Improved Outcomes for Pediatric B-ALL but Inferior Survival for T-ALL Using the Same Treatment Protocol: Results of Consecutive Malaysia-Singapore 2003 and 2010 Childhood ALL Trials

Background

The Malaysia-Singapore (MS) 2003 and 2010 trials are consecutive, international multi-center trials for the treatment of acute lymphoblastic leukemia (ALL) in South-East Asia, utilizing mainly a 3-drug induction and a BFM-based backbone, with identical therapy and risk stratification for both B-ALL and T-ALL. In the successor MS2010, protocol modifications were made, largely in order to further reduce treatment toxicities and shift care towards the outpatient setting to optimize resource utilization. Intensification was planned only in IKZF1^del patients, where patients were prospectively upgraded to the next risk arm. While these modifications in MS2010 proved immensely successful for B-cell ALL, survival outcomes were conversely distinctly inferior for T-ALL.

Methodology

We evaluated patient characteristics, treatment response and survival outcomes of N=1,117 children and adolescents treated on consecutive Ma-Spore trials. We compared the differences in early treatment response (as measured by minimal residual disease, MRD) and outcomes between B-ALL (N=986) and T-ALL (N=131) across these two successive protocols.

The main changes in MS2010 during induction and consolidation included:

1. Removal of day 8 prednisolone response as a risk stratification criteria.

2. Fractionated and reduced-intensity consolidation.

- The 4-week block of BFM-Ib consolidation was split into 2 shorter blocks of 3 weeks (Ib#1 and Ib#2). The original cyclophosphamide dose of 1g/m² was split into 0.5g/m²/dose over 2 days.

- Total cytarabine dose was reduced by 50%. The omitted cytarabine blocks were replaced by two weekly doses of vincristine (VCR).

- The Ib#1 and Ib#2 were spaced apart by 2 doses high dose methotrexate (HD-MTX) for SR and IR, or a fludarabine block (FLAD) for HR, in between.

For MS2010, other changes during delayed intensification included:

1. Removal of all anthracyclines from reinduction for standard risk patients.

2. Removal of anthracyclines from the 3^rd block of reinduction for intermediate risk patients.

3. Reducing cytarabine by 50% in each reinduction phase and the omitted block was replaced by two weekly doses of VCR.

Results

For B-ALL, both 5-yr event-free survival (EFS, 80.6% to 90.3%, P<0.0001) and 5-yr overall survival (OS, 89.8% to 95.4%, P=0.0003) improved from MS2003 to MS2010. Excluding patients with IKZF1^del who experienced treatment intensification, the remainder of B-ALL without IKZF1^del (N=874) i.e. who underwent the deintensification changes in MS2010, similarly had improved outcomes in MS2010 for both EFS (83.2% vs. 92.1%, P=0.0002) and OS (92.1% vs. 95.9% P=0.014).

However, from MS2003 to MS2010, T-ALL had a distinctly inferior outcomes: 5-year EFS (80.3% vs 51.9%, P=0.008) and OS (87.7% vs 67.7%, P=0.009). This inferiority was seen throughout all treatment arms. Notably, the standard risk arm fared the poorest in T-ALL (EFS 33.3%), whereas the SR arm had distinctly the best outcomes in B-ALL (EFS 96.6%, P<0.001 for T-ALL vs B-ALL in MS2010).

For B-ALL, in multivariable analysis, the MS2003 protocol had inferior EFS (Hazard Ratio [HR], 2.3; 95% confidence interval [CI], 1.5 – 3.4; P<0.0001), along with high-MRD at end-of-induction, IKZF1^del, BCR-ABL1 subtype and low hypodiploidy. For T-ALL, only the MS2003 had independent prognostic effect, and in contrast to B-ALL, MS2003 conferred a more favorable EFS (HR, 0.3; 95% CI, 0.1 – 0.8; P=0.016).

We assessed differences in treatment response at end-of-consolidation (EOC) from MS2003 to MS2010. For B-ALL, EOC-MRD improved significantly: proportion of patients with negative MRD, 68% to 85%; low-positive MRD, 21% to 9%; and high-positive MRD, 5% to 3% (P=2.43E-10). In contrast, MS2010 was significantly worse for T-ALL: EOC MRD-negative proportions decreased from 47% to 41%, low-positive MRD increased from 21% to 33%, and dramatic worsening of high-positive MRD proportions from 2% to 13% (P=0.038).

Conclusion

In our MS2010 trial with fractionated consolidation and reduced overall treatment intensity, results of B-ALL improved. However, these same changes were instead detrimental for T-ALL. Our results highlight that T-ALL should be treated differently from B-ALL. This may help inform in the design of future lineage-specific ALL treatment protocols.