Correlation of responses to venetoclax combination treatment and AML mutations or BCL-2 expression

Acute myeloid leukemia (AML) is a genetically heterogenous disease with multiple cytogenetic markers used as outcome predictors that help stratify patients in different risk groups and guide treatment.¹ Some of these markers include, NPM1, TP53, FLT3, and IDH1/2, while overexpression of the B-cell leukemia/lymphoma 2 (BCL-2) protein has been associated with poor responses and even resistance to AML chemoimmunotherapy.² The selective BCL-2 inhibitor, venetoclax, has been approved by the U.S. Food & Drug Administration (FDA) in combination with a hypomethylating agent (HMA; azacitidine or decitabine) or low-dose cytarabine (LDAC) for the treatment of previously-untreated older patients with AML, where standard chemotherapy is not the best treatment option (Read more here).³

During the 61^st American Society of Hematology Meeting & Exposition on Monday 9^th December 2019, Brenda Chyla from AbbVie Inc., IL, US, presented an exploratory biomarker analysis from two phase Ib/II studies that examined the combination of venetoclax with HMA (NCT02203773) or LDAC (NCT02287233) in older patients with AML, who were ineligible for induction chemotherapy. The primary objectives of this study were to evaluate baseline genomic data from these two clinical trials and to correlate them to patient outcomes.⁴

Study design⁴

• The study design and analysis of that venetoclax + HMA clinical trial (NCT02203773) has been summarized here by the AML Global Portal and the relevant information on the venetoclax + LDAC trial (NCT02287233) can be found here
• Data cut-off dates for the exploratory analysis:
  • Venetoclax + HMA group: August 31, 2018
  • Venetoclax + LDAC group: August 22, 2018
• Evaluable data from 209 patients included in two phase Ib/II trials that explored either:
  • Venetoclax (400mg) + HMA (azacitidine [20mg/m²] or decitabine [75mg/m²]) or
  • Venetoclax (600mg) + LDAC
• Median age (range): 74 (61-90) years
• Patient responses were assessed based on the International Working Group (IWG) AML criteria⁵
• AML mutations (IDH1/2, TP53, and FLT3) were identified in bone marrow aspirates (BMAs) taken before treatment initiation
• BCL-2 expression was determined by qPCR in patients with ≥ 50% AML blasts at baseline

Key findings⁴

• Recurrent AML mutations (NPM1, IDH1/2, TP53, and/or FLT3) were identified in:
  • 119 out of 127 patients who received venetoclax + HMA (Table 1)
  • 79 out of 82 patients who received venetoclax + LDAC (Table 2)

Table 1. Correlation of baseline molecular markers with responses to venetoclax plus HMA⁴

CR, Complete response; Cri, CR with incomplete hematologic recovery

AML mutation CR (%) Cri (%) CR + Cri (%) NPM1 (n= 21) 67 14 81 IDH1 (n= 19) 37 37 74 IDH2 (n= 14) 71 22 93 FLT3 (n= 17) 41 6 47 TP53 (n= 30) 30 27 57 Other mutation (n= 18) 49 27 76 All patients (N= 119) 46 26 72

 Table 2. Correlation of baseline molecular markers with responses to venetoclax plus LDAC⁴ 

CR, Complete response; Cri, CR with incomplete hematologic recovery

AML mutation CR (%) CRi (%) CR + Cri (%) NPM1 (n= 9) 78 11 89 IDH1 (n= 9) 71 21 93 IDH2 (n= 11) 55 18 73 FLT3 (n= 16) 19 25 44 TP53 (n= 13) 8 31 38 Other mutation (n= 21) 22 28 50 All patients (N= 79) 25 28 53

• Patients with IDH1/2 mutations achieved high response rates with or without co-occurring NPM1 mutations with complete response (CR) or CR with incomplete hematologic recovery (Cri) rates of 91% or 74%, respectively
• Patients with FTL3 and NPM1 concurrent mutations achieved high response rates with a CR/CRi rate of 73%
• Conversely, patients with FLT3 mutations but without an NPM1 driver mutation achieved lower response rates with a CR/CRi of 32%

For further details on the effect of both treatment options on the median overall survival (mOS) and one-year overall survival (OS) see Table 3.

 Table 3. mOS and 12 month OS estimate in both treatment groups by mutation

mOS, median overall survival; NR, not reached; OS, overall survival

Mutation mOS (95% CI), months 12 month OS estimate (95% CI), % NPM1 (n= 30) NR (12.5-NR) 73 (53-85) IDH1 (n= 28) 15.9 (10.7-NR) 67 (46-81) IDH2 (n= 25) NR (19.4-NR) 76 (54-88) FLT3 (n= 33) 9.9 (4.8-17) 47 (29-63) TP53 (n= 43) 6.4 (3.8-8.9) 23 (11-37) FLT3^mut NPM1^wt(n= 22) 5.9 (3-14) 35 (16-54) FLT3^wt NMP1^mut NR (11-NR) 74 (48-88) FLT3^mut NMP1^mut(n= 11) 27.8 (4.8-NR) 72 (35-90)

BCL-2 expression ⁴ 

• BCL-2 mRNA expression in BMAs with ³ 50% AML blasts was quantified in:
  • 42 out of 127 patients who received venetoclax + HMA
  • 34 out of 82 patients who received venetoclax + LDAC
• BCL-2 mRNA expression was variable within different molecular subgroups with a median of 0.9 (range, 0.03-9)
  • Patients with IDH1/2 had higher BCL-2 expression (p= 0.04) compared to their wildtype counterparts than those seen in the NPM1, TP53, or FLT3 groups
• BCL-2 levels were not an independent predictor of response to treatment with venetoclax plus HMA or LDAC (p= 0.75) as assessed by a multivariate generalized additive model analysis (GAM) that took into account; age, cytogenetic risk, OS and AML type for BCL2 expression
• No associations were identified in BCL-2 or BCL-XL and MCL1 expression levels, and responders versus non-responders

Conclusion

Patients with NPM1 or IDH1/2 mutations achieved high response rates (> 70%) and prolonged OS irrespective of the venetoclax combination therapy they received. Interestingly, outcomes (OS and CR/CRi rate) were better in patients with concurrent FLT3 and NPM1 mutations than those with a FLT3 mutation alone. Patients with TP53 mutations managed to achieve good responses but they were not sustained, with a mOS of 6.4 months. BCL-2 expression was not prognostic of response to venetoclax combination therapy (HMA or LDAC) in this naïve, older patient population that is not eligible for standard induction therapy.