Paper Details

ABSTRACT

Approximately 5–15% of children and adolescents with acute lymphoblastic leukemia (ALL) develop acute neurotoxicity during therapy, including seizures, posterior reversible encephalopathy syndrome (PRES), and cerebrovascular events. These events cause short- and long-term morbidity and may increase relapse risk through treatment interruptions or modifications. Prior studies have largely been small or single-institution, and most have been restricted to a single toxicity (usually seizures) or a single agent (usually methotrexate) rather than capturing neurotoxicity across an entire treatment protocol. To describe the frequency and types of neurotoxic events and identify predictors of acute neurotoxicity, we analyzed prospectively collected data from three consecutive DFCI ALL Consortium protocols for pediatric patients (pts) with newly diagnosed ALL.

Methods: Between 2006-2022, 1481 evaluable pts aged 1 to <22 years with newly diagnosed B- or T-ALL were enrolled on DFCI protocols 05-001 (n=731), 11-001 (n=222), and 16-001 (n=528). Each protocol included 1-2 doses of high-dose methotrexate (5 g/m²) and a similar risk-adapted IT chemotherapy schedule. Cranial radiation was given to 23.8% of pts on 05-001 (12 Gy: 154; 18 Gy: 20), 18.5% on 11-001 (12 Gy: 38; 18 Gy: 3), and 1.9% on 16-001 (18 Gy: 10). Neurotoxicity was defined as any seizure regardless of grade, or any other Grade 3+ CNS toxicity per CTCAE v3.0 (05-001) or v4.0 (11-001, 16-001). Candidate predictors were age, sex, race/ethnicity, initial WBC, immunophenotype, final risk group, CNS status, Down syndrome, severe infection (Grade 3+), and cranial radiation. Proportions between groups were compared using chi-square or Fisher's exact tests. Logistic regression models evaluated associations between covariates and neurotoxicity. Variables significant at p≤0.10 in univariate analysis entered an initial multivariable model, with backward selection for final retention (p≤0.05). Event-free survival (EFS) was estimated by the Kaplan-Meier method.

Results: 93 pts (6.3%) developed acute neurotoxicity, with similar incidence across protocols (05-001: 7.1%; 11-001: 5.4%; 16-001: 5.5%). Seizures were most common (57.0% of events), followed by CNS hemorrhage (11.8%), encephalopathy (7.5%), PRES (4.3%), and ischemic stroke (2.2%). There were 4 fatal events (4.3%), all occurring during induction (CNS hemorrhage:3, encephalopathy:1). Of the 93 pts, 31% experienced a neurotoxic event during induction, 14% within 3-weeks following high-dose methotrexate, and 12% during maintenance, which included IT chemotherapy every 9-18 weeks. In univariate analyses, presenting features associated with neurotoxicity were older age (p=0.001), Hispanic ethnicity (p=0.016), T-ALL (p=0.006), Down syndrome (p=0.002), and CNS status at diagnosis (CNS-2, p=0.035; CNS-3, p=0.032). Standard Risk (SR) pts were less likely than High or Very High Risk pts to develop post-induction neurotoxicity (p=0.0015). Prior Grade 3+ infection and prior cranial radiation were each associated with an increased risk of subsequent neurotoxicity (p=0.0009 and p=0.002, respectively). In multivariable analysis, independent predictors of neurotoxicity were Down syndrome (OR 4.29, 95% CI 1.76–10.29, p=0.001), Hispanic ethnicity (OR 2.04, 95% CI 1.21–3.44, p=0.007), CNS-2 (OR 1.88, 95% CI 1.13–3.15, p=0.016), CNS-3 (OR 2.40, 95% CI 1.08–5.33, p=0.031), and age ≥10 yrs (OR 1.97, 95% CI 1.28–3.04, p=0.002). 5-year EFS was 81.1% (95% CI 71.3–87.8) in pts with neurotoxicity vs 86.9% (95% CI 85.0–88.7) in those without (p=0.19).

Conclusions: To our knowledge, this is one of the largest multicenter studies to identify predictors of acute neurotoxicity during pediatric ALL therapy. Down syndrome, Hispanic ethnicity, CNS-2 or CNS-3 status, and older age at diagnosis were independently associated with acute neurotoxicity. SR pts were less likely to develop post-induction neurotoxicity. Prior Grade 3+ infection and cranial radiation were each associated with subsequent neurotoxicity. There was a trend toward inferior EFS in pts with neurotoxicity, possibly due to its association with HR presenting features. Ongoing analyses are examining long-term neurological and neurocognitive sequelae after acute neurotoxicity and whether CSF biomarkers can identify pts at highest risk. Future studies should also evaluate changes in frequency and predictors of neurotoxicity as blinatumomab is incorporated into frontline therapy for ALL.