Subgroup Analysis Can Optimize the Relapse-Prediction Cutoff Value for WT1 Expression After Allogeneic Hematologic Stem Cell Transplantation in Acute Myeloid Leukemia.

High WT1 expression after allogeneic hematologic stem cell transplantation (allo-HSCT) can strongly predict relapse in acute myeloid leukemia (AML). However, the cutoff values obtained have been inconsistent. Precise cutoff values may be optimized through subtype analysis; the RUNX1-RUNX1T1 fusion transcript provides an ideal reference. RUNX1-RUNX1T1 and WT1 transcript levels were simultaneously measured in 1299 bone marrow samples serially collected from 176 t(8;21) AML patients after allo-HSCT. The upper limit of the normal bone marrow WT1 level was 0.6%, which we previously reported to be the cutoff value for significant relapse prediction in AML as a whole. WT1 cutoff values of 0.6%, 1.2%, and 1.8% significantly differentiated patients in relapse after allo-HSCT. Nonetheless, patients with WT1 levels of 0.6% to 1.2% and those with levels of >1.2% and 1.8% after HSCT had rates of cumulative incidence of relapse similar to those with a continuous WT1 level of ≤0.6%, and both were significantly lower than that in patients with a WT1 level of >1.8%. WT1 expression was significantly related to RUNX1-RUNX1T1 transcript levels at WT1 levels of >1.8% but not at levels of 0.6% to 1.2% or >1.2% to 1.8%. Therefore, subgroup analysis can optimize the relapse-prediction cutoff value of WT1 expression. A cutoff level of 1.8% more accurately differentiates t(8;21) AML patients in relapse after allo-HSCT than does a cutoff level of 0.6%.