HMA + venetoclax may improve outcomes in patients with splicing mutations

Spliceosome mutations, including SRSF2, U2AF1, SF3B1, and ZRSR2, are identified in approximately one-third of patients with myelodysplastic syndromes (MDS) and in almost 50% of patients with acute myeloid leukemia secondary to MDS (sAML).¹ Splicing mutations are associated with poor outcomes, though many patients with these mutations also have cohesin complex mutations (STAG1, STAG2, SMC3, SMC1A, RAD21), the prognostic impacts of which are unclear.

For older patients with AML who are ineligible for intensive treatment, the standard of care includes venetoclax incorporated into a low-intensity regimen of low-dose cytarabine or hypomethylating agents (HMA + Ven). Some molecular subgroups (NPM1, IDH1/2, TP53) have been identified as predictive of venetoclax sensitivity or resistance, but outcomes in other molecular subgroups remain to be defined and, though spliceosome mutations have historically been linked to poor outcomes, venetoclax-based therapy may attenuate these negative prognostic effects.

Though improved responses have been seen in prospective trials of venetoclax-based therapy, long-term outcomes for patients with splicing and/or cohesin complex mutations have not been elucidated. It is important to understand the responses within these molecular subgroups and stratify patients accordingly. To this end, Lachowiez et al. conducted a study published recently in Blood Advances,¹ the key points of which are summarized below.

Study design and methods

This cohort analysis included patients ≥18 years of age with a diagnosis of AML who had been treated with frontline HMA + Ven-based therapy at the University of Texas MD Anderson Cancer Center; the population included patients treated on prospective clinical trial protocols and those treated off-protocol. Patients with FLT3-ITD mutations treated with FLT3 inhibitors in combination with HMA + Ven on protocol were included, patients with acute promyelocytic leukemia were excluded.

Patients were stratified into spliceosome-mutated and wildtype cohorts. Those with cohesin complex mutations were part of a subgroup analysis within the two larger cohorts. Patients were also stratified into risk groups using 2017 European LeukemiaNet (ELN) guidelines and response criteria were defined using International Working Group criteria for AML. Time-to-event endpoints included OS, event-free survival (EFS), and duration of response.

Cytogenetics testing was performed using standard metaphase karyotype analysis, molecular analysis was performed using next-generation sequencing, and measurable residual disease (MRD) was assessed using multiparameter flow cytometry.

Results

The analysis included 119 patients treated with HMA + Ven; patient demographics are shown in Table 1. Of note, patients with splicing mutations were significantly older and more likely to be male, and the majority of patients in both cohorts were in the adverse ELN risk group.

Table 1. Patient demographics*

AML, acute myeloid leukemia; d, day; ECOG, Eastern Cooperative Oncology Group; ELN, European LeukemiaNet; MDS, myelodysplastic syndromes; MPN, myeloproliferative neoplasms; sAML, secondary AML; tAML, therapy-related AML; WBC, white blood cell. *Data from Lachowiez et al.1
Demographic	All patients (N = 119)	Spliceosome (n = 39)	Wildtype (n = 80)	P value
Median age, years (range)	72 (40–89)	75 (56–85)	70 (40–89)	0.047
Male, n (%)	65 (55)	33 (85)	32 (40)	<0.01
ECOG score, median (range)	1 (0–3)	1 (0–3)	1 (0–3)	0.11
Median WBC, × 103/μL (range)	3.1 (0.3–81)	3.6 (0.3–81)	2.9 (0.5–43)	0.61
Median treatment duration, months (range)	5.7 (0.1–43)	5.5 (0.5–43)	5.7 (0.1–32)	0.90
Treatment regimen, n (%)
Azacitidine + venetoclax	14 (12)	3 (8)	11 (14)	0.54
5-d decitabine + venetoclax	5 (4)	2 (5)	3 (4)	0.66
10-d decitabine + venetoclax	100 (84)	34 (87)	66 (83)	0.60
FLT3 inhibitor	14 (12)	2 (5)	12 (15)	0.14
Disease, n (%)
De novo AML	70 (59)	24 (62)	46 (58)
sAML                   secondary to MDS                   secondary to MPN	20 (17) 17 3	8 (21) 7 1	12 (15) 10 2	0.44 — —
tAML	29 (24)	7 (18)	22 (28)	0.36
ELN risk group, n (%)
Favorable	28 (24)	11 (28)	17 (21)	0.49
Intermediate	19 (16)	7 (18)	12 (15)	0.79
Adverse	72 (61)	21 (54)	51 (64)	0.32
Cytogenetic group, n (%)
Diploid	39 (33)	14 (36)	25 (31)	0.68
Other intermediate	30 (25)	14 (36)	16 (20)	0.07
Adverse/complex	50	11 (28)	39 (49)	0.047

Considering cytogenetic characteristics:

• Compared to the spliceosome cohort, more patients in the wildtype cohort had adverse risk cytogenetics/complex karyotype (49%, p = 0.047) (see Table 1).
• Mutational differences were noted between the spliceosome and wildtype cohorts, with IDH mutations (36% vs 15%) significantly associated with spliceosome mutations and TP53, TET2, and EZH2 mutations significantly associated with the wildtype cohort (p < 0.05). Mutations in DNMT3A, ASXL1, TET2, TP53, RUNX1, RAS, and NPM1 were present in ≥15% of patients in each cohort.
• The spliceosome cohort was enriched with mutations in methylation genes (64% vs 33%; p < 0.01), while the wildtype cohort was enriched with chromatin modifiers and tumor suppressor mutations (45% vs 23%, p = 0.03), and there was a trend toward enrichment for active signaling mutations in the wildtype cohort (55% vs 26%, p = 0.05).
• There were no significant differences observed in transcription factor or cohesin complex mutations, and the median mutation burden was not significantly different between the spliceosome and wildtype cohorts (4 vs 4.5) or between spliceosome mutations.

Within the spliceosome cohort:

• SRSF2 mutations frequently co-occurred with mutations in IDH2, TET2, NPM1, and RUNX1.
  • In fact, IDH2 mutations occurred almost exclusively with SRSF2 mutations compared to other splicing mutations (88% vs 11%).
• SF3B1 mutations frequently co-occurred with mutations in RUNX1, NPM1, DNMT3A, TP53, and IDH1.
• U2AF1 mutations co-occurred with mutations in RAS (which were enriched in patients with U2AF1 mutations; p < 0.01), TET2, and DNMT3A.

Cohesin complex mutations, which have demonstrated preclinical sensitivity to venetoclax-based treatment, were identified in 14% (n = 17) of patients in the study cohort, 13% (n = 5) of patients in the spliceosome cohort, and 15% (n = 12) of patients in the wildtype cohort.

Response rates and survival outcomes

The ORR for the entire patient cohort was 81%. Patient outcomes, including composite CR (CRc) and responses based on specific mutations, are shown in Table 2.

Table 2. Patient outcomes*

CR, complete response; CRc, composite CR; CRi/CRh, CR with incomplete/partial hematologic recovery; MLFS, morphologic leukemia-free state; NR, not reached; ORR, overall response rate. *Data from Lachowiez et al.1 †p value presented is comparison between spliceosome and wildtype cohorts for the respective demographic.
Response Data are n (%) or n/N (%)	All patients (N = 119)	Spliceosome (n = 39)	Wildtype (n = 80)	P value
ORR	97 (81)	34 (89)	63 (79)	0.32
CRc	91 (76)	31 (79)	60 (75)	0.65
CR	71 (60)	20 (51)	51 (64)	0.23
CRi/CRh	20 (17)	11 (28)	9 (11)	0.03
MLFS	6 (5)	3 (10)	3 (4)	0.39
NR/died	22 (18)	5 (13)	17 (21)	0.32
CRc by molecular group				P value†
NPM1 (n = 29)	27 (93)	9/11 (82)	18/18 (100)	0.02
IDH1 (n = 10)	6 (60)	3/5 (60)	3/5 (60)	0.24
IDH2 (n = 16)	14 (88)	8/9 (89)	6/7 (86)	0.35
FLT3-ITD/TKD (n = 20)	16 (80)	4/5 (80)	12/15 (80)	0.78
TP53 (n = 36)	21 (58)	3/6 (50)	18/30 (60)	0.01
Adverse/complex cytogenetics (n = 50)	32 (64)	5/11 (45)	27/39 (69)	0.01

Of note, 74 of the 91 patients achieving a CRc had samples that were evaluable for MRD analysis, and MRD‾ CR was achieved in 40 of those patients, 10 of whom were in the spliceosome cohort and 30 of whom were wildtype. There was a 71% CRc rate among patients with isolated cohesin complex (n = 12) or co-occurring cohesin complex and spliceosome mutations (n = 5).  

Median follow-up was 24 months, and the median EFS and OS for the entire cohort were 10 months and 14 months, respectively. Survival outcomes are shown in Table 3.

Table 3. Survival outcomes*

AML, acute myeloid leukemia; CI, confidence interval; ELN, European LeukemiaNet; NA, not available; NR, not reached; sAML, secondary AML; tAML, therapy-related AML. *Data from Lachowiez et al.1 †p value presented is comparison between spliceosome and wildtype cohorts for the respective demographic.
Demographic	Event-free survival, months (95% CI)	Overall survival, months (95% CI)	P value†
All patients (N = 119)	10 (9–15)	14 (12–NR)	—
Spliceosome (n = 39)	10 (6–NR)	35 (13–NR)	0.58
ELN risk group
Favorable	NR (3–NA)	NR (3–NA)	0.48
Intermediate	10 (8–NA)	NR (13–NA)	0.97
Adverse	7 (3–NA)	14 (5–NA)	0.29
Cytogenetic group
Diploid	NR (10–NA)	NR (–)	0.57
Other intermediate	17 (9–NA)	35 (13–NA)	0.40
Complex/adverse	3 (2–NA)	5 (3–NA)	0.29
AML type
De novo AML	18 (8–NA)	NR (–)	0.35
sAML	7 (6–NA)	10 (6–NA)	0.80
tAML	6 (5–NA)	8 (6–NA)	0.69
Wildtype (n = 80)	11 (9–19)	14 (10–NA)
ELN risk group
Favorable	31 (13–NA)	NR (17–NA)	—
Intermediate	32 (7–NA)	NR (12–NA)	—
Adverse	9 (5.0–NA)	10 (8–18)	—
Cytogenetic group
Diploid	31 (20–NA)	NR (–)	—
Intermediate	11 (7–NA)	19 (9–NA)	—
Complex/adverse	5 (5–10)	9 (6–12)	—
AML type
De novo AML	19 (11–NA)	22 (14–NA)	—
sAML	5 (3–NA)	9 (6–NA)	—
tAML	6 (5.0–10)	10 (5–NA)	—

 Factors impacting outcomes of HMA + Ven treatment

NPM1 mutations were significantly associated with improved OS in the entire patient population, though the difference was only significant in wildtype patients compared with spliceosome patients (likely due to the low number of patients). There was, however, a trend toward improved OS in spliceosome patients with methylation mutations that was largely driven by IDH mutations. Tumor suppressor mutations were significantly associated with adverse outcomes in both cohorts, while no differences were observed in OS between patients with chromatin, active signaling, or transcription factor mutations.

Median OS was 35 months in patients with spliceosome mutations and/or cohesin complex mutations; there was no significant difference observed between those with co-occurring cohesin and spliceosome mutations, isolated spliceosome mutations, or isolated cohesin mutations. A 1-year survival of 80% was seen in patients with co-occurring mutations, while isolated mutations had a 1-year OS similar to that seen in wildtype patients (spliceosome, 58%; cohesin, 55%; wildtype, 55%).

Co-occurring spliceosome and IDH1/2 mutations were associated with especially favorable OS compared to spliceosome mutations in the absence of IDH1/2; 1- and 2-year OS in patients with co-occurring spliceosome and IDH1/2 mutations were 92% and 83%, respectively. Further analysis showed that IDH1/2 mutations were enriched in patients with SRSF2 mutations, and 1- and 2-year OS rates for patients with SRSF2/IDH1/IDH2 mutations were 100% and 88%, respectively. Additionally, RAS mutations were enriched in patients with U2AF1 mutations, and this was associated with lower CRc and MRD‾ CR rates, though there was no significant difference in survival when compared to patients in the wildtype cohort with RAS mutations.

Conclusion

Spliceosome mutations have been linked, historically, to inferior outcomes with intensive chemotherapy. In addition, these mutations are often identified in older patients with myeloid malignancies—particularly sAML—who are more likely to be unfit for intensive chemotherapy. In this study, the authors found that, in patients treated with HMA + Ven combinations, those with spliceosome mutations had similar outcomes to wildtype patients. This suggests that HMA + Ven may be an effective regimen for this population, particularly in patients with IDH2-enriched SRSF2 mutations. Other mutational pairs, such as U2AF1/RAS, may be able to identify patients with poorer responses to HMA + Ven, though further cohort analyses are needed.