Browse by author
Lookup NU author(s): Ozcan Gulbey, Dr Amir EnshaeiORCiD, Melvin JoyORCiD, Dr Sujith Samarasinghe, Professor Anthony MoormanORCiD
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License (CC BY-NC-ND).
Background: T-cell acute lymphoblastic leukaemia (T-ALL) is an aggressive disease with high relapse rates and poor survival post relapse. There are few established risk factors and only end of induction (EOI) Measurable Residual Disease (MRD) is widely used to risk stratify patients. Therefore, identifying additional risk factors is crucial to improving risk stratification and treatment in T-ALL. Aims: 1. Evaluate a wide range of potential prognostic factors (demographic, clinical, genetics and response) in 3 independent trials. 2. Develop and validate a prognostic model to stratify patients to clinically relevant risk groups. Methods: We used data from 916 T-ALL patients aged 1-24 years old recruited to consecutive trials (UKALL97/99, -2003 and -2011) between 1997 and 2018 with a median follow-up time ranging 5-11 years. We evaluated a wide range of potential risk factors including sex, age, white blood cell count, CNS status, somatic genetics and treatment response measured at various timepoints by morphology (blast percentage) and IgMRD. Wherever possible, we used continuous rather categorical data, explored the utility of transformed variables and imputed missing data. Survival analysis employed standard endpoints – relapse rate (RR) and overall survival (OS) – and all rates are quoted at 5 years. Modelling was based on Cox regression and employed backward and forward selection with C-index used to assess model fitness. Results: Overall survival was: UKALL97/99 (n=209) 73% (95% CI 66-78), UKALL2003 (n=386) 86% (82-89) and UKALL2011 (n=321) 84% (80-88). Univariable analysis failed to identify any significant and reproducible prognostic effect for sex, age, white blood cell count, CNS involvement and genetics across the trials. Multivariable modelling of the UKALL97/99 cohort revealed that only treatment response was predictive of outcome. Although day 8 and 28 marrow blast percentage were associated with an increased risk of relapse (D8BM% hazard ratio=1.31 (95% CI 1.14-1.49), p<0.001 and D28BM% 1.29 (1.05-1.58), p=0.016), only D8BM% remained in final model. We used the coefficient of D8BM% model to develop a prognostic index (range 0-2.7), and threshold analysis revealed a cut-off of 1.7 defined standard and high-risk groups accounting for ~80% and 20% cases. Patients in the high-risk group had a significantly inferior RR and OS (47% v 17% and 50% v 80%, both p<0.001). Crucially, the D8BM% model validated in both UKALL2003 [HR/SR 21%/79%, RR 22%/11% p=0.06 and OS 76%/89% p=0.01] and UKALL2011 [HR/SR 24%/76%, RR 30%/15% p=0.007, OS 69%/90% p=0.0005]. In UKALL2003, EOI MRD<0.01% was used to randomise patients to post-induction therapy whereas in UKALL2011 EOI MRD<0.005% was used to assign post-induction therapy. Hence, we evaluated our D8BM% model within the context of EOI MRD. In UKALL2003, the D8BM% model identified a large group of non-refractory (MRD<5%) patients with MRD≥0.01% who had a rapid early response (D8BM%- SR) and had significantly lower RR (Figure 1A & 1B) and better OS compared to those who had a slower early response (D8BM%-HR). Similarly, in UKALL2011, MRD risk patients (MRD≥0.005%) could be sub-divided into prognostic subgroups based on the D8BM% model. Summary/Conclusion: Risk stratification to T-ALL is challenging because the risk factors and thresholds identified in BCP-ALL have limited value. Our study provides strong evidence that early treatment response is a critical predictor of outcome, even when crudely measured by morphology. Modelling outcome at very early time-points provides an opportunity to identify high-risk patients and adapt therapy before the EOI.
Author(s): Gulbey O, Delft FV, Enshaei A, Joy M, Kirkwood AA, Moppett J, Samarasinghe S, Vora A, Moorman AV
Publication type: Conference Proceedings (inc. Abstract)
Publication status: Published
Conference Name: EHA2024 Hybrid Congress
Year of Conference: 2024
Pages: 36-37
Online publication date: 14/06/2024
Acceptance date: 23/04/2024
Date deposited: 15/08/2025
ISSN: 2572-9241
Publisher: Wiley
URL: https://doi.org/10.1002/hem3.104
DOI: 10.1002/hem3.104
Notes: Abstract number S116
Series Title: HemaSphere
