Real-World Outcomes of Relapsed/Refractory Core-Binding Factor Acute Myeloid Leukemia: A COMMAND Registry Study

Abstract

Core-binding factor acute myeloid leukemia (CBF-AML) is characterized by the presence of inv(16)/t(16;16) or t(8;21) and is classified as favorable risk by the 2022 European LeukemiaNet (ELN) guidelines [1]. We have previously reported on outcomes of patients with newly diagnosed CBF-AML treated with intensive chemotherapy (IC) regimens [2] and despite the favorable risk status, approximately 50% of patients experienced relapse. Prior analyses have shown limited survival after relapse [3-5]. Khan et al. reported on 92 patients with relapsed/refractory (R/R) CBF-AML treated from 1990 to 2014 with a median overall survival (mOS) of 12 months; type of therapy and allogeneic stem cell transplant (alloSCT) did not impact survival [3]. Hospital et al. evaluated 145 patients with CBF-AML from 1994 to 2011 who underwent IC on a variety of French AML Intergroup studies and reported a second complete remission (CR2) rate of 88% with 5-year disease-free survival of 50% for the entire cohort [5]. Additionally, 53% of the patients included in this study proceeded to undergo alloSCT in CR2, with the study noting a survival benefit for patients with inv(16) but not for t(8;21). Finally, Kurosawa et al. analyzed 139 patients treated from 1999 to 2006 with R/R CBF-AML and reported a 3-year overall survival (OS) of 48% along with a survival benefit seen with alloSCT in those with t(8;21) [4].

While these prior studies offer valuable insight into the outcomes of R/R CBF-AML, the therapeutic armamentarium for AML has greatly expanded since 2017 with the re-approval of the CD33-directed antibody drug conjugate gemtuzumab ozogamicin (GO) and guideline recommendations for its frontline use in intensive IC-eligible patients with CBF-AML, and the incorporation of targeted therapeutics and the BCL2 inhibitor venetoclax into management [6]. Furthermore, access to alloSCT has expanded given the increasing utilization of haplo-identical as well as mismatched donors when appropriate. Given these advances, we sought to characterize the treatment patterns and outcomes of R/R CBF-AML patients in a timeframe inclusive of these advances.

Here we report on 68 patients with CBF-AML who experienced relapse or were refractory to frontline IC from the Consortium on Myeloid Malignancies and Neoplastic Diseases (COMMAND) registry and received treatment for their R/R disease. Patients with AML harboring inv(16)/t(16;16) or t(8;21) who were treated with IC from January 2010 through April 2023, across the seven participating institutions, were included. The study was approved by the institutional review boards at institution participating in the COMMAND registry. Survival analysis was done using Kaplan–Meier survival estimates and log-rank tests of significance, with additional modifications as noted below.

The median age of patients at diagnosis with CBF-AML was 48.5 years (range 20–72 years), with 59% male patients and 24% of patients identifying as under-represented minorities (African American, Hispanic, or other). Fifty percent of patients had AML harboring t(8;21) while 50% had AML with inv(16), with 18% harboring other mutations and cytogenetic abnormalities that would have otherwise categorized them as adverse risk by ELN22 guidelines when excluding CBF status. Six percent of patients were categorized as having therapy-related AML. All patients were treated with frontline IC, with 30% of patients also receiving GO with their frontline induction therapy. Six percent of patients (4/68) were refractory to induction therapy, and 74% received post-relapse treatment from 2017 onwards. Additional characteristics are summarized in Table S1.

The median time to relapse was 11.0 months (inter-quartile range [IQR]: 8.5–14.5 months). At relapse, 29% of cases harbored additional cytogenetic abnormalities; 7% (5/68) harbored complex cytogenetics. Fifty percent of patients (n = 34) had next-generation sequencing (NGS) testing done at time of relapse, with KIT mutations being the most common at 7/34 (21%). Among the seven patients with AML harboring KIT mutations, three did not have a KIT mutation detected at original CBF-AML diagnosis. Other most common mutations detected at relapse included: NRAS (5/34), FLT3 (4/34), IDH2 (3/34), and KRAS (2/34). Altogether among patients with NGS testing at relapse 12/34 (35%) would have been categorized as adverse risk by ELN22 criteria after excluding CBF-AML status. Mutations by treatment group at time of relapse (IC vs. hypomethylating agent (HMA) + venetoclax vs. HMA alone) are further detailed in Table S3.

All included patients received additional chemotherapy after relapse. Seventy-five percent (51/68) received intensive chemotherapy with 6/68 (9%) receiving venetoclax in addition to intensive chemotherapy. Sixteen percent (11/68) of patients received an HMA with venetoclax, and 7% (5/68) received HMA monotherapy. Additionally, among all patients 9% (6/68) received an additional targeted therapeutic (e.g., IDH2 inhibitor) and 3% (2/68) received therapy with the addition of GO.

The median time from initiation of therapy to response assessment was 1.2 months (IQR: 0.9–1.7 months). Among 67 evaluable patients, 75% (50/67) achieved a complete response (CR) or CR with incomplete hematologic recovery (CRi), and 18% (12/67) had progressive disease. The overall CR/CRi rate for all patients was 75%, 82% (42/51) for those receiving intensive chemotherapy, 55% (6/11) for HMA + venetoclax, and 40% (2/5) for HMA monotherapy. Of patients achieving a response to relapse therapy, 71% (39/55) went on to receive an alloSCT. Among patients initially achieving response to therapy, 29% (16/55) experienced subsequent disease relapse at a median time of 8.3 months (IQR: 4.7–12.4 months) after first relapse.

The median time of follow-up for all included patients after relapse or refractory disease was 13.9 months (IQR: 6.0–39.3 months). The 12-month event-free survival (EFS) from time of first relapse was 53% (95% CI: 42%–67%), and 12-month OS was 65% (95% CI: 54%–78%) (Figure 1A,B). To evaluate survival outcomes by intensity of therapy, we censored patients at time of receipt of alloSCT. For those receiving IC, 12-month EFS was 26% (95% CI: 14%–49%) compared to 12% (95% CI: 20%–68%) for those receiving lower-intensity therapy (p = 0.66); 12-month OS for those receiving intensive chemotherapy was 28% (95% CI: 16%–51%), compared to 31% (95% CI: 12%–73%) for those receiving lower-intensity therapy (p = 0.76) (Figure 2A,B).

FIGURE 1

Open in figure viewerPowerPoint

(A and B) Whole cohort EFS and OS.

FIGURE 2

Open in figure viewerPowerPoint

(A and B) Survival outcomes by intensity of therapy for R/R CBF-AML (censored at time of alloSCT).

We also evaluated survival outcomes by alloSCT consolidation with landmark analysis at 3 months after first relapse. For patients who received a consolidative alloSCT, the 12-month EFS was 73% (95% CI: 60%–89%), while those who did not have alloSCT consolidation experienced a 12-month EFS of 24% (95% CI: 9%–64%) (p < 0.01). Similarly, those who underwent alloSCT experienced a 12-month OS of 81% (95% CI: 70%–95%) compared to 52% (95% CI: 30%–89%) for those who did not undergo alloSCT (p < 0.01) (Figure 3A,B).

FIGURE 3

Open in figure viewerPowerPoint

(A and B) Survival outcomes by alloSCT with 3-month landmark analysis.

In a multivariable Cox proportional hazards model for EFS, consolidation with alloSCT after first relapse was associated with significantly improved EFS (hazard ratio [HR] 0.25, 95% CI: 0.12–0.48, p < 0.01). In the multivariable model for OS, consolidation with alloSCT after first relapse was also associated with significantly improved OS (HR 0.19, 95% CI: 0.05–0.75, p = 0.02) as well as the presence of mutations categorized as adverse risk by ELN22 guidelines when excluding CBF status (HR 4.2, 95% CI: 1.4–12.4, p < 0.01) (Table S2).

In this modern, multi-center study of CBF-AML patients who relapse after frontline IC, our data suggest that there is no survival benefit to intensive therapies over less intensive ones when censoring for alloSCT. However, CR/CRi rates were higher for those with intensive chemotherapy, suggesting that there may be benefit in intensive chemotherapy in bridging patients to alloSCT for definitive treatment. While CBF cytogenetics (inv[16] vs. t[8;21]) were associated with EFS and OS on univariate analysis, this did not persist in the multivariate model, where only receipt of alloSCT was associated with significantly improved EFS and OS.

Compared to prior retrospective studies [3-5] as discussed above, we did not find a significant difference in the benefit of alloSCT between patients with R/R CBF-AML harboring inv(16) versus t(8;21). Our study population had a higher utilization of alloSCT as a consolidative strategy at 71%. This may be reflective of a more modern treatment era which includes advances in the availability of alloSCT donors for more patients, along with the inclusion of a broader set of academic institutions nationally. We also did not find differences in survival outcomes between intensive versus lower-intensity treatments when censored at time of alloSCT in contrast to prior studies, which is also likely reflective of improvements in lower intensity treatment options, including the addition of venetoclax to HMA therapy, as well as targeted inhibitors for other co-occurring mutations such as IDH2.

Our study has limitations, most notably the limited sample size and thus ability to pursue further subgroup analyses. While we note significant benefit by consolidation with alloSCT, we did not have additional details on donor HLA matching status and conditioning regimens available. Baseline patient factors at time of relapse such as performance status and additional hematologic variables including blood cell counts and blast percentages were also unavailable.

Patients with R/R CBF-AML after frontline induction therapy should be strongly considered for alloSCT consolidation upon achieving a response. The emergence of strategies which allow increased tolerance of haplo-identical or HLA-mismatched donors promises to expand the pool of suitable donors for patients in need of alloSCT. We demonstrate a clear benefit in our patient cohort of alloSCT for CBF-AML patients after first relapse, irrespective of other factors including CBF cytogenetics in the modern era, and thus this treatment goal should be a main therapeutic goal for management of relapsed CBF-AML in patients eligible for alloSCT.