Results in Pediatric T-ALL Patients Treated in Trial AIEOP-BFM ALL 2009: Exploring Prognostic Factors in the Context of Modern Risk Adapted Therapy

Introduction

The prognosis of children with T-cell acute lymphoblastic leukemia (T-ALL) has dramatically improved over time, but the outcome of relapsed and refractory patients remains dismal. Trial AIEOP-BFM ALL 2000 showed the benefit of dexamethasone (DXM) during induction in patients with prednisone good response (PGR) and that high level of bone marrow disease (≥10% blasts) by flowcytometry at induction day 15 (d15 FCM-MRD HR) was associated with poor prognosis. The subsequent AIEOP-BFM ALL 2009 trial included these interventions for T-ALL: (1) use of d15 FCM-MRD HR as new high risk (HR) criterion for stratification, (2) DXM for PGR patients in induction, (3) addition of cyclophosphamide (1 g/m²) on induction day 10 in patients with prednisone poor response (PPR), (4) substitution of native asparaginase with PEG-ASNASE, (5) a randomized study to receive or not 4 doses of PEG-ASNASE (2500 IU/m²) during Consolidation Phase IB in patients classified as HR by end of induction (EOI). Patients were stratified as HR if they had one of the following: PPR, d15 FCM-MRD HR, no EOI-CR or MRD ≥5×10^-4 at end of consolidation (EOC). Here we describe the outcome of T-ALL patients into this protocol and explore the prognostic factors for future refinement of risk group classification.

Methods

Between 06/2010 and 02/2017, this trial enrolled 6136 patients, aged 1-17, of whom 872 (14.2%) with T-ALL. The Cox regression model on event-free survival (EFS) evaluated the effect of factors in terms of Estimated Hazard Ratios (EHR) in different scenarios for patient stratification (significance was p<0.05).

Results

The 5-year EFS and overall survival for T-ALL patients were 79.9±1.4% and 84.9±1.2%, respectively, and the cumulative incidence of relapse (CIR) and death were 13.7±1.2% and 6.3±0.9%, respectively.

Non-HR patients (n=470) had an EFS of 86.8±1.6% and CIR of 8.7±1.3%, respectively, with excellent outcome in those MRD negative at EOI (n=137, EFS 91.1±2.5%). For HR patients (n=402), EFS and CIR were 71.9±2.3% and 18.0±1.9%, respectively, with a markedly inferior outcome in those assigned to HR group due, in hierarchical order, to EOC-MRD ≥5x10^-4 (n=79, EFS 58.7±5.6%) or no EOI-CR (n=26, EFS 57.4±9.7%) compared to those assigned to HR due to PPR (n=228, EFS 75.8±2.9%) or d15 FCM-MRD HR only (n=65, 83.0±4.7%). The randomized intensification with PEG-ASNASE did not improve outcome, as already reported (Rizzari C, Hemasphere 2023). Other interventions adopted in this study contributed to the improved outcome compared to trial AIEOP-BFM ALL 2000 (EFS 76.3%±2.0%, Schrappe M, Blood 2011).

In univariate analysis, WBC count ≥300 x10⁹/L (n=155, EFS 70.2±3.7%), central nervous system disease (CNS3, n=78, EFS 63.6±5.7%), PPR, d15 FCM-MRD HR or no EOI-CR were associated with worse prognosis. Poor outcome was also observed for patients with EOI-MRD ≥5x10^-2 (n=90, EFS 64.5±5.2%) compared to those with lower levels or negative (n=656, EFS 78.5±3.9% – 92.3±2.3%). Similarly, the outcome was worse for patients with EOC-MRD ≥5x10^-4 (n=79, EFS 58.7±5.6%) compared to those with lower levels or negative (n=667, EFS 79.6±2.5% - 88.1±1.6%).

A Cox model investigated if factors for optimizing stratification could be identified at EOI. In this model, CNS3 (EHR 2.36), PPR (EHR 1.77) and high level of EOI-MRD (EHR 4.75 for ≥5x10^-2 compared to negative) retained a significant impact on EFS. A 2^nd model included also EOC-MRD, with a significant impact of levels ≥5x10^-4 and <5x10^-3 (EHR 2.52) and of ≥5x10^-3 (EHR 6.28); here EOI-MRD lost significance, while CNS3 and PPR retained their prognostic value. A 3^rd model investigating the dynamic of MRD showed an increased hazard for patients maintaining a high level at EOC (≥5x10^-4) regardless of EOI level, and similar results for CNS3 and PPR. Age, WBC count, d15 FCM-MRD HR did not significantly affect outcome in these 3 models.

Conclusions

Results of T-ALL patients in AIEOP-BFM ALL 2009 were favorable, yet future patient stratification can be optimized. Although EOC-MRD remains the strongest prognostic factor, PPR remains a risk factor and CNS3 disease is identified as a novel HR factor for stratification. EOI-MRD alone has prognostic value, with increasing levels associated with worsening outcome, and identifies two patient subgroups, those with levels ≥5x10^-4 and <5x10^-2 with a poor prognosis and those with MRD ≥5x10^-2 with a dismal prognosis, that could be the targets for early experimental interventions.