Identification and validation of a prognostic risk-scoring model in drug-resistant ALL and evaluation of immune micro-environment

ALL is the commonest childhood hematological malignancy; modern multi-agent chemotherapy cures about 80% of patients, yet roughly 20% experience primary or secondary chemotherapy resistance that sharply reduces long-term survival. The molecular circuits driving this resistance remain poorly defined, and validated biomarkers for early risk stratification are lacking. Growing data implicate reprogramming of the immune microenvironment as a critical determinant of treatment response and relapse risk. We therefore interrogated transcriptomic profiles from GEO and TARGET datasets to identify genes differentially expressed between chemosensitive and chemoresistant ALL, distilled a 13-gene resistance-associated prognostic index (RAPI-13) by LASSO regression, and validated the signature in independent cohorts. Functional enrichment, immune-infiltration deconvolution and single-cell RNA-seq revealed that high-risk patients harbor activated oncogenic signaling (IL-2/STAT5, Notch, PI3K/Akt/mTOR) together with immune-evasion signatures. Mechanistic studies showed that CD40 agonist treatment polarizes tumor-associated macrophages toward the pro-inflammatory M1 state, restores anti-leukemic immunity and re-sensitizes resistant blasts to chemotherapy. In cell-line-derived xenografts and humanized mice, triple therapy combining CD40 agonist, daunorubicin and PD-L1 blockade achieved superior leukemia control, prolonged survival and fostered immune memory formation. These findings provide a clinically applicable prognostic tool and pre-clinical rationale for integrating CD40-centered immunotherapy into frontline protocols to overcome chemotherapy resistance in childhood ALL.