Impact of the NGS-based and the PredicT-ALL classifiers in pediatric patients with T-ALL from the AALL0434 trial. (A) HR-NGS-based classifier patients in the AALL0434 trial had a 5-year DFS of 83% (95% CI, 79-86) compared with 92% (95% CI, 89-95) in the low-risk NGS-based classifier (LR-NGS) group (P < .001). (B) 5-year OS was 87% (95% CI, 84-90) in the HR-NGS group vs 93% (95% CI, 91-95) in the LR-NGS group (P < .001). (C) The 5-year DFS in patients from the PredicT-ALL_High group was 80% (95% CI, 76-85), 87% (95% CI, 84-90) in the NGS-PredicT-ALL_Int patients, and 95% (96% CI, 93-99) in the PredicT-ALL_Low. PredicT-ALL_High vs PredicT-ALL_Int, P = .03, PredicT-ALL_Low vs PredicT-ALL_Int, P = .01, and PredicT-ALL_Low vs PredicT-ALL_Int, P < .001. (D) 5-year OS was 84% (95% CI, 80-88) in the PREDICT-ALL_High vs 98% (95% CI, 96-99) in the PredicT-ALL_Low, and 90% (95% CI, 88-93) in the PredicT-ALL_Int (P < .001). PredicT-ALL_High vs PredicT-ALL_Int, P = .002; PredicT-ALL_Low vs PredicT-ALL_Int, P < .001; and PredicT-ALL_Low vs PredicT-ALL_High, P < .001.