Combination of NCCN Risk Stratification System and Pre-Transplant Minimal Residual Disease Levels for the Prognosis of Acute Myeloid Leukemia after Allogeneic Hematopoietic Stem Cell Transplantation

Background

Numerous studies have confirmed that National Comprehensive Cancer Network (NCCN) risk stratification or pre-transplant minimal residual disease (MRD) levels can predict the risk of recurrence and survival after transplantation. But it is unclear whether combining these two parameters can more accurately predict prognosis.

Methods

We retrospectively analyzed 85 patients with acute myeloid leukemia (AML) who underwent allogeneic hematopoietic stem cell transplantation (allo-HSCT) and constructed a new risk stratification tool combining NCCN risk stratification and pre-MRD levels. All patients were grouped by NCCN risk stratification (favorable/intermediate prognosis and poor prognosis group), pre-MRD levels (MRD (-) group (<0.1%) and MRD (+) group (≥0.1%)) and a combination of the above (low, intermediate and high risk groups), and graft-versus-host disease (GVHD) and prognosis were compared between groups.

Results

Relative to the favorable/intermediate prognosis group, OS and RFS were poorer in the poor prognosis group (71% vs 82%, P= .156; 60% vs 74%, P= .101) and CIR (29% vs 20%, P= .229) and NRM (23% vs 14%, P= .200) were better. The incidence of aGVHD and cGVHD was slightly lower in the favorable/intermediate prognosis group than in the poor prognosis group (38% vs 46%, P= .415; 10% vs 11%, P=. 572). Relative to the MRD (+) group, the MRD (-) group had significantly better OS and RFS (89% vs 59%, P= .002; 79% vs 50%, P= .003), lower CIR and NRM (15.1% vs 37.5%, P= .011; 11.3% vs 28%, P= .040), and a lower incidence of cGVHD (6% vs 19%, P= .022). The new risk stratification tool stratified patients into low, intermediate and high risk groups. Patients in the high-risk group had the highest incidence of aGVHD and cGVHD (42% vs 35% vs 53%, P= .606; 6% vs 11% vs 20%, P= .157). The difference in cGVHD between the low- and high-risk groups was significant (P= .038). Three-year OS was 93.9%, 70% and 60% (P= .011) and RFS was 85%, 62% and 46.7% (P= .009) for low-, intermediate- and high-risk patients, respectively. The differences in OS and RFS between the low- and intermediate-risk groups were statistically significant (P= .010; P= .025), as were the differences in OS and RFS between the low- and high-risk groups (P= .001; P= .001). Patients in the high-risk group had the highest CIR and NRM relative to those in the low- and intermediate-risk groups (9% vs 32% vs 33.3%, P= .027; 6% vs 24.3% vs 26.7%; P= .059). The differences in CIR (P= .012) and NRM (P= .028) were statistically significant in both the low-risk and intermediate-risk groups, as well as in the low- and high-risk groups (CIR: P= .028; NRM: P= .021). Multivariate analysis indicated that time to ANC recovery, time from diagnosis to transplantation, and novel risk stratification were independent prognostic factors.

Conclusions

Both pre-MRD levels and NCCN risk stratification predict AML prognosis after allo-HSCT, and combining the two can more accurately predict post-transplant prognosis.