The Predictive Factors of Favorable Prognosis after Allo-HSCT for Refractory Acute Leukemia

Background:

For patients with chemotherapy refractory acute leukemia (refractory AL), allogeneic hematopoietic stem cell transplantation (allo HSCT) is the only treatment that has curative potential. However, the long term overall survival (OS) of patients with refractory AL who received allo HSCT was reported to be less than 25% from CIBMTR. In contrast, the OS of patients who received allo HSCT on complete remission (CR) was reported to be approximately 60%. Although some refractory AL patients can achieve CR and survive for a long time, their characteristics remain to be elucidated. These are required in order for efficient use of transplant resources to find the subpopulation of refractory AL patients who can receive larger benefits from allo HSCT.

Objectives:

The purpose of this study was to find the predictive factors of favorable prognosis among patients with refractory AL prior to allo HSCT.

Patients and Methods:

Fifty refractory AL patients (AML n=40, ALL n=10), who underwent allo HSCT between January 2000 and January 2016 at Niigata University Hospital and Nagaoka Red Cross Hospital, were analyzed retrospectively. The median age of the patients at the time of allo HSCT was 38 years (range: 18-64), and the median follow up period was 8.3 months (range: 0.5-144.6). All patients were evaluated with bone marrow (BM) aspiration within one month prior to allo-HSCT. Thirty-two patients received myeloablative conditioning and 18 patients received reduced-intensity conditioning. Donor sources were siblings (n=12), unrelated (n=16), haploidentical (n=12) or cord blood (n=10). According to the NCCN guidelines of cytogenetic risk status, AML patients were classified as low risk (n=2), intermediate risk (n=22) or high risk (n=16). ALL patients were classified as standard risk (n=7), high risk (n=2) or unknown (n=1). Relapse free survival (RFS) and OS were estimated by the Kaplan-Meier method. Multivariate Cox regression was used to identify the independent prognostic factors.

Results:

The median blast percentage in BM before allo HSCT was 18.0% (range: 0.8-93.6%). Non-relapse mortality was 18.0%. The 1y-RFS and OS were 32.2% and 45.8%. The 5y-RFS and OS were 25.3% and 25.7%, respectively.

First, to predict relapse based on the optimal threshold value of blast percentage in BM, we calculated receiver operating characteristics (ROC) analysis and the largest areas under the curve (AUC). ROC analysis for blast percentage in BM and relapse revealed that ≤32% was the optimal threshold value (AUC 0.677) to predict relapse. Univariate analysis revealed that patients with HCT-CI ≦2 (p<0.01) and BM blast ≤32% (p=0.019) had significantly better RFS. In addition, HCT-CI ≦2 (p=0.035), BM blast ≤32% (p<0.01) and primary diagnosis (AML) (p<0.01) were also correlated with better OS. On the other hand, age, graft type, primary refractory AL or not, conditioning regimen and cytogenetic risk group had no influence on RFS and OS.

To investigate the correlation between multiple factors and the outcome, we scored the patients according to these three favorable factors (AML, HCT-CI≦2 and BM blast ≤32%), and stratified them into 3 groups as follows; score=3 (n=23), score=2 (n=20) and score=0-1 (n=7). The patients with score=3 exhibited better 1y-RFS compared with the other groups (p<0.001, 54.8%, vs. 6.9% (score=2) and 14.3% (score=0-1)). The 1y-OS and 5y-OS of patients with score=3 were also significantly better than patients with score=2 (68.5% and 52.3% vs 28.9% and 0%) or score=0-1 (14.3% and 0%), respectively (p<0.0001). The hazard ratio (HR) for RFS of patients with score=2 and score=0-1 were 5.02 (p<0.001) and 3.47 (p=0.013). The HR for the OS of patients with score=2 and score=0-1 were 5.12 (p<0.001) and 6.54 (p<0.001), respectively.

On multivariate analysis adjusted by age, cytogenetic risk, primary refractory or not, donor source and conditioning regimen, score=3 was an independent favorable prognosis factor for RFS (HR 0.17 (p<0.001)) and OS (HR 0.12 (p<0.00001)).

Conclusions:

We found three favorable prognostic factors (AML, HCT-CI≦2 and BM blast ≤32%) in this study. Refractory AL patients with all three factors may receive larger benefits from allo-HSCT and the 5y-OS may be comparable to allo HSCT with CR.