Bone Marrow Cellularity Is a Single Most Important Independent Prognostic Factor In AML Patients with t(8;21)

Abstract 1707

Core binding factor AML including t(8;21) and inv(16) have been associated with a relatively favorable prognosis compared with patients with normal or adverse karyotypes, and treated similarly. However, both t(8;21) and inv(16) AML seem to differ with respect to several biologic features and several reports demonstrated inferior outcome of t(8;21) compared with inv(16). Advanced age, higher WBC or granulocytic count, as well as CD56 expression or granulocytic sarcoma have been reported as poor prognostic factors in t(8;21) patients. Higher bone marrow (BM) blasts, lower platelets, and non-white race in t(8;21) AML adversely affected the probability to achieve CR. The KIT mutation is associated with poor prognosis in AML1-ETO-positive AML. Five-year survival rate was only around 40% in patients with t(8;21) having poor prognostic factors. Several chemotherapeutic strategies have been reported, among which high-dose cytarabine (HDAC) is generally the most effective option for successful postremission therapy. Furthermore, none of the randomized studies disclosed an advantage of allogeneic SCT (alloSCT) in this group of patients, given the relatively high treatment-related death (TRD) rate. Patients with t(8;21) AML with unfavorable prognosis may benefit from intensive postremission therapy such as early hematopoietic SCT. We conducted a retrospective study to investigate whether postremission therapies impact on survival according to prognostic factors in 132 AML patients with t(8;21) achieving first CR. Univariate analyses of prognostic factors for survival were performed in the patients with t(8;21), as well as more limited population of chemotherapy (CTx) group according to postremission therapies. The BM cellularity was a single most important independent prognostic factor on survival when using BM cellularity cutoffs as 90%. The 5-year overall survival (OS) in patients with t(8;21) and CTx group were significantly lower at 49.7% and 44.3% in patients with ≥ 90% BM cellularity, compared with 81.4% and 81.9% in those with < 90% BM cellularity, respectively (P = 0.001 and 0.027, respectively). The only other prognostic factor that influenced OS in CTx group was WBC count with cutoffs as 9.1 × 10⁹/L. High WBC count was trend towards poor OS in CTx group (P = 0.067). In multivariate analysis, BM cellularity appeared to be the only independent prognostic factor for OS in either AML patients with t(8;21) (P = 0.002) or CTx group (P = 0.055). Interestingly, we found positive correlation between BM cellularity and WBC count (P = 0.013), peripheral blood (PB) blast percentage (P = 0.001) and serum LDH level (P = 0.017) but not hemoglobin level and BM blast percentage in a linear regression model. And also, we confirmed negative correlation between BM cellularity and platelet count (P = 0.009). It is speculated that BM cellularity represents on poor prognostic factors including WBC and platelet counts, and PB blast percentage in patients with t(8;21). By combining dichotomized WBC count and BM cellularity in a univariate analysis for OS in CTx group, three risk groups could be established: low risk group, WBC count less than 9.1 × 10⁹/L and BM cellularity less than 90%; intermediate risk group, WBC count ≥ 9.1 × 10⁹/L and BM cellularity less than 90%; high risk group, BM cellularity ≥ 90%. In CTx group, 5-year OS was 81.9% in low risk group, 64.8% in intermediate group, and 32.1% in high risk group (P = 0.041). In alloSCT group, 5-year OS was 94.1% in low risk group, 29.1% in intermediate risk group, and 77.8% in high risk group (P = 0.042). In low risk group, 5-year OS was 81.9% in CTx group, 65.6% in autologous SCT (autoSCT) group, 94.1% in alloSCT group. In intermediate risk group, 5-year OS was 64.8% in CTx group, 29.1% in alloSCT group. In high risk group, 5-year OS was 32.1% in CTx group, 52.5% in autoSCT group, and 77.8% in alloSCT group. We found that BM cellularity was the most powerful independent prognostic factor in AML patients with t(8;21). The newly proposed model using BM cellularity and WBC count demonstrated a simple and valid measurement as main prognostic factor. We suggest a risk-adapted postremissin strategies based on this prognostic model for AML with t(8;21) such as low and intermediate risk patients receiving three cycles or more than three cycles of HDAC CTx and high risk patients undergoing SCT in first CR as postremission therapy.