Clinical and molecular characteristics of NUP98 fusion genes
We screened a large pediatric cohort from the COG-National Cancer Institute (NCI) TARGET AML cohort. Among 885 AML patients, 57 had NUP98 rearrangements, with a total frequency of 6.4%. In this study, we identified six kinds of NUP98 fusion genes. The most common NUP98-fusion gene was NUP98/NSD1(59.6%, 34 of 57), followed by NUP98/KMD5A (17.5%, 10 of 57), NUP98/HOXD13 (7% ,4 of 57), NUP98/HMGB3, NUP98/HOXA9, and NUP98/PHF23 (5.3%, 4 of 57).
Comparison of the clinical and molecular characteristics of NUP98-positive and NUP98-negative patients revealed no significant differences in gender, median age, or median WBC count at diagnosis (Table 1). However, the NUP98/NSD1 fusion gene patients had greater median WBC of 100.8*10^9/L than did the NUP98/NSD1 negative patients (P=0.002) (Table 2). Regarding cytomorphology status, there were significant differences between the NUP98-positive group and the NUP98-negative group (50.9% vs. 2.3%, P<0.001) (Table 1).
Compared with NUP98-negative patients, NUP98-positive patients more often exhibited the FAB M7 phenotype (14% vs. 2.3%, P=0.001). We also found that NUP98 fusion gene cases highly overlapped with FLT3/ATD-positive cases (46% vs. 6.1%, P<0.001) (Table 1) and that NUP98-NSD1fusion gene cases are highly overlapped with FLT3/ATD positivity (64.7% vs. 6.7%, P<0.001) (Table 2).
Prognosis of NUP98 fusion genes carriers
We initially evaluated the impact of NUP98 fusion gene positivity on prognosis. In terms of treatment response (Table 1), 80.1% (378 of 472)of NUP98-negative patients achieved a favorable CR in course 1 which was greater than the 59.6%(34 of 57)NUP98-positive patients (P<0.0001). After the second course of induction therapy, CR was achieved in 19 of 57 (33.3%) patients with NUP98 gene fusion and in 236 (50%) of 472 patients without NUP98 gene fusion (P=0.017).
The survival data of all the 885 pediatric AML patients in the TARGET group and 529 patients who were gene fusion positive were analyzed. Survivors were followed up for a median of 5.6years among all 529 patients with and without the NUP98 fusion gene positivity. As shown in Figure 1(A), patients harboring a NUP98 fusion gene had an inferior EFS (32.4%) compared with wild-type patients (49.2%, P=0.002). Figure 1(B) illustrated a trend toward worse OS in NUP98 fusion gene patients(48.1%)than in NUP98 wild-type patients(67.0%, P=0.006).
Stratification analysis of FLT3/ITD and NUP98 fusion genes
NUP98 fusion gene patients had a high frequency of 45.6% (26 of 57) in cooperating genetic alterations with FLT3/ITD mutation. In the 55 cases of gene fusion positive, which carried FLT3/ITD mutation,the CR rate 1 in patients carried positive of NUP98 gene fusion was lower at 34.6% (9 of 26) compared to the rate of 79.3% (23 of 29) observed in patients without NUP98 gene fusion (P=0.0008),as well the CR rate 2 in dual positive of NUP98 gene fusion and FLT3/ITD was 42.3%(11/26)compared with cases without NUP98 gene fusion 89.7% (26 of 29) (P=0.0002)( Supplementary Table1).
As shown in Figure 2, with respect to the FLT3/ITD mutation, patients with NUP98-positive and FLT3/ITD mutation had the worst prognosis with a 5-year OS of 41.4% and a 5-year EFS of 13.5% compared with patients with wild-type FLT3/ITD without the NUP98 gene fusion patients (OS 68.0%, P=0.002; EFS 50.3%, P<0.0001). According to the NUP98-positive patients and FLT3/ITD status, the survival curves were grouped into 4 subgroups (Figure 2). In FLT3/ITD wild-type cases, NUP98-positive patients did not significantly affect 5-year OS (54.0±9.1% NUP98+ vs. 68.0±2.3% NUP98-; P = 0.187) or EFS (48.0±9.0% NUP98+ vs. 50.3±2.4% NUP98-; P = 0.862).
According to the NUP98 gene fusion status, FLT3/ITD-positive patients had a negative impact on the 5-year EFS rate (13.5± 7.0% NUP98+vs. 31.0 ± 8.6% NUP98-; P = 0.008) but not on the 5-year OS rate (41.4± 10.3% NUP98+vs. 55.0 ± 9.3% NUP98-; P= 0.133). A negative impact on the 5-year EFS rate (31.0± 8.6% FLT3/ITD positive vs. 50.3± 2.4% FLT3/ITD wild-type; P = 0.056) was shown for NUP98-negative patients, but a negative impact not on the 5-year OS rate (68.0± 2.3% FLT3/ITD positive vs. 55.0 ± 9.3% FLT3/ITD wildtype; P = 0.242) was not detected.
Stratification evaluation of FLT3/ITD in NUP98/NSD1 patients
Among 26 patients positive for both NUP98 gene fusion and FLT3/ITD, 84.6% of patients harbored NUP98/NSD1 and FLT3/ITD(Table2). The prognostic effect of NUP98/NSD1 in patients who carried FLT3/ITD was then evaluated. The CR rate after the first course of induction therapy in those with both NUP98/NSD1 and FLT3/ITD was 27.3% (6 of 22) compared with 78.8% (26 of 33) for FLT3/ITD mutation patients without NUP98/NSD1 (P=0.0001).The CR rate after second course of induction therapy in patients carrying NUP98/NSD1 and FLT3/ITD was 36.4% (8 of 22) compared with 87.9%(29 of 33)for patients with FLT3/ITD mutation but without NUP98/NSD1 (P=0.0001) (Supplementary Table 1).
As shown in Figure 3(A), in all 529 patients,patients harboring NUP98/NSD1 had an inferior EFS (26.1%) compared with patients who were NUP98/NSD1-negative (48.9%, P<0.001). Additionally, NUP98/NSD1 patients displayed a trend toward worse OS(50.1%)compared to NUP98/NSD1 negative patients (66.0%, P=0.029) (Figure 3(B)). In analysis restricted to 57 NUP98 fusion genes patients,there were no significant differences in OS between NUP98-NSD1(+) and NUP98-NSD1(-) patients, with EFS being poor in both (26.1% vs. 41.9%, P=0.026) (Figure3(C, D)).
When examining only patients who were NUP98/NSD1 positive and harbored FLT3/ITD mutation, we found unfavorable 5-year OS (41.7%) and 5-year EFS (13.6%) outcomes compared with those of FLT3/ITD wild-type cases without NUP98/NSD1(OS 67.1± 2.3%, P=0.004; EFS 50.1± 2.4%, P<0.0001) (Figure 4B). Survival curves were generated based on NUP98/NSD1 positivity and FLT3/ITD status (Figure 4(A, B)),revealing that NUP98/NSD1 patients had no impact on 5-year EFS (50.0 ± 14.4% NUP98/NSD1+vs. 50.1± 2.4%NUP98/NSD1-; P =0.705) or OS (65.6± 14.0% NUP98/NSD1+vs 67.1± 2.3% NUP98/NSD1-; P = 0.900) in FLT3/ITD-negative patients.
Overall, FLT3/ITD mutation patients had a negative impact on 5-year EFS (13.6± 7.3% NUP98/NSD1+vs 29.0 ± 8.0% NUP98/NSD1-; P = 0.005) but not on 5-year OS (41.7± 11.0% NUP98/NSD1+vs 53.0± 8.9% NUP98/NSD1-; P = 0.147) according to NUP98/NSD1 status.
Stratification evaluation of the thresholds FLT3/ITD AR and NUP98/NSD1
As illustrated in Figure5(A, B), patients harboring NUP98/NSD1gene fusion with a high FLT3/ITD AR had poor prognosis with a 5-year OS of 33.3 ± 13.6%, however, there was no significant difference compared with the 55.6±16.6 % in patients with low FLT3/ITD AR. The reason may be the small sample size. In analysis restricted to high FLT3/ITD AR, patients had poor 5-year EFS prognosis, regardless of the presence or absence of NUP98/NSD1 (26.5± 9.9% NUP98/NSD1+vs. 23.1 ±9.9 % NUP98/NSD1-, P =0.214).
Figure5(C, D) indicated that patients with high FLT3/ITD AR and NUP98/NSD1 gene fusion positivity had a shorter 5-year OS of 33.3 ± 13.6%, which was not significantly different from the 55.5±11.1% of patients without NUP98/NSD1 fusion.
As illustrated in Figure5(E, F), for patients with low FLT3/ITD AR, NUP98/NSD1 had no impact on prognosis, with a 5-year OS of 55.6±16.6% compared with 48.6±14.8% for patients without NUP98/NSD1 (P=0.796).
The impact of SCT in pediatric AML patients carrying NUP98 gene fusion
Among the 57 NUP98-positive cases, only 11 patients received allogeneic SCT, and data were available for 40 patients. Evaluation of the effect of SCT on NUP98 gene fusion patients revealed that the therapy did not have a significant effect on the clinical outcome with respect to 5-year EFS (36.4 ±14.5% SCT vs. 43.0 ± 9.5% no SCT, P = 0.962) or OS (50.5 ± 15.8% SCT vs. 53.1±9.6% no SCT, P = 0.923) rates when compared with chemotherapy (no SCT) Figure6(A, B).
The survival curses depicted In Figure 6(C, D), demonstrated that SCT had no significant effect on the 5-year EFS (42.9 ±18.7% of patients with SCT vs. 40.4 ± 14.6% of patients with non-SCT, P = 0.717) or OS (68.6± 18.6% of patients with SCT vs. 65.6±14.0% of patients with non-SCT, P = 0.802) compared to chemotherapy (no SCT) in NUP98/NSD1 patients.
Multivariate analysis of prognostic factors for overall and event-free survival in pediatric AML patients
Multivariable Cox regression analysis including NUP98-NSD1+/FLT3/ITD+, NUP98-NSD1+/FLT3/ITD-, NUP98-NSD1-/FLT3/ITD+, NPMI status, WTI status, age, WBC, and risk group was performed to determine whether NUP98-NSD1+/FLT3/ITD+ is an independent prognostic factor and concluded that NUP98-NSD1+/FLT3/ITD+ is an independent prognostic factor for poor EFS (HR=4.8, 95% CI: 1.1-21.6, P = 0.041), though it did not have a significant impact on OS. Regardless of NUP98-NSD1+/ FLT3/ITD -or NUP98-NSD1-/ FLT3/ITD + status, EFS and OS were not affected. As expected, we showed WT1mutation was significantly associated with inferior EFS and OS. Moreover, a WBC count >=100×10^9/L, but not OS, had a significant negative influence on EFS. Conversely, neither NPM1 mutation nor risk group had significant impact on EFS or OS (Table 3).