Findings

In acute promyelocytic leukemia (APL), PML-RARa transcript is used as a marker for minimal residual disease (MRD), but the marker is not useful for pre-emptive management since its positivity directly indicates relapse. High Wilms’ tumor gene 1 (WT1) expression was related with subsequent relapse in acute myeloid leukemia, and Hecht et al. recently reported that high initial WT1 expression was associated with more relapse in APL [1–3].

We confirmed APL by chromosome analysis and PML-RARα reverse transcriptase polymerase chain reaction (RT-PCR) method. All were treated with idarubicin (12 mg/m2, days 1, 3, 5, and 7) and all-trans retinoic acid (ATRA; 45 mg/m2/day) [4, 5]. After achievement of hematological complete remission (CR), all received three courses of consolidation—first, idarubicin (7 mg/m2, days 1–4); second, mitoxantrone (10 mg/m2, days 1–4); and third, idarubicin (12 mg/m2, day 1–2)—followed by 2-year maintenance using 6-mercaptopurine (50 mg/m2/day) plus ATRA [5–7]. The molecular studies were performed at diagnosis and 1 month after chemotherapy, and every 3 months after maintenance. Quantification of PML-RARα and WT1 were performed using the real-time quantitative (RQ)-PCR methods (Real-Q PML-RARα quantification kit, Biosewoom, Korea, and WT1 ProfileQuant™ kit, Ipsogen, France) presenting a similar sensitivity of 4.5 log.

We initially identified 142 APL patients from 2009 to 2014 but finally focused on 117 patients (median age 44 years old (range 19–70 years)) who underwent at least 2 cycles of consolidation after hematological CR. All patients were in complete molecular response (CMR) at the time of enrollment (Additional file 1: Figure S1, Table 1). Relapse was identified in 16 (13.7%) patients with a median duration of 22.8 months (range, 4.3–64.0). After median follow-up of 46.0 months (range, 14.7–86.3), 4-year cumulative incidence of relapse (CIR), non-relapse mortality (NRM), disease-free survival (DFS), and overall survival (OS) rates were 16.2, 1.2, 82.6, and 92.5%, respectively. We identified that high-risk Sanz-criteria, peak leukocyte count >40.0 × 109/L, and FLT3 mutation were predictive for relapse.

Table 1 Baseline characteristics of enrolled patients
Full size table

We compared the level of WT1 between relapsed and non-relapsed group during the course of treatment (Additional file 1: Figure S2) and identified that median WT1 was significantly different at post 2nd consolidation (171.5 vs. 76.3, P = 0.049), at post 3rd consolidation (156.0 vs. 67.6, P = 0.013) and at 3 months post-maintenance (162.0 vs. 59.1, P = 0.002). We found that WT1 post-maintenance 3 months was the most significant parameter for relapse prediction at the cutoff of ≥120.0 copies/104 ABL.

We calculated subsequent CIR and DFS rates in 116 patients after excluding 1 patient with early relapse. Patients with WT1 post-maintenance 3 months higher than 120.0 copies/104 ABL showed higher 4-year CIR (30.5 vs. 6.9%, P = 0.0002) and inferior 4-year DFS (62.8 vs. 91.4%, P < 0.0001) rates (Fig. 1a, b). Also in the high-risk subgroup, high WT1 post-maintenance 3 months showed higher 4-year CIR (43.3 vs. 11.1%, P < 0.0001) and inferior 4-year DFS (55.5 vs. 86.4%, P = 0.0015) rates (Fig. 1c, d). In FLT3 positive and negative subgroup, high WT1 post-maintenance 3 months showed higher 4-year CIR (51.4 vs. 0.0%, P < 0.0001 and 21.5 vs. 8.6%, P = 0.0434) and inferior 4-year DFS (46.7 vs. 100.0%, P = 0.0018 and 69.6 vs. 89.3%, P = 0.0154) rates (Fig. 1e, f).

Fig. 1
figure 1

Treatment outcomes according to WT1 expression level (<120 vs. ≥120 copies/104 ABL) at 3 months post-maintenance (PM-WT1). a Four-year CIR rates. b Four-year DFS rates. c, d Four-year CIR and DFS rates according to WT1 expression level in the high-risk subgroup. e, f Four-year CIR and DFS rates according to the status of PM-WT1 and FLT3-ITD mutation

Full size image

Multivariate analysis (Additional file 1: Table S1) revealed that 4-year CIR was significantly higher in patients with high peak leukocyte count (HR = 6.414; 95% CI, 2.1–19.3, P < 0.001) and high WT1 post-maintenance 3 month (HR = 7.533; 95% CI, 2.3–24.8, P < 0.001), and 4-year DFS was significantly inferior in patients with high peak leukocyte count (HR = 5.275; 95% CI, 1.9–14.7, P = 0.001) and high WT1 post-maintenance 3 month (HR = 8.241; 95% CI, 2.3–29.1, P = 0.001).

Unfortunately, our chemotherapy regimen was not differently specified for high-risk APL and the standard treatment of APL is now changed to a combination therapy using ATO. Therefore, current results may not be applicable in the treatment course using ATO and another validation is needed. Conclusively, high post-remission WT1 expression is a reliable marker for prediction of subsequent relapse in APL patients treated with conventional chemotherapy. For patients with high-risk of relapse, early intervention using WT1-specific therapy may prevent relapse and improve survival outcomes [8, 9].