Refined risk stratification helps guiding transplantation choice in adult BCR::ABL1-positive acute lymphoblastic leukemia

Dear Editor,

The integration of tyrosine kinase inhibitors (TKIs) into multiagent chemotherapy regimens has markedly improved the prognosis of BCR::ABL1-positive acute lymphoblastic leukemia (BCR::ABL1 + ALL). Allogeneic hematopoietic stem cell transplantation (allo-HSCT) remains a backbone of contemporary curative treatments for adult BCR::ABL1 + ALL.

IKZF1^plus is a genotype characterized by the coexistence of IKZF1 deletions with deletions in CDKN2A, CDKN2B, PAX5, or PAR1 in the absence of ERG deletion [1]. In 2023, the National Comprehensive Cancer Network (NCCN) ALL guideline officially stratified BCR::ABL1 + ALL into poor-risk [IKZF1^plus or antecedent chronic myeloid leukemia (CML)] or standard-risk (non-IKZF1^plus and without antecedent CML) category [2]. Simultaneously, minimal residual disease (MRD) is a well-recognized prognostic factor of ALL. However, there have been no studies on the synergy of the two prognostic factors (IKZF1^plus genotyping and MRD assessment) in risk stratification. Particularly, although increasing evidence suggests that allo-HSCT may not benefit all patients, the criteria for determining which patients should undergo or be spared from allo-HSCT remain elusive.

In this study, we aimed to refine the risk stratification by integrating IKZF1^plus genotyping and MRD assessment, thus improving the management of allo-HSCT in BCR::ABL1 + ALL therapy. Additionally, we preliminarily explored the molecular characteristics of IKZF1^plus ALL concerning its high-risk phenotype through targeted exome sequencing and RNA sequencing (RNA-seq), providing new insights into mechanistic research.

Here, we analyzed 156 adult patients with newly diagnosed BCR::ABL1 + ALL in three clinical trials conducted from 5 June 2014 to 15 November 2022. Till 31 December 2023, the overall median follow-up period was 34.6 months (range, 7.5–116.5). The patients were treated with a TKI-based standardized regimen, including imatinib or flumatinib (a second-generation TKI). In order to reduce high-dose chemotherapy-associated toxicity and early deaths, the relatively low intensity of chemotherapy was applied [3, 4]. There was no significant difference in baseline characteristics and survival outcomes between imatinib and flumatinib cohort, allowing a joint prognostic analysis (Supplemental Table 1, Supplemental Fig. 1A–C). In accordance with the Declaration of Helsinki, the study was approved by the Ethics Committee of Ruijin Hospital. [Clinical Trial Registration Number: ChiCTR-ONRC-14004968 [5], ChiCTR2100042248 [6, 7], and ChiCTR2100044308].

A landscape of gene deletions, mutations, and clinical data of the 156 patients was illustrated in Fig. 1A and Supplemental Fig. 2A, B. The patients were classified into IKZF1^plus group (n = 54) or non-IKZF1^plus group (n = 102) [composed of IKZF1 deletion alone (n = 65) and no IKZF1 deletion (n = 37)], according to multiplex ligation-dependent probe amplification (MLPA) results (Fig. 1A). Internal validation (Supplemental Table 4) and whole genome sequencing (Supplemental Fig. 2C) were performed to validate the MLPA technique. Clinical information was provided in Supplemental Tables 2 and 3. The IKZF1^plus group revealed micro-deletions in CDKN2A/2B (34.6%), PAX5 (31.4%), ERG (0%), and PAR1 region (0%) gene loci.

Fig. 1: The role of allo-HSCT for three distinct risk groups according to the refined stratification system combining IKZF1 genotype and MRD status in adult BCR::ABL1 + ALL.

A The profile of gene deletions, gene mutations, and clinical characteristics of 156 adult BCR::ABL1 + ALL patients. Patients are divided into three groups according to IKZF1 genotypes. Gray columns indicate patients, and rows include three panels: IKZF1^plus panel (IKZF1, CDKN2A/2B, PAX5, ERG gene, and PAR1 region), deletion panel (BTG1, RB1, ETV6, and EBF1 gene), and mutation panel (SETD2, RUNX1, PAX5, IKZF1, and TP53 gene). Alteration frequencies in each are shown on the right. The last heatmap shows clinical information. B–D Overall survival, event-free survival, and cumulative incidence of relapse for patients (aged 18–64 years) grouped by revised risk stratification: low-risk (non-IKZF1^plus/MRD−), intermediate-risk (non-IKZF1^plus/MRD+) or high-risk (IKZF1^plus), respectively. E–G Effects of transplantation on overall survival for three distinct risk groups (aged 18–64 years), respectively. P values are calculated using the log-rank test. The number at risk is based on all evaluable patients. Abbreviations: MLPA, multiplex ligation-dependent probe amplification; TES, targeted exome sequencing. *MRD− refers to patients achieving CMR (evaluated by RT-qPCR with a sensitivity of 0.001%) at 3 months of treatment, while MRD+ refers to patients failing to achieve CMR at 3 months in this study.

139 patients (aged 18-64 years) were included in the survival analysis. The OS, EFS, and CIR of IKZF1^plus group (n = 48) were markedly worse than those observed in non-IKZF1^plus group (n = 91) (Supplemental Fig. 3A–C).

MRD evaluation was performed for 125 patients at 3 months of treatment by RT-qPCR (sensitivity: 0.001%). Outcomes of patients who failed to achieve complete molecular remission (CMR) at 3 months (MRD+, n = 68) were significantly worse than those with CMR at 3 months (MRD−, n = 57) (Supplemental Fig. 3D–F).

Next, incorporating the two risk factors, 125 patients (aged 18–64 years) with available data were stratified into four subgroups [non-IKZF1^plus/MRD− (n = 37), non-IKZF1^plus/MRD+ (n = 43), IKZF1^plus/MRD− (n = 20) and IKZF1^plus/MRD+ (n = 25)] (Supplemental Fig. 4A–C). The non-IKZF1^plus/MRD− patients were classified into the low-risk group, demonstrating the best survival outcomes and the lowest relapse rate. In contrast, IKZF1^plus/MRD−, IKZF1^plus/MRD+, and IKZF1^plus/MRD not available (n = 3) subgroups had the worst outcomes and highest relapse rates, therefore combined into the high-risk group (IKZF1^plus, n = 48), indicating MRD assessment did not contribute to the risk stratification of IKZF1^plus patients. The remaining non-IKZF1^plus/MRD+ subgroup was stratified as an intermediate-risk group (n = 43). Finally, compared to the intermediate-risk or high-risk group, the low-risk group had better OS (P = 0.10 or P < 0.001), EFS (P = 0.005 or P < 0.001), and lower CIR (P = 0.01 or P < 0.001), respectively (Fig. 1B–D, Supplemental Table 5). Overall, the refined risk stratification could better discriminate the prognosis of adult BCR::ABL1 + ALL patients.

Therefore, based on our refined risk stratification system, 128 patients (aged 18-64 years) eligible for transplantation were analyzed. In low-risk group, OS of allo-HSCT subgroup was similar to non-HSCT subgroup (P = 0.74) (Fig. 1E), suggesting that low-risk patients did not benefit from transplantation. Among 23 patients transplanted, three died of infections post allo-HSCT. Among 14 patients without transplantation, one succumbed to heart failure and another due to leukemia progression.

For high-risk patients, OS of allo-HSCT subgroup was significantly higher than non-HSCT subgroup (P < 0.001) (Fig. 1G), confirming the necessity of transplantation in high-risk group. Finally, among intermediate-risk patients, OS of allo-HSCT subgroup showed a more favorable trend (P = 0.08) (Fig. 1F). Without allo-HSCT, OS of low-risk group was significantly superior to intermediate-risk and high-risk groups, respectively (Supplemental Fig. 5E). While after allo-HSCT, OS of three groups reached a comparable level (P = 0.85) (Supplemental Fig. 5F). Overall, our data suggest that allo-HSCT should be spared for low-risk group but is recommended for the other two groups (Supplemental Fig. 6).

Multivariate analysis demonstrated that IKZF1^plus genotype (all P < 0.001) and MRD positivity at 3 months (P = 0.01, P < 0.001, P < 0.001) were independent risk factors for OS, EFS, and RFS, while allo-HSCT (all P < 0.001) was an independent protective factor (Supplemental Fig. 7).

Based on RNA-seq data from 137 patients, the developmental stages of all patients were inferenced by using diffusion map-based dimensionality reduction (Supplemental Fig. 8) and lineage signatures (top 100 genes) in Bastian et al. [8]. It is consistent with previous reports, the non-IKZF1^plus group and IKZF1^plus group enriched in distinct hematopoietic lineage, including B lymphoid/myeloid lineage, early pro-B cells, and pre-B I/II large stages (Fig. 2A–C) [9, 10]. Using multiparameter flow cytometry [11], similar results at the protein expression level were validated (Fig. 2D).

Fig. 2: Transcriptome landscape of IKZF1^plus in adult BCR::ABL1 + ALL.

A Diffusion map for visualization of cell lineage differentiation through dimensionality reduction. The top 100 differentially expressed genes (DEGs) between multilineage and lymphoid cluster found by Bastian et al. [8] are subjected to diffusion map analysis, and the first three diffusion components are displayed using a three-dimensional plot. Each point stands for a patient (n = 137). Top left panel labels points with IKZF1^plus (n = 49) or non-IKZF1^plus (n = 88) group. Other panels show scaled scores calculated based on single sample gene set enrichment analysis (ssGSEA) method, including B lymphoid lineage markers (CD19, MS4A1, CD22), myeloid lineage markers (MECOM, MPO, and ANPEP) and CDKN2A gene. B ssGSEA score between IKZF1^plus and non-IKZF1^plus group, using gene signatures associated with three transcriptomic classes that correspond to early pro-B, pro-B and pre-B I differentiation stages in Kim et al. [9]. C ssGSEA score between IKZF1^plus and non-IKZF1^plus group, using seven B lymphopoiesis stage-specific gene sets identified by Beder et al. [10]. D Box plots illustrating expression of B lymphoid and myeloid lineage antigens between IKZF1^plus and non-IKZF1^plus group, as measured by multiparameter flow cytometry (n = 135, 2 patients without flow cytometry results). E Volcano plot shows DEGs between IKZF1^plus and non-IKZF1^plus group. Each dot represents one gene. Genes significantly upregulated in IKZF1^plus are colored in red, and downregulated in IKZF1^plus are colored in blue. F Lollipop chart visualizes enriched pathways or terms that are upregulated and downregulated in IKZF1^plus group, utilizing Gene Ontology, Kyoto Encyclopedia of Genes and Genomes, and Reactome database. G Enrichment score of various cell types between IKZF1^plus and non-IKZF1^plus group, using cell type-specific genes from single-cell RNA sequencing data of human normal hematopoiesis cells reported by Zhang et al. [12]. Abbreviations: n.s. no significant, HSC/MPP hematopoietic stem cell/multi-potential progenitor, GMP granulocyte–macrophage progenitor, E/B/M eosinophil/basophil/mast cell progenitor, MEP megakaryocyte and erythroid progenitor, immature Neu immature neutrophil, mature Neu mature neutrophil, cDC conventional dendritic cell, pDC plasmacytoid dendritic cell. P values are calculated using Wilcoxon rank-sum test. *P < 0.05, **P < 0.01, ***P < 0.005, and ****P < 0.001.

Several unreported signature genes of IKZF1^plus group were identified in this work for the first time (Fig. 2E), including upregulated (TUNAR, GPR68, TAFA1, CD24, and TUBB3) or downregulated genes (MARCKS, CDKN2A, and TBL1X) (Supplemental Table 6). IKZF1^plus genotype was associated with activation of proliferative pathways [PI3K-Akt and Ras signaling pathways] and inhibition of hematopoietic/immune functions [T cell activation and other immune system process] (Fig. 2F). The above findings then were validated by additional gene sets [12] and immunocyte deconvolution (xCell algorithm) (Fig. 2G, Supplemental Fig. 9A) [13]. Consistently, peripheral blood samples of IKZF1^plus group showed significantly lower expression of a T cell antigen CD3 (P = 0.04) (Supplemental Fig. 9B). Therefore, better immune function may be a positive factor contributing to the improved prognosis of non-IKZF1^plus group.

This study is the largest cohort to date focusing on IKZF1^plus aberration and MRD simultaneously in adult BCR::ABL1 + ALL. Integrating both IKZF1 genotype and MRD status at 3 months, BCR::ABL1 + ALL could be further classified into three distinct risk groups: low-risk, intermediate-risk, and high-risk. The refined stratification system not only essentially meets the criterion for 2023 NCCN’s poor-risk category, but also allows for a subdivision of NCCN’s standard-risk category into two separate entities: low-risk and intermediate-risk.

Low-risk group (non-IKZF1^plus/MRD−), for the first time, was identified in adult BCR::ABL1 + ALL. Prior studies indicated that BCR::ABL1 + ALL patients attaining CMR at 3 months had favorable survival and did not benefit from transplantation [14, 15]. In contrast, the scenario for IKZF1^plus/MRD− patients was distinct: despite achieving 3-month CMR, prognosis of these high-risk patients was still significantly worse than non-IKZF1^plus patients, and transplantation remarkably improved their OS (Supplemental Fig. 5G). This underscored that for IKZF1^plus genotype with an inherent propensity for relapse, achieving MRD negativity may not be sufficient to mitigate the adverse impact of a poor genetic profile. The classification of low-risk group was only restricted to non-IKZF1^plus patients who were MRD negative at 3 months, which did not benefit from allo-HSCT and should avoid it.

Within intermediate-risk group (non-IKZF1^plus/MRD+), compared to non-HSCT subgroup, there was a promising trend towards better OS in allo-HSCT subgroup. It is advisable to employ targeted immunotherapies to eradicate MRD, and subsequently bridge allo-HSCT. To more comprehensively elucidate the role of allo-HSCT in this group, further studies with larger sample size and extended follow-up are warranted.

High-risk group actually consisted of all IKZF1^plus patients, irrespective of MRD at 3 months. Allo-HSCT subgroup consistently exhibited significantly better OS than non-HSCT counterparts, therefore allo-HSCT was crucial for high-risk patients by reducing relapse risk (Supplemental Fig. 5H).

In conclusion, the refined risk stratification based on IKZF1^plus genotyping and MRD assessment helps to guide transplantation decisions for adult BCR::ABL1 + ALL patients. Allo-HSCT should be spared for the low-risk group, but is recommended for intermediate-risk and high-risk groups (Supplemental Fig. 6). More comprehensive clinical studies are warranted for further validation of our findings.