Prognostic Implications of WT1 and BAALC in Adult Patients with Core-Binding Factor Positive or Normal Karyotype AML: Consecutive Analyses From Diagnosis to Hematopoietic Stem Cell Transplantation

Numerous molecular prognostic markers, in addition to specific chromosomal aberrations, have been identified to confer disease pathogenesis and survival outcomes in patients with adult acute myeloid leukemia (AML). Many efforts have been made to identify genetic mutations (e.g., c-kit, FLT3-ITD/TKD, NPM1, WT1, MLL-PTD) and modulated gene expression levels (e.g., BAALC and WT1) that can be used to predict outcomes in patients with cytogenetically normal AML (CN-AML). However, few data associated with core-binding factor AML (CBF-AML) and BAALC or WT1 expressions have been reported, and a consecutive analysis of BAALC and WT1 from diagnosis to hematopoietic stem cell transplantation (HSCT) has not yet been reported. We analyzed serial BAALC and WT1 data in patients with CN-AML and CBF-AML using the real-time quantitative polymerase chain reaction method.

In this single center retrospective study, 1008 newly diagnosed patients with AML (age, 15–85 years; median 45.2 years) with variable karyotypes were initially enrolled from January 2005 to December 2010. Ultimately, we analyzed 335 patients with CN-AML and 156 with CBF-AML. Patients that underwent conservative treatment were excluded. Among 491 patients, we identified FLT3-ITD results in 316 patients (positive in 56 and negative in 260) and NPM1 results in 202 patients (positive in 50 for CN-AML only), and obtained BAALC and WT1 results at the initial diagnosis in 177 patients (126 in CN-AML and 51 in CBF-AML) since we started these laboratory examinations in 2008. In 126 of the patients with CN-AML, 68 underwent HSCT and 58 were treated with chemotherapy alone. In the CBF-AML group, 42 were treated with HSCT and 9 were treated with chemotherapy alone. We analyzed BAALC and WT1 serial data at the initial diagnosis, after the first cycle of induction chemotherapy, and before and after HSCT. An additional analysis included the decreased ratio in the logarithm between diagnosis, pre-HSCT, or post-HSCT. Overall survival (OS) and relapse free survival (RFS) were associated with BAALC and WT1 expression levels along with the treatment timeline, and concomitantly with FLT3, NPM1, and c-kit mutations. Survival and other clinical correlates and prognostic relevance of mutations in patients with CN-AML and CBF-AML were evaluated using the log-rank test, Cox proportional hazard model, and chi-square analyses.

In both the CN- and CBF-AML groups, BAALC and WT1 levels at the initial diagnosis had no definite influence on OS when patients were treated with HSCT. In patients with CN-AML, only the chemotherapy-treated group with higher BAALC expression at diagnosis showed poor remission status (p = 0.008) and inferior OS (p = 0.013). Higher BAALC expression at diagnosis resulted in inferior OS in patients with CN-AML and the NPM1 (n = 7 vs. 38, p = 0.002) and FLT3-ITD mutations (n = 14 vs. 19, p = 0.052) in each group, as well as those with both mutations (n = 4 vs. 15, p = 0.000). In contrast, higher WT1 expression at diagnosis resulted in inferior OS in patients with CN-AML without the NPM1 (n = 91, p = 0.022) and FLT3-ITD (n = 92, p = 0.012) mutations in each group, as well as without both mutations (n = 69, p = 0.092). Higher pre-HSCT WT1 expression also resulted in inferior OS (p = 0.002) and RFS (p = 0.015) after HSCT in patients with CN-AML. Pre-HSCT BAALC expression levels and the decrease in the BAALC ratio before HSCT was not significant, but the decrease in the WT1 ratio over 1 logarithm prior to HSCT resulted in superior OS (p = 0.001) and RFS (p = 0.002) in patients with CN-AML. In patients with CBF-AML, the c-kit mutation-positive group showed inferior OS (p = 0.044), and their mean BAALC expression level was exclusively higher than that of other karyotypes (73.4 vs. 19.8, p = 0.000). However, only higher BAALC expression levels at diagnosis were associated with an increased relapse rate and shortened RFS after HSCT (p = 0.023) in our cohort. Other BAALC and WT1 expression level parameters were not significant in patients with CBF-AML.

The higher pre-HSCT WT1 expression and a decreased WT1 ratio before HSCT in patients with CN-AML, and higher BAALC expression at diagnosis in patients with CBF-AML resulted in significant changes in OS and RFS after HSCT. Therefore, the impact of combined molecular marker analysis on outcomes of allogeneic HSCT for AML should be considered intensively in order to produce a more defined HSCT plan.

No relevant conflicts of interest to declare.