Prognostic Factors for T-Cell Acute Lymphoblastic Leukemia in Children and Young Adults: ALL-T11 Study Conducted By Japan Children’s Cancer Group and Japan Adult Leukemia Study Group

Background: T-cell acute lymphoblastic leukemia (T-ALL) has a cellular biology distinct from that of B-cell precursor ALL; therefore, a different treatment approach is required. The Japan Children’s Cancer Group and Japan Adult Leukemia Study Group conducted a phase 2, nationwide multicenter trial ALL-T11 for T-ALL in children, adolescents, and young adults (aged <25 years at diagnosis) and reported excellent outcomes (Sato, 2023). Currently, we investigated the prognostic impact of the initial patient characteristics and early treatment responses within the trial.

Methods: AIEOP-BFM 2000-backbone chemotherapy was used. We measured end-of-induction (timepoint [TP]1) and end-of-consolidation (TP2) minimal residual disease (MRD) using PCR targeting TCR/immunoglobulin rearrangement, and only TP2 PCR-MRD was used for stratification. Until the measurement of TP2 PCR-MRD, all patients were treated with an identical therapy that was intensified with dexamethasone instead of prednisolone in remission induction therapy IA and L-asparaginase in consolidation therapy IB. Patients were stratified into three risk categories; standard, high, and very high risk, according to prednisolone response, TP1 remission status, central nervous system (CNS) status, and TP2 PCR-MRD. Nelarabine (650 mg/m2 daily for 5 days) was administered in patients with high (6 courses) and very high risk (1 or 2 courses). Hematopoietic cell transplantation was scheduled for patients with very high risk. CNS leukemia (CNS-3) status was classified into CNS-3a (>5 white blood cells per microliter with blasts; without traumatic tap), -3b (>5 white blood cells per microliter with blasts; with traumatic tap), and -3c (with signs and symptoms of CNS leukemia infiltration and detection of CNS disease by computed tomography or magnetic resonance imaging). As prophylaxis for CNS leukemia, patients received protracted intrathecal therapy while sparing cranial radiotherapy (CRT) only for those with overt CNS-3. Patients with CNS-3 were treated with a high- or very high-risk protocol and received nelarabine. We compared event-free survival (EFS) and overall survival (OS) based on initial characteristics and treatment response indices using a Log-rank test. The Cox proportional hazard model was used for (EFS) and OS, and the Fine–Gray model was used to calculate the cumulative incidence of relapse (CIR). These analyses used time from TP2, and the model included factors selected based on clinical importance.

Results: Between 2011 and 2017, 349 patients (median age; 9 years) were eligible. Among them, 16 were excluded from the per-protocol set (PPS) because of violations, self-discontinuation, or treatment errors. Presently, we used the PPS (n=333). The 5-year EFS (5y-EFS), OS (5y-OS), and CIR for the PPS cohort were 84.8% (95% confidence interval [CI]; 80.3–88.3％), 89.8% (85.9–92.6%), and 10.1% (6.9–14.0%), respectively. Among the initial characteristics, such as sex, age, white blood cell counts, NCI risk, CNS status, mediastinal leukemia infiltration, and testicular involvement, only CNS-3 status (n=27) indicated a worse outcome compared with CNS-1/2, with 5y-EFS of 70.4% ([49.4–83.9%] vs 86.3% [81.7–89.8%], P=0.004) and 5y-OS of 70.4% ([49.4–83.9%] vs 91.8% [87.9–94.4%], P<0.001). We further investigated whether CNS-3a/3b or -3c status affect outcomes and found that the EFS, OS, and CIR for CNS-3c (n=15) were inferior to those for CNS-1/2 (n=304) and CNS-3a/3b (n=12). EFS and OS had no difference with the indicators of early treatment responses, including PSL response, day 15 bone marrow blast, TP1 bone marrow blast, and TP1 PCR-MRD, while high TP2 PCR-MRD load (≥10*-3) indicated poorer 5y-EFS (54.3% [28.8–74.1%] vs 89.0% [84.8–92.1%], P<0.001) and 5y-OS (72.2% [45.6–87.4%] vs 93.4% [89.8–95.7%], P<0.001). In the Cox proportional hazard model, we found that CNS-3c and TP2 PCR-MRD had significant impacts for poorer EFS (HR=3.37, 95% CI=1.30–8.70, p=0.012, HR=4.59, 95% CI=2.10–10.06, p<0.001) and OS (HR=5.48, 95% CI=1.83–16.41, p=0.002, HR=4.61, 95％ CI＝1.69–12.58, p＝0.003) and only CNS-3c was significant for higher CIR (HR=3.97, 95% CI=1.30–12.08, p=0.015).

Conclusion: This study found a prognostic impact of CNS-3c on EFS, OS, and CIR, even with the use of CRT and regimens with nelarabine, whereas patients with CNS-3a/3b had a comparable prognosis to those with CNS-1/2.