent hazard in this analysis. Presence of an informative sample was of prognostic value, too (2-year-OS/ NRM 54.7%/ 80.4% for “evaluable” versus 20%/ 36.9% for “not evaluable” due to low cellularity, p<0.001). Cellularity at d100 showed no significant correlation with survival outcomes. We found no correlation of either dysplasia or cellularity with the pretransplant cytogenetic risk group and CMV serostatus. In this study, patients with AML who achieved normal cellularity early in the post-transplant period had improved survival outcomes and a reduced relapsed incidence as compared to patients with abnormal cellularity in bone marrow aspirates.

These data suggest that cellularity of BM cytomorphology at d30 after allogeneic SCT aids to assess risk of relapse and NRM in transplant recipients with AML. At this time, it can only be speculated whether underlying persistent leukemia below the microscopic level might be associated with disturbed BM cellularity.

Haferlach:MLL Munich Leukemia Laboratory: Employment, Equity Ownership.