To the Editor:

Advancements in tyrosine kinase inhibitors [1], novel agents such as blinatumomab [2], and allogeneic hematopoietic cell transplantation (allo-HCT) have improved the survival of patients with Philadelphia chromosome (Ph)-positive acute lymphoblastic leukemia (ALL) [3]; however, outcomes remain poor for those who experience relapse, especially after HCT [4]. Recent advancements in genetic techniques have deepened insights into gene mutations linked to ALL development and resistance [5]. Key genetic alterations in lymphoid differentiation, tumor suppression, and cell cycle regulation suggest their potential as prognostic markers. Among these, IKZF1 deletions (IKZF1del) have attracted significant attention due to their association with impaired lymphoid development and a high relapse rate [6]. Another notable genetic alteration is the deletion of cyclin-dependent kinase inhibitor 2 (CDKN2) [7], a known tumor suppressor gene. However, while most studies report a negative clinical impact, some discrepancies remain.

We aimed to evaluate the prognostic significance of IKZF1del and CDKN2del in a well-defined cohort of adults with Ph-positive ALL who underwent allo-HCT after receiving imatinib-based intensive chemotherapy as described in previous studies [8]. Between April 2018 and December 2022, a total of 156 adult patients were retrospectively analyzed. Genetic profiling included MLPA to detect deletions in IKZF1, CDKN2A, CDKN2B, PAX5, BTG1, EBF1, ETV6, JAK2, RB1, and the PAR1 region, as well as NGS targeting 73 genes. Measurable residual disease (MRD) monitoring for BCR::ABL1 transcript was conducted by RT-qPCR at key treatment milestones, with a sensitivity of 10−5. Relapses included a significant MRD increase of at least 1-log. Poor molecular response (PMR) was defined with values exceeding 0.1% of detected MRD, while complete molecular response (CMR) was defined as absence of detectable MRD. Major molecular response (MMR) was characterized by values between these CMR and PMR criteria. Allo-HCT was the standard post-remission therapy for high-risk adult ALL including Ph-positive ALL and other cytogenetics like complex karyotype and MLL rearrangement, and delayed remission after at least 2 cycles of induction chemotherapy. Thus, all patients in this study were candidates for allo-HCT, but it was not performed when a suitable donor was unavailable or when the patient’s performance status was insufficient. The primary end points were overall survival (OS), disease-free survival (DFS), cumulative incidence of relapse (CIR), and non-relapse mortality (NRM) in patients who underwent allo-HCT after successful chemotherapy. All patients provided written informed consent approved by the Institutional Review Board of The Catholic University of Korea (KC16TISI0438, KC23RASI0119, KC24RISI0772). This study was conducted in accordance with the Declaration of Helsinki.

Baseline characteristics of the entire cohort are detailed in Supplementary Table 1. The most frequent gene deletion detected by MLPA was IKZF1del (80.8%), followed by CDKN2del (39.1%), and PAX5del (38.5%). Among the 73 targeted genes analyzed by NGS, the most frequently observed gene mutation was in SETD2 (7.1%), followed by RUNX (5.8%) and IKZF1 (5.8%). One hundred and twenty-seven patients (81.4%) underwent allo-HCT, all of whom were in complete remission (CR) at the time of allo-HCT. Pre-HCT MRD assessment showed CMR in 69 (54.3%), MMR in 31 (24.4%), and PMR in 27 (21.3%). Among those genetic alterations, our data revealed CDKN2del was the only genetic alteration associated with poor survival outcomes (Supplementary Fig. 1). Also, we identified that the survival outcome of IKZF1del was poor only when it was combined with CDKN2del (Supplementary Fig. 2).

Accordingly, we moved our focus to CDKN2del. As a large proportion of CDKN2del co-occurred with IKZF1del, we classified all patients into three subgroups – Group 1 with CDKN2del(−)/IKZF1del(+) (n = 72), Group 2 with CDKN2del(−)/IKZF1del(−) (n = 23), and Group 3 with CDKN2del(+)±IKZF1del (n = 61) to validate the prognostic impact of the combinatorial effect of IKZF1del and CDKN2del. Among the three subgroups, CR rate and MRD response at the time of allo-HCT were not significantly different, whereas concurrent genetic alterations and mutations showed some differences (Table 1). The median follow-up of the entire cohort was 43.4 months (range 2.6-77.5). Significant differences in OS (p = 0.010) and DFS (p = 0.049) were observed among the three subgroups, primarily due to the poor outcomes of Group 3 (CDKN2del regardless of IKZF1del). The 4-year OS and DFS rates were 65.9% and 38.5% in Group 1, 52.6% and 22.0% in Group 2, and 33.8% and 11.8% in Group 3, respectively. A trend toward higher CIR in Group 3 was observed (p = 0.156), while no significant differences were found in NRM (p = 0.599) (Supplementary Fig. 3).

Table 1 Baseline characteristics according to the CDKN2del and IKZF1del combinatorial subgroups.
Full size table

Since allo-HCT was recommended as a post-remission therapy for all eligible patients, we analyzed outcomes in the transplanted cohort. The estimated 3-year OS, DFS, CIR, and NRM rates were 63.4%, 42.9%, 40.8%, and 16.3% after allo-HCT, respectively. Survival outcomes were significantly different among the three groups (Supplementary Fig. 4) – 3-year OS rates of Group 1, 2, and 3 were 73.6%, 60.9%, and 43.5% (p = 0.017), and 3-year DFS rates were 56.4%, 46.3%, and 23.5% (p = 0.006), respectively. CIR was highest in Group 3 (57.0%) compared to Group 1 and 2 (28.4% and 43.7%, respectively; p = 0.024). To further assess the adverse impact of CDKN2del, we additionally analyzed the outcomes in patients who underwent allo-HCT in CMR (Fig. 1). Group 3 still showed significantly inferior DFS and higher CIR than the other groups, with 3-year DFS rate of 27.9% compared to Group 1 and 2 (64.2% and 60.6%; p = 0.009) and CIR rate of 47.2% compared to Group 1 and 2 (21.2%, 13.6%; p = 0.013).

Fig. 1: Survival outcomes of the patients treated with allo-HCT in CMR according to the CDKN2del and IKZF1del combinatorial subgroups.
Fig. 1: Survival outcomes of the patients treated with allo-HCT in CMR according to the CDKN2del and IKZF1del combinatorial subgroups.
Full size image

A Disease-free survival. B Cumulative incidence of relapse.

We performed multivariate analyses in the entire cohort (Supplementary Table 2) and transplantation cohort (Supplementary Table 3) to analyze the factors affecting survival outcomes. In the transplantation subgroup, high leukocyte counts at diagnosis (HR 2.02, 95% CI: 1.10–3.69, p = 0.023) and CDKN2del (HR 2.51, 95% CI: 1.27–4.95, p = 0.008) were significantly associated with poorer OS. Similarly, CDKN2del (HR 2.27, 95% CI: 1.34–3.84, p = 0.002) and pre-HCT PMR (HR 1.88, 95% CI: 1.06–3.33, p = 0.030) were associated with poorer DFS and higher CIR rates (CDKN2del – HR 2.29, 95% CI: 1.21–4.34, p = 0.011 and pre-HCT PMR – HR 3.01, 95% CI: 1.56–5.79, p = 0.001). Older age remained as a significant risk factor for high NRM (HR 3.13, 95% CI: 1.24–7.87, p = 0.015).

With the increasing availability of diverse treatment options for adult ALL, the assessment of genetic abnormalities has become more crucial in establishing a well-defined risk stratification system. Our study identified CDKN2del as a significant genetic alteration associated with a higher relapse rate and poorer survival outcomes, even more than the well-recognized IKZF1del. Furthermore, we found that IKZF1del alone did not impact survival outcomes unless concomitant CDKN2del was present. This demonstrated that CDKN2del might be considered a primary prognostic indicator regardless of IKZF1del status, as reflected in Group 3. Also, this finding indicates that the adverse prognostic significance traditionally attributed to IKZF1del may actually stem from the frequent co-occurrence of CDKN2del, which appears to primarily drive the poor outcomes observed. Nevertheless, the clinical significance of CDKN2del should be interpreted with caution, as CDKN2del showed a higher frequency of concomitant deletions, including PAX5, BTG1, and EBF1. Although limited by a small number of patients, our data suggest concomitant genetic alterations did not substantially contribute to the poor survival outcomes observed in patients with CDKN2del (Supplementary Fig. 5). These findings highlight the prognostic relevance of CDKN2del.

Our data finally suggested that CDKN2del was associated with poor survival outcome regardless of the impact of IKZF1 co-deletion. This result was based on the poor outcomes of both CDKN2del alone and CDKN2del(+)/IKZF1del(+) groups, showing that all CDKN2del alone relapsed even after allo-HCT with a higher CIR rate (Supplementary Fig. 6). However, the small number of patients with CDKN2del alone (n = 7) may limit clear conclusions and further studies are needed to validate the results by clarifying the role of IKZF1 co-deletion. In a previous study by Pfeifer et al. [9], CDKN2del was identified as the sole genetic abnormality linked to poor survival outcomes and an independent predictor of a shorter remission duration, consistent with our findings. However, a notable distinction in our study is that the adverse prognostic impact of CDKN2del persisted even in patients who achieved pre-HCT CMR. Despite the favorable response to chemotherapy, the high relapse rate observed in patients with CDKN2del may be attributed to its role in promoting clonal evolution [10], enabling leukemic subclones with enhanced survival advantages to emerge and drive relapse. Additionally, the increased relapse rate following allo-HCT may be partially attributed to higher PD-L1 expression in patients with CDKN2del [11], potentially contributing to the immune escape process in ALL. Moreover, our study assessed MRD using RT-qPCR, but recent studies suggest that NGS-based detection offers superior sensitivity [12]. Further validation is needed to refine MRD assessment and better predict relapse risk in patients with CDKN2del.

In this study, IKZF1del did not correlate with prognosis, especially in patients without CDKN2del. This may suggest a weaker prognostic impact of IKZF1del in Ph-positive ALL in the Korean population or may be due to the lack of a detailed IKZF1 isoform analysis in our study. Fedullo et al. [13] reported that only dominant-negative IKZF1del isoforms had prognostic significance, while a study from our institute [14] found no significant survival differences between dominant-negative and haplo-insufficient IKZF1del subtypes. Further research is needed to clarify the impact of IKZF1del through a more detailed isoform analysis.

In conclusion, this study underscores the substantial prognostic impact of CDKN2del in Ph-positive ALL. As our patients were treated with imatinib-based treatment, our results may not apply to a current ponatinib-based treatment era, but recent studies showed that IKZF1del alone did not affect survival even with ponatinib [15] Nevertheless, concomitant genetic alterations such as CDKN2del alongside IKZF1del, which is now defined as IKZF1-plus, was associated with significantly poor outcomes. These findings highlight the need for tailored strategies for this high-risk group with genetic alterations.