Central nervous system involvement in pediatric Acute Myeloid Leukemia: A comprehensive report from the AML-BFM group

Abstract Background The involvement of the central nervous system (CNS) affects approximately 10% of pediatric patients with acute myeloid leukemia (pedAML). However, its prognostic impact remains unclear. Since 2011, prophylactic cranial irradiation has been abandoned in the AML-BFM group. Current treatment strategies aim to reduce exposure by limiting both cranial irradiation and intrathecal chemotherapy, with mono- or triple intrathecal therapy administered only as necessary. The NOPHO group has eliminated cranial irradiation without observing any negative impact on survival rates. Our aim was to evaluate whether this treatment change involving the discontinuation of prophylactic cranial irradiation over the past two decades has influenced patient outcomes. Methods We conducted a retrospective analysis of patients diagnosed with de novo AML between January 2012 and December 2022 in Germany. Data were derived from the AML-BFM 2012 and 2017 registries, and the samples were centrally analyzed. Patients with Acute Promyelocytic Leukemia, Myeloid Leukemia associated with Down Syndrome, Mixed Phenotype Acute Leukemia, and therapy-related AML were excluded. CNS assessment data were available for 811 of the 815 remaining pediatric patients, while eleven had an early death. Pearson chi-square or Fisher's exact test and Mann-Whitney U test were used for comparison of the two groups. Overall survival (OS) and event-free survival (EFS) were analyzed using Kaplan-Meier method and the log-rank test. Results The mean follow-up was 3.8 years (0.8-11.8). The 5y-OS and 5y-EFS of the total cohort were 78±1.6 % and 58.7±1.9 %, respectively. Sixty-nine patients (8.6 %) had CNS involvement at the time of diagnosis. The mean age at diagnosis was 8.3±6.1 years for CNS negative (CNSneg) patients and 7.1±6.1 years for CNS positive (CNSpos) patients (p=0.12). A similar distribution was noticed for sex (female: CNSneg 48.2 % vs. CNSpos40 %, p=0.28). Higher leukocyte count at diagnosis was significantly correlated with CNS involvement (CNSneg 49.2x109/L vs. CNSpos 92.6x109/L, p<0.001). Haemoglobin levels and platelet counts were comparable at diagnosis in both groups. FAB M5 was observed more often in CNSpos patients (22% vs. 33%,p=0.04). Survival analysis revealed no significant difference in 5-y-OS (CNSneg78.3±1.7% vs. CNSpos 75.5 ± 6.1%, p=0.64) or 5-y-EFS (58.9±2.0% vs. 56.4±6.3%, p=0.65). However, subgroup analysis based on risk stratification revealed a notable trend in the high-risk group, where CNSpos patients had lower survival rates (5y-OS: 66.8±4.0% vs. 49.4±12.6%, p=0.25; 5y-EFS: 48.2±4.1% vs. 32.2±11.2%, p=0.27). No relevant differences in survival were observed among standard- and intermediate-risk patients. CNSpos patients showed a significantly higher frequency of trisomy 8 (7.8% vs. 18.5%, p=0.017), inv(16) (8.8% vs. 18.5%,p=0.040), and t(11;19)(q23;p13.1) (2.3% vs. 9.3%,p=0.016) and a trend towards FLT3-ITD mutations (15 % vs. 24.1%,p=0.069).Within the CNSpos cohort, fifteen patients with hyperleukocytosis required pre-phase cytarabine prior to the initial lumbar puncture. The majority of patients with CNS involvement achieved clearance of cerebrospinal fluid (CSF) blast after three or fewer intrathecal treatments. Persistent CSF blasts were detected in only four patients after the fourth lumbar puncture and in two patients following the fifth puncture. MRI scans showed radiological abnormalities in 22 patients, including orbital involvement in four cases. Of these, five patients received cranial irradiation (18–22 Gy) due to persistent radiological abnormalities post-treatment. Overall, 15 patients underwent cranial irradiation due to persistence of CNS disease; among them, nine relapsed, and two of them died due to disease progression. Conclusion Our data confirm the previously reported frequency of CNS involvement in pedAML and do not show a general association with inferior survival outcomes. CNS involvement is frequently associated with leukocytosis and monocytic FAB subtype. No statistically significant differences in outcomes were observed for CNSpos patients for different risk subgroups. In patients with radiological evidence of CNS involvement, due to high risk of relapse, close monitoring during and after treatment is warranted; however, there is no indication that cranial irradiation provides a survival benefit.