Hierarchical Evaluation of Baseline Poor-Prognosis Factors and Response to Initial Therapy in Younger Adults with Ph-Negative Acute Lymphoblastic Leukemia (ALL) Treated in the Risk-Adapted GRAALL-2003 Study.

Several prognostic predictors, including baseline ALL features and response to initial therapy, have been described in adult ALL raising the issue of whether these predictors might be redundant and which must be considered for treatment stratification. In the GRAALL-03 prospective Phase 2 study, we aimed to hierarchize the high-risk factors in Ph-negative ALL patients (pts): 1) baseline (BL): WBC≥ 30G/L in B-lineage, CNS involvement, MLL-AF4 and E2A-PBX fusions, haploidy/near-triploidy; 2) early response (ER): corticoresistance after prophase (CsR), chemoresistance at Day 8 (ChR); all CsR and/or ChR pts were planned to receive higher doses of cyclophosphamide (HyperC) at Day 15 of induction; 3) induction response (IR): no CR or Ig-TCR minimal residual disease (MRD) ≥ 10⁻² after standard or HyperC induction. Allogeneic stem cell transplantation was proposed to pts with a donor and at least one BL, ER, or IR factor. A total of 144 pts are included in this analysis (median age, 31y, 15–59; T/B-lineage, 48/96; CNS+, 8; MLL-AF4, 12; E2A-PBX, 5; median follow-up, 12 months). After induction, CR rate was 127/144 (88%; 98/83% in T/B-lineage, P=.01; 95/87% in standard/HyperC induction, P=.15), 3 others pts reaching CR after additional salvage (overall CR rate, 91%). There were 11 early deaths (8%), associated with age > 40y but not with HyperC reinforcement. Among the 130 CR pts, 13 relapsed and 6 died in CR. Overall, estimated 18-month DFS and OS were 65% and 74% (OS 93/66% in T/B lineage, P=.04), respectively. BL criteria: the 41/144 pts with BL factors (28%) tended to be older and had worse DFS (P=.04) despite similar CR rate. ER criteria: 74/143 tested pts (52%) had poor ER (CsR/ChR, 23; CsR/ChS, 15; CsS/ChR, 36; P=.007). T-ALLs were more frequently CsR (42 vs 19%; P=.005) but not ChR. Poor ER did not correlate with age or BL factors (of note, 9/12 MLL-AF4 and 4/5 E2A-PBX ALL had good ER), but was strongly associated with MRD+ in CR patients tested (P=.03). Overall, 97/144 pts (67%) had at least one BL or ER high-risk factor. In the setting of our ER-adapted HyperC induction strategy, poor ER did not predict lower CR rate (93 vs 90%) or lower DFS (P=.59) even after censoring allografted pts at transplant time. IR criteria: 10/70 CR pts tested (14%) were MRD+, 8 were poor early responders and the 2 remaining had CD10-negative B-lineage ALL and TLX3/NUP214-ABL T-lineage ALL, both potential high-risk BL features. In conclusion: 1) ER to initial therapy is not correlated to BL high-risk factors and should be considered in further adult ALL studies; 2) morphological ER evaluation of CsR/ChR correlates closely with post-induction MRD measurement; 3) HyperC reinforcement in poor ER pts appears to be a promising approach to improve their outcome and will be randomly tested in the next GRAALL-05 study.