Comparison Of Different Pretransplant Predictive Scores In Patients With Refractory Acute Myeloid Leukemia After Allogeneic Stem Cell Transplantation Including Highdose Melphalan: Results Of a Double-Center Observational Study

Patients (pts) with refractory acute myeloid leukaemia (AML) have a particularly dismal prognosis using conventional supportive therapy and chemotherapy. In this double-center retrospective observational study, sequential high-dose melphalan (HD-Mel) as part of the conditioning regimen for allogeneic hematopoietic stem cell transplantation (aSCT) was administered in 162 adult pts (median age 55, range 17 to 71 years) with overt refractory AML. For every patient pretransplant assessment of transplantation (PAM) score, European BMT (EBMT) score, and Hematopoietic Cell Transplantation-Comorbidity Index (HCT-CI) was calculated. Furthermore, by adjusting the threshold of 30 points PAM score was stratified into low and high risk group. 107 pts (66%) underwent a total body irradiation (TBI) based conditioning regimen and 55 pts (34%) a chemotherapy based regimen after a median interval of 5 days (range 1-18 days) following HD-Mel application. Dose of melphalan was 140 mg/m² in 150 pts (93%), and 200 mg/m² in 12 pts (7%). 48 pts (30%) were transplanted with an identical sibling donor (ISD), and 114 pts (70%) with matched unrelated donors (MUD). Pts had mostly intermediate (51%) or adverse (46%) cytogenetic risk subgroups.

In 150 pts (93%) a stable engraftment was observed and a complete chimerism could be detected in 124 pts (83%). Acute GvHD grades 0-I developed in 98 pts (60%), whereas higher grades II-IV occurred in 64 pts (40%). The median calculated PAM score was 31 (range 22-89). The corresponding median overall probability of death within 2 years after transplant was 70 % (range 44-95%). Median HCT-CI was calculated with 2 (range 0-9) characterizing a low risk comorbidity. Finally the European BMT (EBMT) score was calculated with more than 5 (range 3-7) in 166 pts (72%) transplanted in this cohort. The cumulative incidence of treatment failure estimate was 68% (95%-CL: 60-76%) after transplant resulting in a cumulative relapse estimate of 35% (95%-CL: 28-43%), and an overall non-relapse mortality (NRM) of 38% (95%-CL: 30-46%). After a median follow-up of 35 months after aSCT among surviving pts, the 3-year overall survival estimate (OS) was calculated 37% (95%-CL: 35-45%), and 5-year OS 30% (95%-CL: 25-40%). Time-dependent multivariate analysis on the primary endpoints OS, relapse-free survival (RFS), relapse incidence, and NRM demonstrated an influence of chronic GvHD on EFS (hazard ratio [HR] 0.443, p=0.0003), NRM ([HR] 0.369, p=0.0032), and relapse incidence ([HR] 0.554, p=0.0467). Furthermore, MUD donors tended to have a favourable impact on relapse risk ([HR] 0.590, p=0.0595) compared to sibling donors. PAM score demonstrated significant influence on OS (p=0.0381) and RFS (p=0.0121). For low and high risk groups stratified PAM score was shown to have significant influence on OS ([HR] 1.175, p=0.0056), RFS ([HR] 1.003, p=0.0160), and NRM ([HR] 1.283, p=0.0049), respectively. No influence could be observed for EMBT score and HCT-CI on none of the primary endpoints. In addition the factors age and cGvHD were found to have significant influence on relapse incidence (p=0.011, and p=0.0051) in multivariate analysis.

The results of this large study confirm an association between PAM score risk stratification and outcome after aSCT following sequential conditioning regimen with HD-Mel in patients with refractory AML. We could show, that by using a stratified PAM score with the threshold of 30 points a separation between prognostic risk groups is possible. We can summarize that this protocol is a feasible therapy option for patients with refractory AML.