Key Points
-
•
Survival outcomes in AML-MRC defined by ICC 2022 vary by mutation type, revealing significant heterogeneity in this adverse-risk group.
-
•
AML-MRC by ICC 2022 had worse EFS but better posttransplant OS vs non-MRC, highlighting distinct treatment responses in this group.
Visual Abstract
Abstract
Acute myeloid leukemia with myelodysplasia-related changes (AML-MRC) is a high-risk AML subtype defined by cytogenetic or genetic changes. We evaluated the relevance of AML-MRC classifications in a contemporary cohort of intensively treated patients with AML. We retrospectively reviewed 496 consecutive patients receiving induction chemotherapy for AML between 2016 and 2023 at a single center. Patients were categorized using both World Health Organization 2017 (MRC-2017) and International Consensus Classification 2022 (MRC-2022) criteria. Outcomes were compared using χ2 or Kaplan-Meier. MRC-2017 and MRC-2022 criteria identified 113 (22.8%) and 147 (29.6%) patients, respectively; 42.1% (77/183) met both definitions. Patients with AML classified by either definition had worse event-free survival (EFS) vs all patients with AML (MRC-2017: median, 1.4 vs 11.6 months; P < .001; MRC-2022: median, 1.9 vs 10.5 months; P = .019). Overall survival (OS) was inferior for MRC-2017 vs all patients with AML (median, 19.7 vs 53.1 months; P < .001) but not for MRC-2022. In the MRC-2022 mutation group, patients with a splice or transcription factor (TF) mutation had worse EFS vs mixed or chromatin modifier mutations (1.10 vs 5.35 vs 7.40 months, P = .002). Landmark analysis after allogeneic hematopoietic cell transplantation (allo-HCT) showed better 2-year OS for MRC-2022 vs non–MRC-2022 patients (77.0% vs 62.7%, P = .015). Multivariate analysis confirmed the EFS impact of TF/splice mutations (hazard ratio [HR], 1.65; 95% confidence interval [CI], 1.20-2.27; P = .002) and allo-HCT (HR, 0.49; 95% CI, 0.34-0.69; P < .001). These findings highlight the heterogeneity within the MRC-2022 group, particularly with mutation type and transplant status. Further refinement of AML-MRC risk stratification is needed.
Introduction
Acute myeloid leukemia (AML) with myelodysplasia-related changes (MRC) is considered an adverse-risk group with worse survival outcomes relative to favorable- or intermediate-risk AML in the European LeukemiaNet (ELN) 2022 risk stratification.1 The definition of AML-MRC has undergone significant changes as part of a broader refocusing of AML classification on genomic rather than clinical features.2 In the World Health Organization (WHO) 2017 scheme, patients could be classified as having AML-MRC based on ≥50% dysplastic cells in ≥2 cell lines, a history of myelodysplastic syndrome (MDS), or an MDS-related cytogenetic abnormality (MRC-2017).3,4 Therapy-related AML was considered as a separate group.3 The current International Consensus Classification (ICC) 2022 guidelines removed the AML-MRC–defining criteria of morphologic dysplasia and previous MDS history, with therapy exposure being considered a diagnostic qualifier. In its place, AML-MRC in ICC 2022 (MRC-2022) was defined by the presence of MRC-related cytogenetic changes (MRC-cyt) or gene mutations (MRC-mut). The 5th edition of the WHO classification follows a similar approach, with some variation in the genomic changes included, and a history of MDS or MDS/myeloproliferative neoplasm (MPN) is retained as a criterion.5 In ICC 2022, patients with TP53 alternations were also grouped separately into a new diagnostic category (AML-TP53). Patients are classified in a hierarchical fashion, with recurrent genomic events taking precedence, followed by AML-TP53, MRC-mut, and MRC-cyt, respectively.1,2 The changes between the MRC-2017 and MRC-2022 classifications are demonstrated in supplemental Table 1.
The validity of the MRC-2022 classification has been demonstrated in several studies on large patient cohorts.6, 7, 8, 9, 10, 11, 12, 13 However, most cohorts used for these validation studies were historical cohorts of patients treated with a variety of induction regimens and including many patients that were treated over a decade before the current treatment era. Several aspects of therapy for patients with AML have changed in this intervening period, such as the introduction of novel induction chemotherapy (IC) regimens and wider application of allogeneic hematopoietic cell transplantation (allo-HCT).14 Several advancements in the application of allo-HCT, including the development of reduced intensity conditioning regimens,15 alternative donor options,16 and better supportive care,17 have resulted in a higher fraction of patients being eligible for, and receiving, allo-HCT,18 particularly among older patients, among whom AML-MRC is seen more commonly. How these changes affect the prognostic ability of the MRC-2022 risk stratification scheme remains largely unclear, particularly for groups that have seen significant shifts in criteria, such as patients with AML-MRC. One study in patients with AML undergoing allo-HCT examined the impact of the MRC-2022 risk group on transplant outcomes; they identified that patients with adverse-risk disease had worse posttransplant survival but did not examine the AML-MRC subgroup.12
In this study, we examine the impact of classification changes between MRC-2017 and MRC-2022 in a real-world, intensively treated cohort of patients from a recent time period with a high rate of allo-HCT. Specifically, we focus on survival outcomes in the newly defined MRC-mut group in this context.
Methods
Patient inclusion and classification
Sequential patients with newly diagnosed AML receiving IC were retrospectively identified at a single institution from 2016 to 2023. Patients received 7+3 (cytarabine 100 mg/m2, daunorubicin 60 mg/m2), liposomal daunorubicin and cytarabine (CPX-351; 44 mg/m2), or other cytarabine-based IC. Patients who received lower-intensity therapies, such as venetoclax and azacitidine, as first-line treatment were excluded, but this regimen was considered a salvage therapy if used after IC. Cytogenetics and next-generation sequencing were done in a clinical laboratory and patients without these were excluded. Next-generation sequencing included the ICC 2022 AML-MRC mutations (SRSF2, SF3B1, RUNX1, U2AF1, ZRSR2, ASXL1, EZH2, BCOR, and STAG2) as well as all other recurrent abnormalities. A chromatin modifier mutation was defined as presence of ASXL1, BCOR, STAG2, or EZH2. A splicing mutation was defined as presence of SRSF2, U2AF1, SF3B1, or ZRSR2. A transcription factor (TF) mutation was defined as presence of RUNX1. Mixed AML-MRC was defined by the presence of mutations from >1 of these 3 categories. AML were retrospectively classified as AML-MRC or AML-TP53 by both MRC-2017 and MRC-2022 criteria, with hierarchical classification of AML-TP53, MRC-mut, or MRC-cyt as per the ELN 2022 guidelines.1, 2, 3, 4 Patients met MRC-2017 clinical criteria if they had morphologic dysplasia meeting the diagnostic threshold or previous MDS or MDS/MPN, without mutated NPM1 or CEBPA. Patients with therapy-related AML were not included in the AML-MRC group per MRC-2017 unless they met other criteria for AML-MRC. Publicly available metadata from the BEAT AML data set was obtained from cBioPortal (https://ww.cbioportal.org).19
Statistical analysis
Clinical end points such as event-free survival (EFS), overall survival (OS), induction mortality, complete response (CR), and CR with incomplete recovery were defined per ELN 2022 response criteria.1 EFS was defined as the time from diagnosis until induction failure after 1 cycle of IC (failure date corresponding to the end of treatment bone marrow), relapse from CR, or death from any cause. OS was defined as the time of diagnosis until death from any cause. Nonrelapse mortality was defined as any death that occurred after allo-HCT without known disease progression. Between-group differences were calculated by χ2 or t test. Survival was calculated by the Kaplan-Meier method. Univariate and multivariate regression was by Cox proportional hazard (PH). Patients were not censored at the time of allo-HCT unless otherwise specified. Statistical analysis was completed using R version 1.4.1717. A P < .05 was considered significant, and pairwise comparisons were adjusted using a Benjamini-Hochberg correction. This study was approved by The University of British Columbia research ethics board (H23-02013) and was conducted in accordance with the Declaration of Helsinki.
Results
Characteristics of study population and changes to MRC classification
A total of 496 patients diagnosed with AML who received IC were included in our analysis. Their baseline characteristics are shown in Table 1. The median age at diagnosis was 62 years (range, 19-78). MRC-2017 criteria were met by 113 of 496 (22.8%) and MRC-2022 criteria met by 147 of 496 (29.6%) patients. For patients meeting MRC-2017 criteria, 79 of 113 (69.9%) met cytogenetic criteria, 37 of 113 (32.7%) had previous MDS or MDS/MPN, and 42 of 113 (37.2%) had morphology dysplasia, with some patients meeting >1 criterion. Among patients meeting MRC-2017 criteria, 26 of 113 (23.0%) carried TP53 mutations with variant allele frequency of >10% and were classified in the MRC-2017 group because of an associated complex karyotype. Using the MRC-2022 hierarchical classification, 272 of 496 (54.8%) patients were classified as having recurrent genetic abnormalities, 28 of 496 (5.6%) as having AML-TP53, 126 of 496 (25.4%) as having MRC-mut, 21 of 496 (4.2%) as having MRC-cyt, and 49 of 496 (9.8%) as having AML not otherwise specified. Among patients meeting MRC-2017 or MRC-2022 criteria, there were only 77 of 183 (42.1%) patients who were classified in both categories (Figure 1A). Patients meeting MRC-2017 criteria were reclassified by MRC-2022 as having AML-TP53 (28/113 [24.7%]), MRC-2022 cytogenetic (21/113 [18.5%]), MRC-2022 mutation (60/113 [53.1%]), and non–MRC-2022 (9/113 [8.0%]). For patients not meeting MRC-2017 criteria, 66 of 383 (17.2%) were reclassified as having MRC-mut, 4 of 383 (1.0%) as having MRC-cyt, and 1 of 383 (0.3%) having AML-TP53 (Figure 1B).
Table 1.
Baseline characteristics of a retrospective cohort of patients with AML receiving IC from 2016 to 2023
| N = 496 | |
|---|---|
| Age at diagnosis, y | |
| Mean (range) | 58 (19-78) |
| Median (range) | 62 (19-78) |
| WHO 2017 classification | |
| MRC-2017 | 22.8% (113/496) |
| Cytogenetic criteria | 69.9% (79/113) |
| Previous MDS or MDS/MPN | 32.7% (37/113) |
| Morphologic dysplasia | 37.2% (42/113) |
| MRC-2017 and AML-TP53 | 23.0% (26/113) |
| ICC 2022 classification | |
| Recurring genetic abnormalities | 54.8% (272/496) |
| AML-TP53 | 5.6% (28/496) |
| AML-MRC | 29.6% (147/496) |
| MRC-Mut | 25.4% (126/496) |
| MRC-Cyt | 4.2% (21/496) |
| AML NOS | 9.8% (49/496) |
| ELN 2022 Risk | |
| Adverse | 40.3% (200/496) |
| Intermediate | 29.4% (146/496) |
| Favorable | 30.2% (150/496) |
| MRC chromosomal abnormalities | |
| Complex karyotype | 10.7% (53/496) |
| Chromosome 7 | 4.6% (23/496) |
| Chromosome 5 | 3.0% (15/496) |
| Chromosome 17 | 1.2% (6/496) |
| Chromosome 8 | 4.6% (23/496) |
| 13q | 0.4% (2/496) |
| 11q | 1.4% (7/496) |
| 3q | 0.8% (4/496) |
| Induction regimen | |
| 7+3 | 87.9% (436/496) |
| CPX-351 | 9.7% (48/496) |
| Other cytarabine-based | 2.4% (12/496) |
| Salvage regimens | |
| First salvage | 46.6% (231/496) |
| Second salvage | 15.7% (78/496) |
| Third salvage | 2.6% (13/496) |
| Venetoclax-based salvage | 10.4% (24/231) |
| Allo-HCT | 56.5% (280/496) |
| MRC-2017 | 54.0% (61/113) |
| MRC-2022 | 67.3% (99/147) |
| Conditioning | |
| Myeloablative | 83.9% (235/280) |
| Reduced intensity | 16.1% (45/280) |
| Donor | |
| Matched unrelated donor | 46.4% (130/280) |
| Matched sibling donor | 29.3% (82/280) |
| Mismatched unrelated donor | 13.2% (37/280) |
| Haploidentical donor | 10.7% (30/280) |
| Cord blood | 0.4% (1/280) |
| Nonrelapse mortality | 11.1% (31/280) |
NOS, not otherwise specified.
Figure 1.
Patient reclassification from MRC-2017 to MRC-2022 criteria. (A) A Euler diagram demonstrating the overlap of patients who met either MRC-2017 or MRC-2022 criteria for AML-MRC. (B) An alluvial diagram demonstrating the MRC-2022 groups into which MRC-2017 AML-MRC group patients were classified. MRC-2017, WHO 2017 myelodysplasia-related changes; MRC-2022, ICC 2022 myelodysplasia-related changes.
Most patients had ELN 2022 adverse-risk disease (200/496 [40.3%]), followed by favorable-risk (150/496 [30.2%]) and intermediate-risk (146/496 [29.4%]) disease. The most common karyotypic abnormality was a complex karyotype (53/496 [10.7%]). Most patients received 7+3 induction (436/496 [87.9%]) followed by CPX-351 (48/496 [9.7%]) and other cytarabine-based IC (12/496 [2.4%]). A salvage regimen was received by 231 of 496 (46.6%) patients, with 24 of 231 (10.4%) receiving a venetoclax-based salvage at some time point. A total of 280 of 496 (56.5%) patients received allo-HCT after their AML-MRC diagnosis, with 235 of 280 (83.9%) receiving myeloablative conditioning and 45 of 280 (16.1%) receiving reduced-intensity conditioning. Among patients meeting MRC-2017 criteria, 61 of 113 (54.0%) underwent allo-HCT; and 99 of 147 (67.3%) of patients meeting MRC-2022 criteria underwent allo-HCT. The most common donor type was matched unrelated donor (130/280 [46.4%]) followed by matched sibling donor (82/280 [29.3%]). The rate of nonrelapse mortality was 11.1% (31/280).
Induction outcomes and responses by MRC-2017 or MRC-2022
Supplemental Table 2 describes induction outcomes by induction regimen and AML-MRC classification status. The overall response rate (ORR) was 61.7 % after induction, including CR and CR with incomplete recovery, in the entire cohort. Irrespective of induction regimen received, the ORR for the entire cohort was 61.7% (306/496) as compared with MRC-2017 at 38.1% (43/113; P < .001) and MRC-2022 at 46.3% (68/147; P < .001). There was no statistically significant difference in ORR between MRC-2017 and MRC-2022 (P = .207). Among patients meeting MRC-2017 criteria, the ORR was 38.4% (28/73) for 7+3 and 41.2% (14/34) for CPX-351. The 30-day mortality for patients meeting MRC-2017 criteria was 2.7% (2/73) for 7+3 and 11.8% (4/34) for CPX-351. Among patients meeting MRC-2022 criteria, the ORR was 48.3% (56/116) for 7+3 and 40.7% (11/27) for CPX-351. The 30-day mortality for patients meeting MRC-2022 criteria was 5.2% (6/116) for 7+3 and 7.4% (2/27) for CPX-351. Supplemental Table 3 describes salvage regimens received for relapsed or refractory disease.
Patients meeting MRC-2017 criteria (n = 113) had a worse EFS as compared with those in the non–MRC-2017 group (median, 1.4 vs 11.6 months, P < .001; Figure 2A). Similarly, patients meeting MRC-2022 criteria (n = 147) had a worse EFS (median, 1.9 vs 10.5 months, P = .019; Figure 2B). Patients meeting MRC-2017 criteria also had a worse OS (median, 19.7 vs 54.1 months, P < .001; Figure 2C). There was no difference in OS for patients who met MRC-2022 criteria vs those who did not (median, 46.5 vs 36.5 months, P = .890; Figure 2D). If patients with TP53 mutations were removed from the MRC-2017 group (n = 26), the effect on EFS was preserved (median, 1.6 vs 11.8 months, P = .003; supplemental Figure 1A), but no difference in OS was seen between those in the MRC-2017 and non–MRC-2017 groups (median, 27.8 vs 54.1 months, P = .140; supplemental Figure 1B). After removing patients with TP53 mutation, the hazard ratio (HR) for EFS was 1.51 (95% confidence interval [CI], 1.15-1.98; P = .003) and for OS was 1.28 (95% CI, 0.92-1.78; P = .138). With censoring at the time of allo-HCT, patients meeting the criteria of either MRC-2017 or MRC-2022 had significantly worse EFS (supplemental Figure 2A-B) and OS (supplemental Figure 2C-D). When comparing patients meeting MRC-2022 criteria with only those with ELN 2022–defined intermediate-risk disease, we observed no differences in EFS (median, 1.9 vs 8.7 months; P = .350; supplemental Figure 3A) or OS (median, 46.5 vs 27.7 months; P = .400; supplemental Figure 3B). Baseline differences in induction and allo-HCT rates between MRC-2017 and MRC-2022 groups are displayed in supplemental Table 4. There was no difference in the rate of allo-HCT between the MRC-2017 and ELN 2022 intermediate-risk (54.0% vs 65.9%, P = .49) or MRC-2022 and ELN 2022 intermediate-risk (67.3% vs 67.2%, P = .81) groups.
Figure 2.
Survival outcomes among 496 intensively treated patients with AML when classified by either MRC-2017 or MRC-2022 criteria. (A) EFS among patients classified with MRC-2017. (B) EFS among patients classified with MRC-2022. (C) OS among patients classified with MRC-2017. (D) OS among patients classified with MRC-2022.
Heterogeneous survival outcomes within patients classified as having AML-MRC based on mutational status
Among patients meeting MRC-2022 criteria, there was no differences in EFS or OS between patients who met criteria for MRC-mut vs MRC-cyt (supplemental Figure 4A-B). Within the MRC-mut group, there were no differences based on the presence of 1, 2, or >2 MRC-defining mutations considering either EFS or OS (supplemental Figure 4C-D). The waterfall plot in Figure 3A details the overlap between mutations in the MRC-mut population, with RUNX1 being the most common mutation. A significant difference in EFS was observed within the MRC-mut group based on the biological pathway of the gene involved. Patients with splice/TF mutations (n = 27) had a worse EFS compared with chromatin modifier (n = 25; median, 1.10 vs 7.40 months; pairwise P = .006) or mixed (n = 74; median, 1.10 vs 5.35 months; pairwise P = .006) mutational patterns (Figure 3B). There was no difference in median OS between patients with splice/TF vs chromatin or mixed mutational groups (Figure 3C). Using the publicly available BEAT AML data set, we identified a significantly different OS based on mutational class (global P = .036). Pairwise comparisons showed no difference for chromatin vs splicing (median, 45.67 vs 26.10 months; pairwise P = .239) but did show differences for chromatin vs mixed (45.67 vs 11.90 months; pairwise P = .004) and chromatin vs TF (including mixed; 45.67 vs 17.10 months; pairwise P = .019) groups (supplemental Figure 5A). Further considering the mixed category in the original data set, we separated this into different combinations of MRC mutations. When comparing patients who had a combination of a chromatin and splice/TF mutation vs only a splice/TF mutation we observed no differences in EFS or OS (supplemental Figure 5B-C).
Figure 3.
Comutation patterns and survival outcomes among patients classified with AML MRC-2022 by mutational criteria with pairwise comparisons. (A) Heat map demonstrating the comutational patterns among patients classified with MRC-2022 by mutational criteria (MRC-mut); (B) EFS stratified by mutation pathway for MRC-mut; (C) OS stratified by mutational pathway for MRC-mut. Benjamini-Hochberg correction was used for pairwise comparisons.
The current ELN 2022 guidelines recommend that patients with favorable-risk disease with a co-occurring MRC-related mutation or cytogenetic change be considered as favorable risk. We did not observe a significant difference in EFS for those with favorable risk with an MRC-related change (n = 34) as compared with those without (n = 116; median, 45.9 vs 19.0 months; P = .830; Figure 4A). Similarly, no differences in OS were seen for patients with favorable risk with a MRC-related change vs those without (median, 45.9 months vs not reached; P = .580; Figure 4B).
Figure 4.
Survival outcomes for patients with favorable-risk AML with co-occurring MRC cytogenetics or mutations. (A) EFS for patients with ELN 2022 favorable-risk AML who have co-occurring MRC-related cytogenetic or mutational changes. (B) OS for patients with ELN 2022 favorable-risk AML who have co-occurring MRC-related cytogenetic or mutational changes.
We examined the impact of reclassifying patients with MRC-mut who carried only a chromatin modifier mutation into the non-MRC group on the risk stratification model. We observed comparable differences in EFS when comparing patients who were classified based on the modified MRC-2022 criteria (median, 1.7 vs 10.5 months; P = .003; supplemental Figure 6A). Similarly, there was no difference in OS when comparing patients classified with the modified MRC-2022 (supplemental Figure 6B). When comparing the binary classification of MRC-2022 status and a linear variable of EFS time with a receiver operator curve, we observed an area under the curve of 0.433 (supplemental Figure 6). This was comparable with the area under the curve observed with the modified MRC-2022 classification, at 0.423 (supplemental Figure 6).
Long-term outcomes based on IC choice, stem cell transplantation, and posttransplant outcomes
Within the MRC-2017 group, we did not identify a statistically significant EFS difference for CPX-351 (n = 34) vs other cytarabine-based induction regimens (n = 79; median 1.8 vs 1.2 months; P = .370; Figure 5A) and, similarly, did not identify a significant OS difference (median, 24.4 vs 19.7 months; P = .690; supplemental Figure 7A). Within the MRC-2022 group, we also did not identify a statistically significant EFS difference for CPX-351 (n = 27) vs other cytarabine-based induction regimens (n = 120; median 1.9 vs 1.6 months; P = .820; Figure 5B) or a difference in OS (median, 26.8 vs 49.5 months; P = .430; supplemental Figure 7B).
Figure 5.
Survival outcomes among 496 intensively treated patients with AML, according to induction regimen received and landmark analysis after allo-HCT. (A) EFS for patients with MRC-2017–classified AML according to induction regimen type; (B) EFS for patients with MRC-2022–classified AML according to induction regimen type; (C) landmark analysis from the time of allo-HCT for patients meeting MRC-2017 criteria with 2-year OS noted; (D) landmark analysis from the time of allo-HCT stratified by MRC-2022 status with 2-year OS noted. Cyt-based, cytarabine based.
To examine the outcomes of patients with AML-MRC who received allo-HCT, we performed a landmark analysis for OS after transplantation. The rate of allo-HCT was lower in patients meeting MRC-2017 criteria (54.0%) vs MRC-2022 (67.3%) and this was statistically significant (P = .03). There was no difference in posttransplantation OS for receipt of CPX-351 or other cytarabine-based IC in the MRC-2017 subgroup (median, 58 months vs not reached; P = .800; supplemental Figure 7C) or the MRC-2022 subgroup (median, 58 months vs not reached; P = .230; supplemental Figure 7D). Patients meeting MRC-2017 criteria vs those not meeting MRC-2017 criteria had a posttransplant 2-year OS of 71.5% vs 66.7% (P = .504, χ2 test). The median posttransplant OS was not reached for either group (P = .430; Figure 5C). Patients meeting MRC-2022 criteria vs those who did not had a posttransplant 2-year OS of 77.0% vs 62.7% (P = .015). The median posttransplant OS was not reached for either group (P = .080; Figure 5D).
Univariate and multivariate Cox PHs analysis for EFS and OS
Univariate Cox PH analysis for EFS identified previous MDS (HR, 1.66; 95% CI, 1.18-2.32; P = .003) and any MRC mutation (HR, 1.31; 95% CI, 1.05-1.65; P = .019) as significant predictors. A multivariate analysis for EFS including allo-HCT as a time-dependent covariate revealed any ELN 2022–defined MRC mutation (HR, 1.55; 95% CI, 1.18-2.03; P = .001) and allo-HCT (HR, 0.46; 95% CI, 0.33-0.66; P < .001) as significant variables. Multivariate analysis for OS with allo-HCT as a time-dependent covariate identified age of >60 years (HR, 1.68; 95% CI, 1.28-2.20; P < .001) and allo-HCT (HR, 0.65; 95% CI, 0.48-0.88; P = .005) as significant. Dividing the MRC mutation group into TF/splice and chromatin modifier in the multivariate model revealed that only TF/splice mutation (HR, 1.65; 95% CI, 1.20-2.27; P = .002) and allo-HCT (HR, 0.49; 95% CI, 0.34-0.69; P < .001) remained significant (Table 2).
Table 2.
Multivariate analysis of EFS and OS for intensively treated patients with AML including MRC and allo-HCT as a time-dependent covariate
| Univariate analysis |
Multivariate analysis |
||||
|---|---|---|---|---|---|
| Variable | HR (95% CI) | P value | Variable | HR (95% CI) | P value |
| EFS | |||||
| Age of >60 years | 1.22 (0.98-1.51) | .073 | Age of >60 years | 1.21 (0.97-1.51) | .098 |
| Therapy-associated | 1.18 (0.76-1.84) | .466 | Therapy-associated | 1.19 (0.75-1.88) | .458 |
| Previous MDS∗ | 1.66 (1.18-2.32) | .003 | >2 MRC mutations | 0.78 (0.51-1.19) | .256 |
| >2 MRC mutations | 1.04 (0.72-1.52) | .823 | Any MRC mutation | 1.55 (1.18-2.03) | .001 |
| Any MRC mutation | 1.31 (1.05-1.65) | .019 | Received CPX-351 induction | 1.24 (0.86-1.77) | .250 |
| Received CPX-351 induction | 1.31 (0.93-1.85) | .120 | Allo-HCT (time-dependent) | 0.46 (0.33-0.66) | <.001 |
| OS | |||||
| Age of >60 years | 1.66 (1.27-2.16) | <.001 | Age of >60 years | 1.68 (1.28-2.20) | <.001 |
| Therapy-associated | 1.21 (0.73-2.00) | .468 | Therapy-associated | 1.02 (0.61-1.72) | .939 |
| Previous MDS∗ | 1.86 (1.28-2.70) | .001 | >2 MRC mutations | 0.89 (0.55-1.47) | .658 |
| >2 MRC mutations | 0.99 (0.64-1.54) | .977 | Any MRC mutation | 0.98 (0.71-1.36) | .912 |
| Any MRC mutation | 0.98 (0.74-1.29) | .894 | Received CPX-351 induction | 1.44 (0.94-2.21) | .096 |
| Received CPX-351 induction | 1.35 (0.90-2.03) | .149 | Allo-HCT (time-dependent) | 0.65 (0.48-0.88) | .005 |
| EFS: MRC mutations decomposed | |||||
| Chromatin modifier only | 0.81 (0.48-1.39) | .449 | Age of >60 years | 1.20 (0.96-1.5) | .111 |
| TF or splice mutation | 1.36 (1.06-1.75) | .017 | Therapy-associated | 1.20 (0.76-1.91) | .435 |
| >2 MRC mutations | 0.71 (0.45-1.12) | .138 | |||
| Chromatin modifier only | 1.00 (0.58-1.72) | .988 | |||
| TF or splice mutation | 1.65 (1.20-2.27) | .002 | |||
| Received CPX-351 induction | 1.24 (0.86-1.79) | .239 | |||
| Allo-HCT (time-dependent) | 0.49 (0.34-0.69) | <.001 | |||
History of previous MDS was not included in the multivariate analysis because of collinearity with other variables.
Discussion
In this real-world, retrospective, cohort study, patients with ICC 2022–defined AML-MRC are seen to form substantively different group than older MRC cohorts, as defined by WHO 2017. Specifically, the separation of the AML-TP53 group removed many patients who would have been previously classed within an AML-MRC group based on a complex karyotype, and the inclusion of MRC-related gene mutations resulted in a large proportion of previously non-MRC cases being reclassified as AML-MRC. The consequence is that less than half of cases are classed as AML-MRC by both criteria. This has significant implications on the interpretation of older literature and clinical trials examining this patient group and suggests that caution must be taken when generalizing results between studies using older classification criteria. We then identified that both the MRC-2017 and MRC-2022 criteria correctly identify patients with AML-MRC as having a worse EFS, mostly driven by higher rates of induction failure, consistent with an adverse-risk classification. However, although patients meeting MRC-2017 criteria also demonstrated a worse OS, patients with MRC-2022–defined disease had an OS comparable with that of the general AML population. This is discrepant with findings from previous validations of the ELN 2022, which largely demonstrate a negative impact on OS for patients classified per MRC-2022 criteria.6, 7, 8, 9, 10, 11, 12, 13 Further examining the MRC-2017 group, we find that the OS difference is largely driven by the inclusion of patients with TP53 mutation in this group.
More detailed assessment of the MRC-mut subgroup demonstrates significant heterogeneity in survival outcomes within this group. Although classification as MRC-mut or MRC-cyt or the number of MRC-related mutations did not have a significant impact on outcomes, we did identify that the biological pathway and specific genes involved is critical in determining the EFS impact of MRC-related mutations. Specifically, we found that patients with mutations in genes affecting chromatin modifier pathways have significantly better outcomes as opposed to patients with splicing or TF mutations. This effect was retained in our multivariate model, reinforcing the differential impact of these mutation types. Given the biological diversity of the genes included in the MRC-mut category, this is not entirely surprising. However, reclassifying patients with chromatin modifier mutations as non-MRC did not change the apparent prognostic ability of the MRC-mut classification, suggesting this relationship is complex but highlights the heterogeneity within the MRC-mut group. We did not identify specific subgroups with different survival in patients with mixed mutational patterns; however, we were limited by the number of patients in this group. Future studies should examine the impact of comutation status in patients with mixed MRC-related mutations. Simply reclassifying patients with chromatin modifier mutations alone as non-MRC did not significantly alter the prognostic model, suggesting that reformulating the MRC-mut subgroup may be more complex than declassifying 1 set of genes. We also did not identify a significant impact for the presence of MRC-related changes in patients with favorable-risk AML, consistent with the current ELN 2022 recommendations that such patients remain classified as favorable risk.
To assess further the reasons for the discrepant OS findings that we identified for the MRC-2022 group with previous literature, we examined clinical factors that have changed significantly in more recent time periods. Most validation studies for ELN 2022 examined patients being treated with conventional cytarabine- and anthracycline-based IC protocols, although there is significant heterogeneity in the regimens used in these studies.6, 7, 8, 9, 10, 11, 12, 13 CPX-351 is a liposomal formulation of cytarabine and daunorubicin that has been examined specifically in the MRC-2017 patient subgroup and has only been used widely since around 2018.14 Real-world studies of patients treated with CPX-351 have demonstrated outcomes comparable with those of the original phase 3 trial; however, only 70% of these patients had AML-MRC, with the remainder having more broadly therapy-related AML.14 We did not identify a significant EFS or OS benefit for CPX-351 over conventional cytarabine-based IC regimens for patients meeting MRC-2017 or MRC-2022 criteria. However, our study is retrospective in nature, which may lead to selection effects between the 2 treatment groups, and the number of patients treated with CPX-351 was small. Further studies on the role of CPX-351 in patients with AML-MRC based on the ICC 2022 criteria are required to clarify its role in this context.
Another significant practice change in recent years has been the wider accessibility of allo-HCT as an option for patients with AML because of the development of novel conditioning approaches and expanded donor options. The rates of allo-HCT in previous validation studies of the ICC 2022 criteria had a mean transplantation rate of 28.3% (range, 6%-45%),6, 7, 8, 9, 10, 11 whereas in our study, 56.5% of the overall population and 67.3% of patients meeting MRC-2022 criteria received allo-HCT. Using a landmark analysis for OS from the date of allo-HCT, we identified that patients meeting MRC-2017 criteria have outcomes comparable with those of the general AML population that received allo-HCT. These findings also may explain the discrepant impact of MRC-2022 classification on EFS and OS, because some patients meeting MRC-2022 criteria who have refractory disease may be successfully salvaged and be able to proceed to an allo-HCT with relatively favorable outcomes after this. The importance of allo-HCT to the overall treatment approach for these patients is further emphasized by our multivariate analysis that demonstrates the differential impact of MRC mutations on EFS as compared with OS and the consistent impact of all-HCT. Interestingly, patients with AML classified as MRC by the ICC 2022 criteria appear to have a better outcome after allo-HCT than the general AML population. This suggests that, although the MRC-2022 group may be at a higher risk of failing upfront therapy, those that are able to receive allo-HCT have relatively favorable outcomes as compared with other patients with AML. This outsized effect of allo-HCT and the high rates of allo-HCT in this group may partially explain the observed discrepancy in OS between our study and the previous literature.
Our study highlights the heterogeneity in outcomes within the ICC 2022 AML-MRC group, and that this is largely driven by differential outcomes within the group of patients with MRC-related mutations. We also identify that outcomes in the MRC-2017–defined group are driven by AML-TP53, which affects interpretation of older data using MRC-2017 criteria. We suggest that grouping these biologically distinct mutations into a single group may result in misclassification of patients as having adverse-risk disease when their outcomes are closer to that of the general AML population. ,Our analysis was not sufficiently powered to examine the impact of mutations at the level of individual genes, and future studies should address this. However, current treatment algorithms would be similar, with most patients selected for allo-HCT if feasible. Furthermore, we identify that allo-HCT has a disproportionate benefit for patients with AML-MRC as classified by the ICC 2022 criteria, as compared with the general AML population. We also suggest that the rate of allo-HCT is an important covariate to be considered when attempting to validate novel prognostic classifiers. Further studies are required to delineate the optimal classification and treatment approaches for patients with AML-MRC.
Conflict-of-interest disclosure: R.J.S. reports honoraria from AbbVie, Astellas, Pfizer, and Sobi; advisory board role with AbbVie, Bristol Myers Squibb, Jazz Pharmaceuticals, Kite/Gilead, and Novartis; speakers bureau participation for Kite/Gilead; and research funding from Jazz Pharmaceuticals. The remaining authors declare no competing financial interests.
Acknowledgments
Authorship
Contribution: R.J.S. conceived and supervised the study; N.C. performed data collection and analysis, and contributed to manuscript and figure preparation; A.F. and A.S. assisted with data collection; and all other authors provided patient data, and reviewed and approved the manuscript.
Footnotes
Presented, in part, in poster form at the 66th annual meeting of the American Society of Hematology, San Diego, CA, 7 to 10 December 2024.
Clinical data cannot be made publicly available for privacy reasons, but the corresponding author, Ryan J. Stubbins (ryan.stubbins1@bccancer.bc.ca), can be contacted to assess reasonable requests.
The full-text version of this article contains a data supplement.
Supplementary Material
References
- 1.Döhner H, Wei AH, Appelbaum FR, et al. Diagnosis and management of AML in adults: 2022 recommendations from an international expert panel on behalf of the ELN. Blood. 2022;140(12):1345–1377. doi: 10.1182/blood.2022016867. [DOI] [PubMed] [Google Scholar]
- 2.Arber DA, Orazi A, Hasserjian RP, et al. International Consensus Classification of myeloid neoplasms and acute leukemias: integrating morphologic, clinical, and genomic data. Blood. 2022;140(11):1200–1228. doi: 10.1182/blood.2022015850. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Arber DA, Orazi A, Hasserjian R, et al. The 2016 revision to the World Health Organization classification of myeloid neoplasms and acute leukemia. Blood. 2016;127(20):2391–2405. doi: 10.1182/blood-2016-03-643544. [DOI] [PubMed] [Google Scholar]
- 4.Döhner H, Estey E, Grimwade D, et al. Diagnosis and management of AML in adults: 2017 ELN recommendations from an international expert panel. Blood. 2017;129(4):424–447. doi: 10.1182/blood-2016-08-733196. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Khoury JD, Solary E, Abla O, et al. The 5th edition of the World Health Organization classification of haematolymphoid tumours: myeloid and histiocytic/dendritic neoplasms. Leukemia. 2022;36(7):1703–1719. doi: 10.1038/s41375-022-01613-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Ruhnke L, Bill M, Zukunft S, et al. Validation of the revised 2022 European LeukemiaNet risk stratification in adult patients with acute myeloid leukemia. Blood Adv. 2025;9(6):1392–1404. doi: 10.1182/bloodadvances.2024013304. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Rausch C, Rothenberg-Thurley M, Dufour A, et al. Validation and refinement of the 2022 European LeukemiaNet genetic risk stratification of acute myeloid leukemia. Leukemia. 2023;37(6):1234–1244. doi: 10.1038/s41375-023-01884-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Mrózek K, Kohlschmidt J, Blachly JS, et al. Outcome prediction by the 2022 European LeukemiaNet genetic-risk classification for adults with acute myeloid leukemia: an Alliance study. Leukemia. 2023;37(4):788–798. doi: 10.1038/s41375-023-01846-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Lachowiez CA, Long N, Saultz J, et al. Comparison and validation of the 2022 European LeukemiaNet guidelines in acute myeloid leukemia. Blood Adv. 2023;7(9):1899–1909. doi: 10.1182/bloodadvances.2022009010. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Sargas C, Ayala R, Larráyoz MJ, et al. Comparison of the 2022 and 2017 European LeukemiaNet risk classifications in a real-life cohort of the PETHEMA group. Blood Cancer J. 2023;13(1):77. doi: 10.1038/s41408-023-00835-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Song GY, Kim HJ, Kim T, et al. Validation of the 2022 European LeukemiaNet risk stratification for acute myeloid leukemia. Sci Rep. 2024;14(1):8517. doi: 10.1038/s41598-024-57295-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Jiménez-Vicente C, Charry P, Castaño-Diez S, et al. Evaluation of European LeukemiaNet 2022 risk classification in patients undergoing allogeneic haematopoietic stem cell transplantation for acute myeloid leukaemia: identification of a very poor prognosis genetic group. Br J Haematol. 2024;205(1):256–267. doi: 10.1111/bjh.19518. [DOI] [PubMed] [Google Scholar]
- 13.Zhou Q, Zhao D, Zarif M, et al. A real-world analysis of clinical outcomes in AML with myelodysplasia-related changes: a comparison of ICC and WHO-HAEM5 criteria. Blood Adv. 2024;8(7):1760–1771. doi: 10.1182/bloodadvances.2023011869. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Lancet JE, Uy GL, Cortes JE, et al. CPX-351 (cytarabine and daunorubicin) liposome for injection versus conventional cytarabine plus daunorubicin in older patients with newly diagnosed secondary acute myeloid leukemia. J Clin Oncol. 2018;36(26):2684–2692. doi: 10.1200/JCO.2017.77.6112. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Scott BL, Pasquini MC, Fei M, et al. Myeloablative versus reduced-intensity conditioning for hematopoietic cell transplantation in acute myelogenous leukemia and myelodysplastic syndromes-long-term follow-up of the BMT CTN 0901 clinical trial. Transplant Cell Ther. 2021;27(6):483.e1–483.e6. doi: 10.1016/j.jtct.2021.02.031. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Luznik L, O'Donnell PV, Symons HJ, et al. HLA-haploidentical bone marrow transplantation for hematologic malignancies using nonmyeloablative conditioning and high-dose, posttransplantation cyclophosphamide. Biol Blood Marrow Transplant. 2008;14(6):641–650. doi: 10.1016/j.bbmt.2008.03.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Jim HS, Syrjala KL, Rizzo D. Supportive care of hematopoietic cell transplant patients. Biol Blood Marrow Transplant. 2012;18(suppl 1):S12–S16. doi: 10.1016/j.bbmt.2011.10.029. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Cusatis R, Litovich C, Feng Z, et al. Current trends and outcomes in cellular therapy activity in the United States, including prospective patient-reported outcomes data collection in the Center for International Blood and Marrow Transplant Research registry. Transplant Cell Ther. 2024;30(9):917.e1–917.e12. doi: 10.1016/j.jtct.2024.06.021. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Bottomly D, Long N, Schultz AR, et al. Integrative analysis of drug response and clinical outcome in acute myeloid leukemia. Cancer Cell. 2022;40(8) doi: 10.1016/j.ccell.2022.07.002. 850.e9-864.e9. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.