Paper Details

ABSTRACT

Background:

The IKZF1^plus genomic classifier, commonly defined as IKZF1 deletion accompanied by selected additional gene deletions (e.g., IKZF1 deletion plus CDKN2A/B and/or PAX5 deletion), is an established adverse prognostic marker in Philadelphia chromosome–positive acute lymphoblastic leukemia (Ph+ALL). However, the relative prognostic contribution of individual partner deletions in combination with IKZF1 deletion remains unclear. We sought to identify prognostically relevant partner deletions in a large cohort of patients (pts) with newly diagnosed Ph+ALL.

Methods:

We retrospectively analyzed adults with newly diagnosed Ph+ALL who received frontline therapy at our institution with genome-wide copy-number profiling performed using SNP array, whole-exome sequencing, and/or optical genome mapping. Recurrent deletions were assessed using univariable Cox regression and Fine–Gray regression with Benjamini–Hochberg false discovery rate (FDR) correction. Multivariable Cox models were adjusted for age, WBC, and frontline treatment (hyper-CVAD plus tyrosine kinase inhibitor [TKI] vs. blinatumomab plus TKI).

Results:

A total of 181 pts with newly diagnosed Ph+ALL underwent genomic profiling. The median age was 54 years (range, 20 – 83) and median WBC at diagnosis was 9.4 x 10⁹/L (range, 0.7 – 658); 53 pts (29%) had WBC ≥ 70 x 10⁹/L. Frontline regimens were hyper-CVAD plus TKI (n=109 [imatinib, n=2; dasatinib, n=56; ponatinib, n=51]) or blinatumomab plus TKI (n=72 [dasatinib, n=15; ponatinib, n=53; asciminib, n=4]). The CR/CRi rate was 94%. The median follow-up was 76 months.

IKZF1 deletion was the most frequently identified deletion, occurring in 120 of 181 pts (66%), followed by CDKN2A/B in 63 (35%), PAX5 in 41 (23%), VPREB1 in 38 (21%), and EBF1 in 24 (13%). Among the 120 IKZF1-deleted pts, the most common co-deletions were CDKN2A/2B in 42 (35%), PAX5 in 33 (28%), VPREB1 in 31 (26%), and EBF1 in 21 (18%).

Among the genomic alterations detected in >10% of pts, only VPREB1 deletion was prognostic. Among the entire cohort (n=181), the 5-year overall survival (OS), event-free survival (EFS) and cumulative incidence of relapse (CIR) for pts with and without VPREB1 deletion were 52% and 73% (p=0.002), 43% and 62% (p=0.01) and 32% and 20% (p=0.10), respectively. In multivariate analysis, VPREB1 deletion remained independently associated with inferior OS (hazard ratio [HR], 2.06; p=0.01), inferior EFS (HR, 1.93; p=0.02), and a higher CIR (subdistribution HR, 3.01; p=0.006).

Within the IKZF1-deleted cohort, VPREB1 co-deletion was the only lesion significantly associated with inferior OS (FDR-adjusted p=0.02), although EFS and CIR were not statistically significant (FDR-adjusted p=0.12 and 0.27, respectively). On multivariate analysis of the entire cohort, IKZF1+VPREB1 co-deletion (vs. all other pts) remained significant for both OS (HR, 2.64; 95% confidence interval [CI], 1.41 – 4.93; p=0.002) and EFS (HR, 2.09; 95% CI, 1.16–3.72; p=0.01), with a trend towards higher CIR (p=0.06). In contrast to the poor prognosis observed with IKZF1+VPREB1 co-deletions, the standard definition of IKZF1^plus was not significantly associated with OS, EFS, or CIR (p=0.76, p=0.19, and p=0.58, respectively).

In the overall cohort, OS, EFS, and CIR differed across the four combined IKZF1/VPREB1 groups (IKZF1+VPREB1 co-deletion, n=31; IKZF1 deletion / no VPREB1 deletion, n=89; VPREB1 deletion / no IKZF1 deletion, n=7; neither deletion; n=54), with the co-deleted group showing the poorest outcomes (log-rank p=0.002, p=0.008, and p=0.02, respectively). The 5-year OS, EFS, and CIR rates for pts with IKZF1+VPREB1 co-deletion were 53%, 42% and 32%, respectively. In contrast, the 5-year OS, EFS, and CIR rates for pts with neither deletion were 76%, 74%, and 9%, respectively.

Conclusions:

In our cohort of pts with newly diagnosed Ph+ALL treated most frequently with ponatinib (57% of overall cohort), the standard definition of IKZF1^plus was not prognostic. In contrast, VPREB1 deletions, which occurred in 21% of pts with Ph+ALL and in 26% of IKZF1-deleted Ph+ALL, significantly impacted prognosis. IKZF1+VPREB1 co-deletions were independently associated with inferior survival and an increased risk of relapse. These findings suggest that individual partner lesions accompanying IKZF1 deletion may have distinct prognostic relevance and that VPREB1 deletions may inform genomic risk stratification in adult Ph+ALL, an observation that should be evaluated in other datasets.