Impact of Ph-like Gene Signatures on Clinical Outcomes in Adult B-ALL Patients in Resource-Constrained Settings

BCR::ABL1 (Ph⁺) fusion gene occurs in approximately 25% of adult acute lymphoblastic leukemia (ALL) cases. This molecular subtype was associated with poor clinical outcomes before the advent of tyrosine kinase inhibitors (TKIs). TKIs have significantly improved overall survival (OS), event-free survival (EFS), and complete remission (CR) rates. For instance, adding the TKI dasatinib to chemotherapy increased complete molecular response (CMR) rates and reduced relapse rates compared to chemotherapy alone. In 2016, the World Health Organization recognized a new provisional diagnostic entity called “Philadelphia-like” (Ph-like) or “BCR::ABL1-like” ALL. This B-ALL subtype, despite presenting a gene signature similar to Ph⁺ ALL, lacks the BCR::ABL1 fusion protein. Patients with this subtype have poor clinical outcomes and exhibit various rearrangements, mutations, and copy number variations involving kinase or cytokine receptor genes, activating JAK2/STAT, ABL1, and RAS signaling pathways. Although clinical trials have evaluated JAK- or ABL-directed TKIs in Ph-like ALL, a standard treatment approach has not been established. Identifying Ph-like ALL patients remains challenging as it requires methodologies that comprehensively evaluate gene expression. Chiaretti and colleagues (J Haematol. 2018;181(5):642-652) proposed a screening tool based on the expression of 10 genes measured by quantitative PCR and a statistical model (10-gene score) for identifying Ph-like ALL patients. This study investigated the expression of Ph-like-related genes in samples from healthy donors (n = 12) and adult patients with B-ALL (n = 83; Ph⁺ n = 33 and Ph^- n = 50) by quantitative PCR and their association with clinical and laboratory characteristics and survival outcomes. The statistical model proposed by Chiaretti et al. was used to calculate the scores. Dichotomization for the 10-gene score and individual gene expression was done using the ROC curve, with metrics such as the area under the curve (AUC) and the Youden index, using the R package cutpoint. Comparisons between groups were made using the Mann-Whitney test, Fisher's exact test, or chi-square test as appropriate. Survival analyses were conducted using the Kaplan-Meier method and Cox regression analysis, with a p-value < 0.05 considered statistically significant. Gene expression levels of CD97, CD99, CRLF2, IFITM1, IFITM2, NUDT4, SEMA6A, SOCS2, and TP53INP1 were higher in B-ALL patients compared to healthy donors (all p < 0.05). Additionally, the 10-gene score was higher in Ph⁺ B-ALL patients than in Ph^- B-ALL patients (p < 0.0001). In our cohort, Ph⁺ or Ph^- status did not predict the prognosis of B-ALL patients, but higher 10-gene scores were associated with reduced overall survival. Among Ph^- B-ALL patients, a high score was also linked to lower overall survival rates (p = 0.04). When evaluating Ph-like-related genes individually, only high IGJ expression was associated with worse clinical outcomes. In the entire B-ALL cohort, higher 10-gene scores were associated with lower LDH levels (p < 0.05). In the Ph^- B-ALL subgroup, higher scores correlated with fewer blasts infiltrating the bone marrow and lower LDH levels (p < 0.05). Notably, in both the entire B-ALL cohort and the Ph^- B-ALL subgroup, the 10-gene score was an independent prognostic factor in multivariate Cox regression analysis with covariates such as gender, age, hematimetric indices, LDH, and cytogenetic risk (p < 0.05). This study suggests that a gene expression-based risk score for adult B-ALL patients could be a feasible tool in resource-limited settings; however, it has limitations. Our B-ALL cohort lacks large-scale gene expression assessment data, preventing us from classifying included patients as Ph-like. Moreover, our cohort is from a single center, necessitating validation by other research groups in future studies. In summary, the differential expression of the evaluated genes warrants further investigation due to potential implications for disease biology and as therapeutic targets. The score based on Ph-like genes may be a valuable tool for risk stratification of adult B-ALL patients in resource-constrained settings. Supported by FAPESP, CAPES, CNPq.