AUTHORS
Hironori Goto, MD, Kentaro Sakamaki, PhD2, Jun Matsubayashi3, Takahiro Aoki, MD, PhD4, Koji Kawaguchi5, Hiroshi Takada6, Shota Kato, MD, PhD7, Shotaro Iwamoto, MD,PhD8, Hidefumi Hiramatsu, MD, PhD9, Kaori Niwa10, Reo Tanoshima11, Yoko Mizoguchi, MD, PhD12, Yasushi Kasahara13, Akifumi Endo14, Akiko M Saito15, Kentaro Ohki16, Hayato Miyachi17, Shin-Ichi Tsujimoto, MD,PhD18, Takao Deguchi, MD, PhD19, Akio Tawa20, Hiroshi Moritake, MD,PhD21, Souichi Adachi, MD,PhD22, Shiro Tanaka23, Daisuke Tomizawa, MD, PhD24
AFFILIATIONS
1 Oita University, Oita, Japan
2 Juntendo University, Tokyo, Japan
3 Shiga University of Medical Science, Otsu, Japan
4 Chiba University Hospital, Chiba, Japan
5 Shizuoka Children's Hospital, Shizuoka, Japan
6 Keio University School of Medicine, Tokyo, Japan
7 The University of Tokyo, Tokyo, Japan
8 Mie University Graduate School of Medicine, Tsu, Japan
9 Kindai University, Sakai, Japan
10 Mie University Graduate School of Medicine, Tsu, Japan
11 Yokohama City University, Yokohama, Japan
12 Hiroshima University, Hiroshima, Japan
13 Niigata Cancer Center Hospital, Niigata, Japan
14 Institute of Science Tokyo, Tokyo, Japan
15 NHO Nagoya Medical Center, Ngoya, Japan
16 National Research Institute for Child Health and Development, Tokyo, Japan
17 Nitobe Bunka College, Tokyo, Japan
18 Yokohama City University Hospital, Yokohama, Japan
19 National Center for Child Health and Development, Tokyo, Japan
20 Higashiosaka Aramoto Heiwa Clinic, Osaka, Japan
21 University of Miyazaki, Miyazaki, Japan
22 Kyoto University, Kyoto, Japan
23 Kyushu University Hospital, Fukuoka, Japan
24 National Center for Child Health and Development, Tokyo, Japan
ABSTRACT
Introduction: Central nervous system (CNS) involvement is observed in 10–30% of children with newly diagnosed acute myeloid leukemia (AML). However, unlike in acute lymphoblastic leukemia, many aspects remain poorly understood, including the optimal CNS-directed therapy and the clinical significance of CNS2 status. Therefore, we evaluated the clinical characteristics and prognostic impact of CNS involvement in children with AML treated in two consecutive Japanese clinical trials.
Methods: We retrospectively analyzed patients enrolled in the AML-05 (2006–2010) and AML-12 (2014–2018) trials. The protocol for this secondary analysis was approved by the Institutional Review Board of Oita University. CNS3 was defined as a cerebrospinal fluid (CSF) white blood cell (WBC) count ≥5/µL with blasts detected by CSF cytology or the presence of clinical symptoms or radiographic findings consistent with CNS leukemia. CNS2 was defined as the presence of blasts in the CSF with a WBC count <5/µL, whereas CNS1 was defined as no blasts with a CSF WBC count <5/µL. In both trials, patients received two induction courses consisting of cytarabine, anthracyclines, and etoposide. Patients who achieved morphological complete remission (CR) subsequently received consolidation therapy based primarily on high-dose cytarabine (HDAC), whereas those classified as high-risk (HR) according to cytogenetic abnormalities and morphological treatment response underwent hematopoietic cell transplantation. All patients received triple intrathecal therapy (TIT; cytarabine, methotrexate, and hydrocortisone) during each course of treatment. Patients with CNS3 received additional TIT during the initial induction course. Baseline characteristics, response to induction therapy, and survival outcomes were compared according to the CNS status.
Results: Among the 802 patients, 675 were classified as CNS1, 76 as CNS2, and 47 as CNS3, and CNS status was unknown in 4 patients. The median age was 7.4 years (range, 0–18 years). Patients with CNS3 were significantly younger than those with CNS1/2, whereas the sex distribution did not differ among the three groups (370 females and 432 males). Patients with CNS2/3 had significantly higher WBC counts at the diagnosis than those with CNS1. FAB M4 and M5 morphologies were more frequent in the CNS3 group [n = 12 (25.5%) and n = 15 (31.9%), respectively], whereas FLT3-ITD positivity was more frequently detected in the CNS2 group [n = 16 (21.1%)]. Cytogenetic and protocol-defined risk classifications did not differ according to CNS status. Traumatic lumbar puncture (TLP+) was significantly more frequent in CNS2/3 than in CNS1, reflecting the classification criteria. A multivariate logistic regression analysis revealed that an initial WBC count ≥10×109/L and FAB M4/M5 morphology were independently associated with CNS2 or CNS3. The CR rates did not differ among the CNS groups. Likewise, the 3-year EFS rates were 58.7%, 59.5%, and 55.1% in the CNS1, CNS2, and CNS3 groups, respectively (p = 0.85), while the corresponding 3-year OS rates were 76.7%, 84.5%, and 71.2%, respectively (p = 0.14). However, among patients in the HR cohort, those with CNS3 had significantly inferior EFS and OS compared with those with CNS2 (p < 0.001 and p = 0.005, respectively). In contrast, TLP+ status was not associated with CR, EFS, or OS.
Discussion: Our findings indicate that a high leukemic burden at diagnosis and FAB M4/M5 morphology are significant risk factors for CNS involvement, consistent with previous reports. Previous pediatric AML studies using treatment strategies involving intensified intrathecal cytarabine for both CNS2 and CNS3 diseases have demonstrated inferior EFS among patients with CNS involvement, particularly among those in the low- and standard-risk cytogenetic groups. In contrast, in the AML-05 and AML-12 trials, which employed multiple HDAC-based consolidations that provided effective CNS penetration and intensified TIT reserved for patients with CNS3, neither CNS2 nor CNS3 was associated with inferior EFS or OS in the overall cohort. These findings suggest that this approach may mitigate the adverse prognostic effects of CNS involvement. The inferior outcomes observed among HR patients with CNS3 should be interpreted cautiously, and further analyses incorporating the underlying cytogenetic or molecular features are warranted. Conversely, routine intensification of CNS-directed therapy for CNS2 diseases may not be necessary.
DISCLOSURES
Hironori Goto, MD: Takeda Pharmaceutical Company Limited - Honoraria