Measurable Residual Disease Assay with WT1 Expression in Acute Myeloid Leukemia Who Underwent Allogeneic Hematopoietic Stem Cell Transplantation; Optimal Threshold, Time Points, and Candidates

Background

Overexpression of WT1 is a surrogate marker of abnormal myelopoiesis and has been evaluated as a potential tool to assess measurable residual disease (MRD) in myeloid malignancies. Given that lack of consensus on clinically relevant WT1 thresholds and time points in the allogeneic hematopoietic cells transplantation (allo-HSCT) setting, WT1 quantification has not yet gained widespread use despite several pieces of evidence demonstrating the possible role for MRD assessment with the limited numbers of patients. To investigate optimal threshold, time points, and candidates of WT1 quantification in AML, we retrospectively analyzed a large cohort of consecutive patients who underwent allo-HSCT at Catholic Hematology Hospital.

Patients and methods

This study included 425 consecutive patients with AML who underwent allo-HSCT at CR state from either a matched siblings (n=199), matched unrelated (n=117) or haploidentical family donors (n=109) from 2012 to 2016. Patients were in the first (n=400) or second (n=25) complete remission with a median age of 48 years (range, 18~70). Favorable, intermediate, poor risk groups by 2017 NCCN criteria were 28% (n=120), 49% (n=206), and 23% (n=99), respectively. Bone marrow WT1 levels before, and at 1 or 3 months after allo-HSCT were determined using real-time PCR using the ELN normalized method. We sought to clarify the prognostic relevance of the WT1 quantification regarding the cumulative incidence of relapse (CIR) and survival outcomes.

Results

With a median follow-up of 39 months (range, two days to 73 months), the 4-year overall survival, disease-free survival, CIR and non-relapse mortality were 63.6%±2.6%, 61.5%±2.6%, 17.9%±2.1% and 24.7%±2.5%, respectively.

Analysis of dynamic changes of WT1 levels demonstrated decreased levels at 1 (n=333, mean 86 copies, range 0~1800) and three months (n=346, mean 101 copies, range 0~1670) after allo-HSCT compared to before allo-HSCT (n=425, mean 219 copies, range 0~9630). Relapsed patients had significantly higher WT1 levels before (P=0.018) and at three months (P=0.041) after allo-HSCT, whereas no difference at one month after allo-HSCT (P=0.167). Even the ROC curve analysis revealed that WT1 levels before allo-HSCT were significantly available to predict CIR after allo-HSCT (P<0.001). Among various cut-off levels of WT1 expression (median, 25% from the top, and cut off by ELN), cutoff by ELN (250 copies) was most effective for predicting CIR. The CIR of MRD positive patients (³ 250 copies) before and at three months after allo-HSCT were 43% (vs. 14%, P<0.001) and 35% (vs. 11%, P<0.001), respectively. In multivariate analysis, the WT1-MRD positivity independently predicted the CIR (before, HR=3.5, P<0.001; at three months, HR=7.4, P<0.001), which translated into inferior disease-free survival (P<0.001) and overall survival (P<0.001). In a subgroup analysis with the WT1-MRD positive patients before allo-HSCT (n=44), the WT1-MRD positivity at three months was significantly effective to identify patients with a higher risk for relapse (100% vs. 26%, P<0.001).

In subgroup analyses in each risk group by 2017 NCCN criteria, the WT1-MRD positivity before allo-HSCT was significantly effective to predict CIR in the intermediate risk group (57% vs. 12%, P<0.001), whereas no significance in both favorable and poor risk groups. On the other hand, the WT1-MRD positivity at three months after allo-HSCT was effective to predict CIR in the poor risk group (60% vs. 21%, P<0.001). In patients with normal karyotype without NPM1 mutations (n=117), the WT1-MRD positivity before allo-HSCT significantly predict CIR (32% vs. 8%, P=0.001), whereas no difference in patients with NPM1 mutations (n=66) or core-binding factor (CBF) AML (n=102).

Conclusions

These data suggest standardized bone marrow WT1 levels using the ELN threshold (250 copies) before and at three months after allo-HSCT provided relevant information to predict relapse in AML with intermediate and poor risk groups by 2017 NCCN criteria, respectively. The validated WT1 MRD assay by ELN was revealed to be particularly available in AML without specific MRD markers, such as NPM1 or CBF-AML, and different significance by times points should be considered for clinical applications to identify high-risk AML for relapse, potential candidates for various immunomodulatory approaches.