RNA-seq Molecular Risk Classification Refines Prognostic Stratification in Childhood B-Cell Acute Lymphoblastic...

Clinical risk group and early minimal residual disease (MRD) guide therapy for childhood B-cell acute lymphoblastic leukemia (B-ALL) but may not capture all molecular heterogeneity. We evaluated whether RNA sequencing (RNA-seq) molecular risk classification improved outcome prediction under Chinese Children's Cancer Group (CCCG)-acute lymphoblastic leukemia (ALL)-2020.

This retrospective cohort included 531 children with newly diagnosed B-ALL treated at the Children's Hospital of Soochow University from 2020 to 2025. RNA-seq subtypes were grouped as favorable, intermediate/other, or adverse risk. The primary endpoint was standard event-free survival (EFS). Real-world EFS additionally counted treatment abandonment and non-protocol discontinuation. Cox models and Harrell's C-index evaluated prognostic performance.

Favorable, intermediate/other, and adverse molecular risk groups included 285, 152, and 94 patients, respectively. The 4-year standard EFS and overall survival were 85.2% and 95.3%. In the integrated Cox model, compared with favorable molecular risk, intermediate/other risk (hazard ratio [HR], 3.23; 95% confidence interval [CI], 1.58-6.60; p = 0.001) and adverse risk (HR, 4.22; 95% CI, 1.94-9.17; p < 0.001) were independently associated with inferior EFS. Adding Day-19 MRD and molecular risk group increased Harrell's C-index from 0.698 to 0.762. Under the real-world EFS definition, adverse molecular risk remained associated with inferior outcomes.

RNA-seq molecular risk classification added prognostic information beyond protocol-assigned clinical risk and early MRD in CCCG-ALL-2020-treated B-ALL. These findings support prospective validation of transcriptomic risk assessment but do not establish RNA-seq as a stand-alone treatment-allocation tool.