Ph-like B-cell acute lymphoblastic leukemia in a Mexican pediatric cohort: Molecular features and prognostic implications

Abstract Introduction: Genomic and transcriptomic studies have highlighted the biological heterogeneity of B-ALL, allowing identification of clinically relevant subtypes. Ph-like B-ALL is a high-risk subtype with a BCR::ABL1-like expression profile. It is the most prevalent subtype among Hispanic/Latino children in the U.S. However, its prevalence, molecular features, and prognostic impact remain undefined in Mexican pediatric patients. This study aimed to characterize the prevalence, molecular landscape, and clinical significance of Ph-like B-ALL in this population. Methods: A total of 323 pediatric patients with de novo B-ALL diagnosed between 2018 and 2024 were included. Diagnoses were based on the 2022 International Consensus Classification of Myeloid Neoplasms and Acute Leukemias. Pretreatment bone marrow samples were used for RNA and DNA extraction. TruSight RNA Pan-Cancer Panel and Cicero, Arriba, and FusionCatcher algorithms were used for fusion detection. CRLF2 expression was assessed with Salmon and grouped via k-means clustering. DNA variants were identified using a custom ArcherDX panel. Cox regression models were applied to evaluate genetic predictors of Event-Free Survival (EFS), defined as time from diagnosis to relapse, death, or induction failure. Results: We performed RNA-seq in 299 patients and DNA-seq in all cases. Ph-like associated fusions were detected in 13.4% (40/299) of cases. Based on the broader criteria by Thai Hoa Tran et al. (PMID: 38657263), patients with CRLF2 overexpression and JAK1/2 or CRLF2 mutations and SH2B3 loss, or rare kinase fusions, were also classified as Ph-like. This raised the Ph-like frequency to 17.4% (52/299). JAK-STAT pathway alterations were the most common in Ph-like cases, present in 76.9% (40/52). These included CRLF2::P2RY8 (40%), CRLF2::IGH (12.5%), JAK2 fusions (5.0%), JAK2 (35%) and JAK1 (2.5%) mutations, CRLF2F232C(22.5%), and SH2B3 loss (2.5%). Several patients had co-occurring alterations, so percentages exceed 100%. ABL-class fusions were found in 19.2% (10/52) of Ph-like. Eight distinct fusions were identified: ETV6::ABL1 (2), ABL1::NUP214 (2), and ABL1::ZMIZ1, ABL1::FOXP1, RCSD1::ABL1, ARAP2::ABL1, ABL2::ZC3HAV1, and PDGFRA::FIP1L1 (each in one patient). ABL1 was the most frequently rearranged gene, present in 80% (8/10) cases. A novel kinase fusion, PRKACA::CCDC13, was detected in one case (1.9%) and classified as Ph-like based on its expression profile. CRLF2 expression was clustered into three distinct groups: low (43.1%, n=129), medium (42.5%, n=127), and high (14.7%, n=44). Among patients with high CRLF2 expression, 70.5% (31/44) had JAK-STAT pathway alterations, most commonly CRLF2::P2RY8 (45.2%), CRLF2F232C (22.6%), JAK2A683 (16.1%), JAK2T875N(9.7%), CRLF2::IGH (16.1%), and SH2B3 mutation (3.2%). In 13 of these cases, no Ph-like alterations were identified; two carried a PAX5::ZCCHC7 fusion, and one had a BCR::ABL1 fusion. Among patients with medium CRLF2 expression, 5 had JAK-STAT–related alterations (2 CRLF2 fusions, 2 JAK2 fusions, and 1 EPOR::IGH. Additionally, 8 out of 10 ABL-class cases and the PRKACA::CCDC13 fusion also exhibited medium CRLF2 expression. To assess the association between EFS and common B-ALL genetic alterations, univariate and multivariate Cox proportional hazards models were used. In univariate Cox regression analysis, the Ph-like subtype was significantly associated with poorer EFS (HR=1.89, 95% CI:1.15–3.09, p=0.04). This association remained significant in multivariate analysis (HR=2.16, 95% CI:1.24–3.78, p =0.03). Conclusion: Ph-like B-ALL is the most prevalent molecular subtype in this cohort and shows a trend toward poorer EFS, warranting further validation, although non-uniform treatment and the absence of tyrosine kinase inhibitor (TKI) therapy may have influenced this outcome. To our knowledge, this is the first report of the PRKACA::CCDC13 kinase fusion, which requires functional characterization. Integrating transcriptomic and genomic diagnostics into standard care is crucial for early identification of Ph-like B-ALL in Mexican patients, potentially enabling TKI-based therapies to improve outcomes.