Chemotherapy with or without oblimersen (G3139) in older patients with AML: results from the phase III CALGB 10201 trial

Patients aged ≥60 years with acute myeloid leukemia (AML) often have inferior responses to standard induction chemotherapy—such as cytarabine- and anthracycline-based regimens—due to adverse cytogenetic and/or molecular abnormalities. One of these adverse molecular abnormalities is overexpression of B-cell leukemia/lymphoma 2 (BCL2), which contributes to chemotherapy resistance by preventing the initiation of apoptosis. Venetoclax is a first-in-class BCL2 inhibitor that is approved in combination with hypomethylating agents or low-dose cytarabine in adults with AML who are ≥75 years or are unfit for intensive induction chemotherapy; approval was based on results of the VIALE-A trial, which were previously reported on the Lymphoma Hub. Combining a BCL2 inhibitor, like venetoclax, with low-dose cytarabine or hypomethylating agents is a treatment strategy that is thought to promote durable remission in patients with BCL2-expressing leukemia, without adding to the toxicity associated with standard induction chemotherapy, which is particularly important in the context of treating older and/or unfit patients with AML.¹

Oblimersen sodium (G3139) is a molecule that blocks the production of BCL2, increasing the susceptibility of cells to cytotoxic agents; it demonstrated an ability to downregulate BCL2 in two phase I trials of patients with acute leukemia and was associated with encouraging complete remission (CR) rates when given in combination with standard induction and consolidation regimens. With these promising phase I results, Walker, et al. investigated the use of G3139 in combination with cytarabine/daunorubicin induction and high-dose cytarabine consolidation in older, untreated patients with AML, the results of which we will discuss here.¹

Study design

The Cancer and Leukemia Group B (CALGB) 10201 study was a phase III randomized trial that investigated the addition of G3139 to standard cytarabine/daunorubicin induction and high-dose cytarabine consolidation therapy. Patients (N = 506) were ≥60 years, had an unequivocal diagnosis of AML, and had not received prior cytotoxic chemotherapy; those with AML secondary to myelodysplastic syndrome were eligible, while those with promyelocytic leukemia were excluded.

Patients were randomized 1:1 to one of two treatment arms as shown in Table 1. Patients who achieved CR and had resolution of all significant induction toxicities to Grade <2 were eligible for consolidation, and patients who completed the first consolidation were eligible for a second cycle following remission-confirming bone marrow biopsy. Patients who experienced a decline in left ventricular ejection fraction to <35% and had been randomized to receive G3139 had it omitted from consolidation therapy.

Table 1. Treatment schedule*

Cytogenetic analysis was performed on pretreatment bone marrow and/or peripheral blood samples, with results confirmed by central karyotype review, and the mutational statuses of NPM1, ASXL1, RUNX1, TP53, CEBPA, and FLT3 were determined. Patients with available data were categorized according to the 2017 European Leukemia Net (ELN) genetic risk classification. The primary endpoint was overall survival (OS), and interim efficacy analyses were planned for OS, disease-free survival (DFS), event-free survival (EFS), and CR rates.

Baseline characteristics

Of the 506 patients who were enrolled, 60.7% were male, and 45.1% were ≥70 years, with a mean age of 69.7 years (Table 2).

Table 2. Baseline characteristics*

Results

Induction

Of the 506 patients, 277 patients achieved remission in response to induction therapy. There was no statistical difference in CR between arm A and arm B (53% vs 56%; p = 0.53; Table 3), even in age (<70 vs ≥70 years), AML type, and 2017 ELN risk group subgroup analyses. There were 64 (26%) patients in arm A and 66 (27%) patients in arm B who received a second induction course.

In terms of treatment toxicities during induction, over 90% of patients in both treatment arms had Grade 4 neutropenia and thrombocytopenia. Febrile neutropenia (73% vs 77%) and infection (57% vs 56%) were the most common Grade 3/4 nonhematologic toxicities and occurred equally between arms A and B.

Left ventricular systolic dysfunction developed in 12 patients in arm A and in eight patients in arm B, and this difference was not statistically significant.

Consolidation

Of the 277 patients who achieved CR during induction, 211 received at least one cycle of consolidation. There were 70 patients (68%) in arm A and 73 patients (68%) in arm B who completed both cycles of consolidation; although, five patients in arm A did not receive G3139 during consolidation, none of whom discontinued due to a decline in cardiac ejection fraction. Patients who did not receive both cycles withdrew due to disease progression (n = 29), toxicity (n = 14), switching to an alternate therapy (n = 4), or death (n = 6).

In terms of toxicities during consolidation, >90% of patients experienced Grade 4 neutropenia and thrombocytopenia. Febrile neutropenia (38% vs 47%) and Grade 3 or 4 infection (27% vs 33%) were common in both arm A and arm B, though not as frequently as during induction. Nine patients (4%) in arm A and two patients (1%) in arm B developed Grade 3 ataxia. Four treatment-related deaths occurred during consolidation, two in each arm.

Survival and long-term follow-up

Median follow-up was 100 months (range, 1–119 months) for the 41 surviving patients (of the 506 who were randomized). There were no significant differences between arm A and arm B in OS, EFS, or DFS (Table 3). No overall differences were observed in CR rates, OS, DFS, or EFS between arms A and B regardless of ELN genetic risk group. There were no significant differences between arm A and arm B in EFS or DFS based on age, though OS for patients <70 years in arm A was significantly longer than OS for patients <70 years in arm B (median, 10 months vs 9 months; p = 0.04). There were no significant differences between arm A and arm B in EFS or OS based on AML subtype, though patients with secondary AML in arm A had significantly improved DFS compared with arm B (median, 10 months vs 6 months; p = 0.04).

Table 3. Treatment response and survival*

Of the 46 patients who underwent stem cell transplantation, most received allogeneic hematopoietic stem cell transplantation from either a matched sibling (n = 25) or a matched unrelated donor (n = 11). There was no difference in the number of transplantation recipients by arm (24 in arm A and 22 in arm B; p = 0.99), and there were no significant differences in OS, EFS, or DFS between the arms for patients who did or did not receive allogeneic hematopoietic stem cell transplantation. Notably, for patients with secondary AML who did not undergo transplantation, those in arm A had improved OS and DFS compared with arm B (p = 0.03 and p < 0.01, respectively).

Conclusion

While adding G3139 to standard chemotherapy did not show significant improvements in CR rates for older patients with AML, improved DFS was seen in patients with secondary AML, suggesting that this subgroup, which tends to be more refractory to conventional treatment, may be more sensitive to BCL2 downregulation. Despite its failure to improve outcomes in older patients with AML, G3139 was well-tolerated in this group and could be safely added to conventional chemotherapy for these patients, and although there has been no further development of G3139, the concept of attacking antiapoptotic proteins may pave the way for future therapeutic development in AML.