Patients’ characteristics
A total of 99 patients (45.0%) harboring single CEBPA mutations and 121 patients (55.0%) with biallelic double mutations were identified. Compared with CEBPAsm AML patients, CEBPAdm AML patients had lower platelet counts (median: 29×109/L vs. 43×109/L, P=0.002), higher hemoglobin levels (median: 93 g/L vs. 82 g/L, P=0.0162), and lower WT1 expression levels (median: 12.385% vs. 42.35%, P=0.000). A total of 52.07% (63/121) of CEBPAdm AML patients presented with the FAB-M2 subtype, which occurred at a distinctively higher rate than that in CEBPAsm patients (52.07% vs. 33.33%, P=0.0053). AML patients with CEBPAsm were more likely to show the M5 subtype than those with CEBPAdm (41.4% vs. 17.35%, P=0.0001). No significant differences were identified regarding sex, age, white blood cell (WBC) count or cytogenetic karyotype between the CEBPAsm and CEBPAdm AML groups. Table 1 lists different categories of clinical information of the enrolled patients.
Frequency and types of WT1 mutations
Overall, 30 WT1 mutations were detected in 29 of the 220 patients (13.18%) screened. WT1 mutations occurred significantly more frequently in CEBPAdm AML (21/121, 17.36%) than in CEBPAsm (8/99, 8.08%) AML patients (P=0.043). Mutations clustered overwhelmingly in exon 7 (16 mutations in 15 patients), but they were also detected in exon 8 (n=6) and exon 9 (n=8). One patient presented with two kinds of mutations (no. 29). The amino acid changes and mutation information of the WT1 gene are shown in Fig. 1 and Table 2.
Association of WT1 mutations with clinical characteristics
WT1-mutated (WT1mut) patients had lower platelet counts (median: 25×109/L vs. 38×109/L, P=0.0481), higher WT1 expression levels (median: 29.72% vs. 16.81%, P=0.0371), and a negative trend for the M5 subtype (median: 17.24% vs. 29.84%, P=0.07) compared to WT wild-type (WTwt) patients. No significant differences were identified regarding sex, age, WBC count, hemoglobin, or cytogenetic karyotype between the WT1mut and WT1wt groups. The clinical characteristics are shown in table 1.
Companion gene mutations
Based on NGS, we detected 59 mutated genes by screening the 112-gene panel. In addition to WT1, 33 of the 59 genes could be categorized as signaling pathways, epigenetic regulators, transcription factors, spliceosomes, tumor suppressors, chromatin modifiers, and cohesin. A total of 94.54% (208/220) of CEBPA mutations were accompanied by additional mutations and frequently co-occurred with FLT3-ITD (n=39, 17.73%), NRAS (n=34, 15.45%), GATA2 (n=34, 15.45%), RUNX1 (n=30, 13.64%), WT1 (n=29, 13.18%), FAT1 (n=29, 13.18%), TET2 (n=23, 10.45%), CSF3R (n=20, 9.09%), KIT (n=21, 9.54%) and NOTCH2 (n=18, 8.18%). Other genes had a mutation prevalence of <8%. Functionally, mutations involved in signaling pathways (n=123, 55.91%) and transcription factors (n=47, 21.36%) were found frequently in CEBPAmut AML patients.
Among WT1-mutated patients, the most common co-mutation was FLT3-ITD (n=7, 24.14%), followed by NRAS (n=5, 17.24%), CSF3R (n=4, 13.79%), GATA2 (n=4, 13.79%), KIT (n=4, 13.79%), and FAT1 (n=3, 10.34%). Fewer common co-mutations (2%–10.25%) were found in NPM1, SF3B1, NOTCH1/2, PTPN11, JAK2 and others. The most frequent functional pathway was signaling pathways in as many as 62.07% of cases. Notably, no mutations in TET2, IDH1/2, or DNMT3A were identified in WTmut patients. The concomitant mutations in epigenetic regulators were inversely correlated with WT1 mutations (P=0.003). There were no differences in the incidences of FLT3-ITD, NRAS, CSF3R, GATA2, PTPN11, SF3B1 and NOTCH1/2 between WTmut and WTwt patients. The difference in the mutational spectrum between WT-mutated and WT1 wild-type AML is shown in table 3. The frequencies of the detected gene mutations and gene functional groups are shown in Figure 2.
Prognostic impact of WT1mutations in the context of CEBPA double mutations
Of 121 patients with CEBPA double mutations, 111 patients were evaluable for complete remission (CR) after the first induction therapy. No differences were observed in the CR rate between patients with or without WT1 mutations in CEBPAdm AML (89.5% vs. 91.3%, P=1.000). The median follow up was 30.5 (range 0.5-60.6) months. We next investigated the impact of WT1 mutations on survival. WT1-mutated patients had significantly inferior RFS, EFS, and OS than WT1 wild-type patients (3-year RFS rate, 40% vs. 70%, median, 8.9 months vs. NR, P=0.002; 3-year EFS rate, 36% vs. 65%, median, 9.9 months vs. NR, P=0.004; 3-year OS rate, 48% vs. 72%, 20.4 months vs. NR, P=0.010) (Fig. 3a/b/c and Table 4).