Tokyo Children's Cancer Study Group Research Articles

The Epigenetic and Genetic Landscapes of Pediatric T-Cell Acute Lymphoblastic Leukemia (T-ALL)

IntroductionAlthough genetic basis of pediatric T-cell acute lymphoblastic leukemia (T-ALL) has been well characterized and we reported recurrent SPI1 fusion cases exhibiting uniformly dismal outcome, no comprehensive studies have yet explored the epigenetic profiles and their potential contribution to clinicopathological features of T-ALL.MethodsA total of 79 cases with pediatric T-ALL samples from Tokyo Children's Cancer Study Group (TCCSG) and Japan Association of childhood Leukemia Study (JACLS) including 7 cases of SPI1 fusions were analyzed in this study. To describe epigenetic landscape of T-ALL, we performed methylation array analysis using Illumina HumanMethylation EPIC array, and conducted combined analysis of methylation data with our previous data of whole transcriptome sequencing (WTS) and targeted-capture sequencing for 158 ALL-related genes.ResultsAfter normalization, 939 probes were selected to identify the most variable methylated probes, when a standard deviation of the beta-values across the samples was above 0.38. Unsupervised consensus clustering clearly identified 4 distinct sample clusters. Combined analyses with WTS data revealed that these 4 clusters were characterized by TAL1 fusions/high expression with intermediate methylation (M1 cluster; n=39), high TLX/HOX expression with intermediate methylation (M2 cluster; n=20), high TLX//HOX expression with high methylation (M3 cluster; n=11), and SPI1 fusions/high expression with low methylation (M4 cluster; n=9), respectively. Interestingly, these methylation clusters were well correlated with genetic features, expression status, and clinical outcomes. In M1 cluster, PTEN abnormalities and UPS7 mutations were especially enriched, and these aberrations were observed mutually exclusive. Furthermore, promoter regions in TAL1 and its target ALDH1A2, important for retinoic acid synthesis, were significantly hypomethylated and highly expressed in this cluster. In M2 cluster, genes related to chromatin remodeling and JAK-STAT were frequently mutated. PRR5L and DEPTOR, both related to AKT1 activation, were differentially methylated and expressed, suggesting an important role of PI3K-AKT1 pathway. M3 cluster was characterized by frequent mutations of epigenetic regulators related genes, such as PHF6, EZH2, and DNM2 . In contrast, mutations in JAK-STAT and PI3K-AKT1 pathways were less frequent. Since the methylation profile of M3 cluster was similar to normal T cells, abnormal DNA methylation status may be not crucial for the leukemogenesis in this group. Intriguingly, all SPI1 fusion cases were classified into M4 cluster, and remaining 2 cases without SPI1 fusions in this cluster showed very high SPI1 expression. Importantly, M4 cluster was characterized by high expression of RASGRP4, a family of Ras guanine nucleotide exchange factors, and low methylation of RASGRP4 promoter. In addition, FGFR1 expression was uniformly repressed and abnormalities of CDKN2A (89%) and KRAS/NRAS (44%) were frequently observed in this group. Furthermore, it should be noted that patients in M4 cluster showed significantly poor prognosis compared to other clusters; eight out of nine cases were died (Log-rank p =4.4×10-7). SPI1 fusion cases exhibited not only unique cytological and gene expression profiles but also distinctive methylation status.ConclusionsBased on DNA methylation profiles, pediatric T-ALL is clustered into 4 distinct subtypes, which exhibited remarkable correlation with genetic signatures, expression features, and clinical outcomes. PI3K-AKT pathway may be essential in M1/M2 cluster, whereas genetic abnormalities of epigenetic regulators were characteristics in M2/M3 clusters. In addition, the involvement of JAK-STAT pathway would be unique in M2 cluster. Given that cases with M4 cluster showed a catastrophic prognosis and very high expression of RASGRP4 with high frequency of RAS mutation, our findings implicate that targeting therapy of RAS may improve the prognosis of the patients in this group. Although our cohort is limited, our results suggested that the biological phenotype of T-ALL is mediated by both genetic and epigenetic regulations, and explorations for aberrant DNA methylation along with genetic alterations might be helpful for a new therapeutic strategy for T-ALL. DisclosuresNo relevant conflicts of interest to declare.

NUDT15 Variants Confer High Incidence of Secondary Malignancies of ALL in Children

Background: Secondary or subsequent malignant neoplasms (SMNs) are one of the most serious late complications in children and adolescents after treatment for acute lymphoblastic leukemia (ALL). Several studies have reported that excess dosing of 6-mercaptopurine (6-MP) therapy and polymorphisms of 6-MP metabolizing enzyme coding gene might influence the risk of SMNs. Recently, nucleoside diphosphate-linked moiety X-type motif 15 (NUDT15) variants have been identified as a cause of high sensitivity to 6-MP, which reflects a high sensitivity to 6MP in East Asians. Therefore, we investigated an association between NUDT15 genotype and incidence of SMNs in pediatric ALL. Methods: Germline samples including bone marrow samples, peripheral blood samples and buccal swab samples during remission were obtained from 13 cases who suffered from SMN after childhood ALL. The prevalence of major NUDT15 non-functional alleles (*2 and *3), which is categorized intermediate or poor metabolizer of 6-MP in Clinical Pharmacogenetics Implementation Consortium Guideline (Relling M et al. Clin Pharmacol Ther 2019), was evaluated by using Sanger sequencing and/or whole exome sequencing. The NUDT15 variants prevalence in SMN patients with primary ALL was compared to those without SMN, who were registered in the TCCSG L04-16 study, which was conducted by the Tokyo Children's Cancer Study Group (TCCSG). Results: Patients' profiles and NUDT15 genotypes were summarized in table 1. Median ages at diagnoses of ALL and SMNs were 4 (1-15) and 13 (5-44) years, respectively. Median duration between ALL and SMNs was 8 (1-30) years. SMNs included acute myeloblastic leukemia (AML) / myelodysplastic syndrome (MDS) (n = 8), brain tumor (n = 1), basal cell carcinoma (n = 1), bladder cancer (n = 1), Ewing sarcoma (n = 1) and osteosarcoma (n = 1). Among the 13 SMN cases, six patients (46.2%) possessed NUDT15 non-functional alleles. The SMNs of these cases included AML / MDS (n = 4), bladder cancer (n = 1) and osteosarcoma (n = 1). Among them, 3 patients, whose SMNs were AML, bladder cancer and osteosarcoma, were treated with radiation therapy against primary ALL. All NUDT15 genotypes were heterozygous variants (3 were *1/*2 and 3 were *1/*3) and seemed to be intermediate metabolizers. In patients without known SMNs in TCCSG ALL L04-16 cohort study, 73 of 437 (16.7%) possessed NUDT15 non-functional alleles. Three of them had bi-allelic variants (2 cases were *3/*3, 1 case was *2/*3), and the others were all heterozygous. The prevalence of NUDT15 non-functional alleles was significantly higher in SMNs compared to that of non-SMN cases (P = 0.012). Conclusion: Our study indicated the NUDT15 hypomorphic variant alleles could be a risk of SMNs after ALL treatment with 6-MP. NUDT15 heterozygotic variants carriers are generally tolerable of 6-MP, and are usually treated with normal or slightly low-dose of 6-MP. However, our results suggested that standard adjustment of 6-MP in maintenance therapy might cause potential overdose of 6-MP for cases with NUDT15 variants, leading to increased risk of SMNs. To certify it, a higher number of study samples and somatic samples analysis of 6-MP induced SMNs are required. To obtain the best long-term event-free survival probability, it might be necessary to reduce the risk of SMNs development without compromising therapeutic effect by further optimal adjustment of 6-MP dose (or alternative agents). Disclosures No relevant conflicts of interest to declare.

Association of Germline Variants of TCF3 and PAX5 with Pediatric Acute Lymphoblastic Leukemia Development

Background: The development of B-lymphocyte is controlled by the coordinated action of transcription factors networks, such as IKZF1, TCF3, EBF1 or PAX5. It is increasingly recognized that germline variants of those genes which disrupt their function are responsible for inborn errors of humoral immunity. Meanwhile, it is well understood genetic alterations of those molecules are frequently altered in B-lineage acute lymphoblastic leukemia (ALL) and are associated with leukemogenesis. However, those involvements are usually attributed to somatic mutation. Germline variations may also trigger cancer development in some cases. For instance, genome-wide association studies (GWAS) have identified B-cell precursor (BCP-) ALL susceptibility variants within IKZF1. Furthermore, germline variants of IKZF1, TCF3, and PAX5 are reported to be associated with B-ALL development. Therefore, we hypothesize that attenuated function due to germline variation in those genes may predispose to ALL development in Japanese population to some extent. Method: Saliva DNA from 583 pediatric ALL patients in remission registered in Tokyo Children's Cancer Study Group (TCCSG) were analyzed in this study. Targeted sequencing for exons and exon-intron boundaries of IKZF1, TCF3, and PAX5 was performed using next-generation sequencer. Frequency of the detected variant alleles was compared to data from the Exome Aggregation Consortium (ExAC) East Asian Cohort (n=4327), Tohoku Medical Megabank Organization (ToMMo) cohort (n=3554), and the Human Genetic Variation Database (HGVD) cohort (n=1208) by chi-square test or Fisher's exact test. Results: In total, 113 variants (28 nonsynonymous, 19 synonymous, 56 intronic, 8 untranslated regions, and 2 deletions) were detected. We could not detect any nonsynonymous variations, insertion or deletion in IKZF1. On the other hand, we detected 8 rare variations in TCF3 and 7 of them are found to be deleterious by in silico analysis. Among them, 1 variant, TCF3 p.S514L, was significantly more common in ALL patients (0.26% in patients vs 0% in ExAC, 0.01% in ToMMO and 0% in HGVD, p=0.0017, 0.0115 and 0.0343 respectively). In PAX5, we found 6 rare variations and 3 of them are found to be deleterious by in silico analysis, However, we could not identify a statistically significant difference in frequency among these variants between TCCSG and normal cohorts. On the other hand, a patient who harbored heterozygous PAX5 p.V26G, which had been reported as a somatic mutation in B-ALL, was diagnosed with primary antibody deficiency after treatment of B-ALL. He is receiving immunoglobulin replacement therapy without relapse of ALL for 15 years. Discussion: Our study implied TCF3 p.S514L is associated with ALL development. Besides, we identified a case of heterozygous PAX5 p.V26G variant who exhibited antibody deficiency and B-ALL. These findings suggest those rare variants impair the normal B-cell development and may predispose to ALL development. GWAS is effective in evaluating common genetic variants. However, it is hard to identify the genes involved Mendelian inheritance or functionally defective rare variants associated with disease susceptibility. Although it is rare, our study suggests that rare variants of genes involved in B cell development could be associated with ALL susceptibility and immunodeficiency, and we revealed the frequency of functionally defective rare variants in Japanese childhood ALL cohort. Disclosures No relevant conflicts of interest to declare.

NUDT15 Genetic Variation Is the Strongest Predictive Marker of Tolerance to 6-Mercaptopurine in Japanese Childhood ALL Patients: A Genome-Wide Association Study

Introduction In Asians, NUDT15 genetic variants have been reported to be associated with low tolerance to 6-mercaptopurine (6-MP). However, variation in 6-MP tolerance is still observed among patients without the NUDT15 variants. Some patients required high doses of 6-MP to obtain treatment effect. We performed a genome-wide association study (GWAS) in Japanese to confirm reported associations as well as to identify additional genetic loci associated with variation of 6-MP tolerance in childhood acute lymphoblastic leukemia (ALL) patients. Methods We included ALL patients enrolled into a Tokyo Children's Cancer Study Group (TCCSG) trial who started maintenance therapy by normal protocol dose (30-50 mg/m2/day). Whole genome single nucleotide polymorphism (SNP) microarray genotyping was performed followed by whole genome SNP imputation (Japanese reference genome). Data for over six million SNPs were available for analysis. Association analysis was performed adjusting for age at diagnosis and considering average 6-MP dose during 168 days from maintenance therapy initiation as the outcome. Results A total of 224 patients were included in the analysis. The median age at diagnosis was 4.7 (range = 0.9-15.7) years old and boys comprised 56.3%. The average 6-MP dose during the 168 days of maintenance therapy was 41.1 (range = 13.3-78.5) mg/m2/day. Variants representing 10 genomic regions showed potential association (P < 5 x 10-6) with average 6-MP dose, in which one locus was genome-wide significant (13q14.2, rs116855232). Variant rs116855232 is located in the previously identified NUDT15 gene and represents the strongest association with average 6-MP dose within our Japanese study (β = -11.45, P = 8.5 x 10-10). In contrast, the TPMT locus, which is predictive of 6-MP intolerance in populations of European ancestry, was not associated with 6-MP dose in Japanese. Among the suggestive loci requiring further follow-up is a locus residing in AFF3 (2q11.2, P = 2.1 x 10-6), a gene previously implicated in childhood leukemia. Conclusion We have validated that NUDT15 rs116855232 is a locus with strong association with 6-MP tolerable dose, and it currently represents the most meaningful clinically predictive marker in Japanese. These findings indicated that NUDT15 genotyping should be considered prior to initiation of 6-MP treatment. Prospects for the identification of additional loci of 6-MP tolerance are high after further validation of suggestive loci observed in this Japanese study. Figure Disclosures No relevant conflicts of interest to declare.

Distribution and Clinical Features of NOTCH1 Signaling Activating Alterations in Pediatric T-Cell Acute Lymphoblastic Leukemia (T-ALL)

IntroductionNOTCH1 and FBXW7 alterations leading to aberrant activation of NOTCH1 signaling, classified into two patterns; ligand-independent activation (LIA) and impaired degradation (ID) of NOTCH1. In general, activation of NOTCH1 axis is a hallmark of T-cell acute lymphoblastic leukemia (T-ALL), though comprehensive studies regarding subclonal mutations inducing NOTCH1 activation are still elusive. In the present study, we explored the clinicopathological relevance of NOTCH1/FBXW7 aberrations considering subclonal alterations.MethodsA total of 176 cases with pediatric T-ALL were enrolled in this study. We reanalyzed our previous data of targeted-capture sequencing (n=176) for 158 ALL-related genes/regions and combined with previous expression profiling data based on whole transcriptome sequencing (WTS; n=121). We defined as a subclonal mutation when variant allele frequency was below 0.15 and/or multiple alterations were found within the same pattern of NOTCH1 activation (LIA or ID). All patients were received Berlin-Frankfurt-Münster based chemotherapies with non-minimal residual disease (MRD) based risk stratification, which were mainly offered from the Tokyo Children's Cancer Study Group (TCCSG) and the Japan Association of Childhood Leukemia Study (JACLS).ResultsIn total, we detected aberrations activating NOTCH1 signaling in 81.3% (143/176) of cases including subclonal mutations. Subclonal alterations were observed in 26.7% (n=47). Single nucleotide variations in the heterodimerization domain (HD-SNV) were the most frequent (43.2%; n=76), followed by PEST domain mutations (33.0%; n=58), FBXW7 mutations (26.1%; n=46), non-frameshift indels of NOTCH1 (19.9%; n=35), and in-frame internal duplication known as juxta-membrane expansion (6.3%; n=11). Amplification of NOTCH1 region and 5' NOTCH1 deletion were not detected in our cohort. Both LIA and ID patterns were detected in 43.2% (n=76). Most mutations were mutually exclusive within each LIA and ID pattern.Intriguingly, we detected four (2.3%) internal deletion of NOTCH1 gene (DEL; missing exon 3-27 (DEL3) or 21-27 (DEL21)), three cases (1.7%) of SNV at 3' untranslated region, and two (1.1%) SEC16A-NOTCH1 fusions. These alterations were previously reported to activate NOTCH1 signaling in breast cancer or chronic lymphoblastic leukemia, except for DEL21. We confirmed that DEL21 strongly activates NOTCH1 signaling by luciferase reporter assay (over 100 times compared to wild type NOTCH1). As previously reported in DEL3 and CUTLL cell line, transcripts might initiate at methionine 1737 located within the NOTCH1 transmembrane domain and seem to be sensitive to γ-secretase inhibitors.Analysis of frequency of detected NOTCH1 activating alterations in each previously reported WTS-based cluster (ETP, SPI1, TLX, TAL1-RA, and TAL1-RB) revealed that alterations were frequently detected in TLX (100%; 24/24) and TAL1-RB (95.1%; 39/41), whereas less frequent in TAL1-RA (61.1%; 11/18). In TAL1-RA, all SEC16A-NOTCH1 fusions were observed despite significantly low rate of HD-SNV (11.1%; 2/18). In SPI1 cluster, PEST domain alterations were frequently detected (71.4%; 5/7). Importantly, cases harboring subclonal NOTCH1/FBXW7 alterations showed significantly worse outcome (log-rank P = 0.01), although there was no prognostic difference between cases with and without NOTCH1/FBXW7 mutations.ConclusionsWe observed NOTCH1 activating alterations in 81.3% of pediatric T-ALL cases and detected rare internal deletion of NOTCH1 gene and NOTCH1 fusions recurrently in T-ALL. Furthermore, the presence of subclonal NOTCH1/FBXW7 mutations might be relevant to unfavorable outcome. Despite several limitations such as non-MRD based treatment, our results might be useful for developing a new anti-NOTCH1 therapeutic strategy for pediatric T-ALL patients. DisclosuresNo relevant conflicts of interest to declare.

The Prognostic Value of TP53 Mutations Depends on Clinical Backgrounds in Pediatric Patients with Acute Lymphoblastic Leukemia

IntroductionTP53 mutations in relapsed cases with pediatric acute lymphoblastic leukemia have been implicated in poor clinical outcomes. However, the prevalence and clinical significance of TP53 mutations at diagnosis have not been fully investigated. Such knowledge is essential for the care of patients, because treatment intensity is tailored to predictive prognosis, where increased attention has been directed toward de-escalation of treatment for the problem of long term effects and second malignancies in childhood cancer survivors.MethodsMutation status of TP53 was detected by targeted-capture sequencing of TP53 coding regions in 1,003 children with B-precursor ALL who had been treated in either of the two prospective clinical trials, JACLS (Japan Association of Childhood Leukemia Study) ALL-02 and TCCSG (Tokyo Children's Cancer Study Group) L04-16. Detection of common fusion genes, including BCR-ABL, ETV6-RUNX1, MLL-AF4, MLL-ENL, MLL-AF9, and TCF3-PBX1, were performed using qPCR assays. We designed SNP baits to analyze copy number status of chromosome 17, and also captured 662 probes tiling the entire IgH enhancer locus to identify IGH-DUX4 rearrangement.ResultIn total, 36 different non-silent coding TP53 mutations were identified in 30 (3%) patients, including 22 missense, 7 frameshift indel, 5 in-frame indel, and 2 nonsense mutations. All missense mutations were found in the core DNA-binding domain (n=21), except for one mutation, which affected the tetramerization motif. Variant allele frequencies (VAF) of TP53 mutations varied from 3% to 97% with 14 mutations showing < 10% VAFs. Showing a significant correlation with mutated TP53 (Odds ratio 20: 95%CI 6.4-61, P<0.001), loss of heterozygosity affecting the TP53 locus was observed in 11 (37%) cases and caused by del(17p) in most cases (n=10). We next evaluated clinical features of TP53-mutated cases. TP53 was most frequently mutated in Hypodiploid ALL (33% n=3), followed by MLL rearrangement (12% n=4), IGH-DUX4 (9% n=5), Others (3% n=8), TCF3-PBX1 (2% n=2), Hyperdiploid (2% n=6), and ETV6-RUNX1 (n=2 0.9%). TP53 mutations were not associated with age or white blood cell count at diagnosis. However, significantly more patients were categorized into National Cancer Institute (NCI) high risk (HR) category (Odds ratio 2.4: 95%CI 1.1-5.3, P = 0.03) and TP53 mutation was associated with a significantly shorter overall survival (OS) among NCI-HR patients (n = 16; HR for death, 6.3; 95% CI, 3.1-13; P<0.001). Five-year OS of NCI-HR patients with TP53 mutations was 44%, suggesting that early treatment intensification or alternative treatment strategies are warranted for these patients. TP53 mutations were also associated with a shorter OS in MLL rearrangement and IGH-DUX4 ALL. Particularly, 67% (n=4/6) of cases with any cause of death harbored TP53 mutation in IGH-DUX4 ALL. In contrast, TP53 mutation was not associated with shorter overall survival in NCI-SR cases. In Hyperdiploid ALL, 5 out of 6 cases with TP53 mutations were categorized into the NCI-SR category and were all alive. Prognostic impact of TP53 mutation was also investigated using recursive partitioning to generate a hierarchical prognostic model for OS by incorporating genetic subgroups and the NCI risk criteria. This model also demonstrated that the NCI risk criteria was the most important prognostic variable and TP53 mutation was used for stratification of patients only in the NCI-HR category.ConclusionTP53 mutations at diagnosis are common in Hypodiploid ALL and also found in a substantial fraction of MLL rearrangement and IGH-DUX4 ALL, where the mutations predict a poor prognosis. TP53 mutation is also found in NCI-SR cases but may not be associated with poor prognosis. [Display omitted] DisclosuresNo relevant conflicts of interest to declare.

Treatment outcome of children with acute lymphoblastic leukemia: the Tokyo Children's Cancer Study Group (TCCSG) Study L04-16.

The survival rate of children with acute lymphoblastic leukemia (ALL) has increased to approximately 90% after substantial progress in risk-oriented treatment strategies. Between 2005 and 2013, the Tokyo Children's Cancer Study Group (TCCSG) conducted a risk-oriented, non-randomized study, L04-16. The principal aim of this study was to assemble background characteristicsand treatment outcomes, and gather genetic information on leukemic cells under central diagnosis. This report outlines the background characteristics and treatment outcomes of 1033 children with ALL treated according to a TCCSG platform. The 5-year event-free and overall survival (OS) rates for all children were 78.1 ± 1.3 and 89.6 ± 1.0%, respectively. The OS rate was significantly higher in children with B-cell precursor (BCP)-ALL (91.9 ± 1.0%, n = 916) than in those with T-ALL (71.9 ± 4.3%, n = 117, p < 0.001). In univariate analysis for BCP-ALL, children aged 1-6years (5y-OS: 94.2 ± 1.0%), with an initial white blood cell count of < 20,000/μL (94.0 ± 1.0%), high hyperdiploidy (95.4 ± 1.6%), ETV6-RUNX1 (97.4 ± 1.2%) or TCF3-PBX1 (96.9 ± 2.1%), and "Day8NoBlasts" (96.4 ± 1.1%) had the best outcomes. Genetic investigation revealed two novel fusion geneswithin this cohort: ETV6-ZNF385A and ZNF362-TCF4. Our study highlighted the clinical aspects of genomic features of ALL in Japanese children. We provide fundamental information for the further molecular investigation of this disease.

TAL1 Super Enhancer Aberration and Stil-TAL1 Fusion in Pediatric T Cell Acute Lymphoblastic Leukemia

Abstract Background Pediatric T-cell acute lymphoblastic leukemia (T-ALL) is an aggressive hematologic cancer accounting for 10 to 15% of newly diagnosed pediatric ALL cases, and is associated with a poor treatment outcome compared with B-cell ALL especially in relapsed cases despite recent improvement with an intensive treatment. TAL1 deletion on chromosome 1p33 which is also known as STIL-TAL1, is found in 20-25% of cases with T-ALL, which results in overexpression of TAL1 under the control of the promoter of STIL located next to TAL1. Recently, TAL1 super enhancer aberration (TAL1-SE) was reported, which introduce binding motifs for the MYB transcription factor in a noncoding region. This somatic mutation results in aberrant expression of TAL1 without STIL-TAL1. Clinical features of TAL1-SE and its difference with STIL-TAL1 cases have been poorly studied, and this could be a new instrument of treatment stratification in T-ALL. Methods Targeted capture sequencing for coding regions of 151 genes using a custom-made bait including the TAL1-SE region was performed in 132 Japanese T-ALL cases under 15 years old to detect somatic mutations. For the further analysis, whole transcriptome sequencing (WTS) was performed in 88 cases. Screening for STIL-TAL1 fusion by reverse transcription PCR was also performed in additional 44 cases. All cases of TAL1-SE and STIL-TAL1 were validated by Sanger sequencing. Our pipelines,"Genomon 2" algorithm, were used for detection of somatic mutations and fusions. Most samples of DNA and RNA were offered from Tokyo Children's Cancer Study Group (TCCSG), and available clinical data were mainly based on the TCCSG clinical study treated with BFM-based ALL treatment protocol. Results TAL1-SE and STIL-TAL1 were found mutually exclusive in 6 (4.5%) and 24 cases (18.1%), respectively. All patients with TAL1-SE were under 10 years old. TAL1-SE cases showed a favorable 3-year overall survival (3y-OS) compared with STIL-TAL1 cases which were reported to show a good prognosis (83.3% vs 85.9% respectively, p = 0.71). However, 3-year disease free survival (3y-DFS) was 22.2% and 72.3% respectively (p = 0.04), indicating higher relapse rate in TAL1-SE cases, and they were probably rescued by second line treatment. Patients with STIL-TAL1 showed significantly poorer prednisolone response (over 1,000 /µl of blast at day 8, p &lt; 0.05), though they showed favorable 3y-OS, which might be resulted from frequent high WBC counts at diagnosis (&gt; 100×106/l, 59%). Comparison of TAL1 expression using read counts of WTS revealed significantly higher TAL1 expression in TAL1-SE cases than STIL-TAL1 cases (p &lt; 0.05). To reveal the difference of gene expression profile between STIL-TAL1 and TAL1-SE, unsupervised consensus clustering of gene expression data in 24 cases performed WTS (6 cases of TAL1-SE and 18 cases of STIL-TAL1) divided into 2 clusters. All cases with TAL1-SE were grouped into the same cluster with 10 STIL-TAL1 cases (Cluster 1). The other 8 STIL-TAL1 cases were classified as Cluster 2. A genetic landscape of these 24 cases showed more frequent NOTCH1 and/or FBXW7 mutations in Cluster 1 (87.5% vs 37.5%, p = 0.02), whereas mutations in genes encoding transcription factor such as RUNX1 and BCL11B were found only in Cluster 1 (n = 1, n = 3, respectively). There was a tendency of better 3y-OS and 3y-DFS in Cluster 2 (73.9% vs 100%, p = 0.14, 55.6% vs 85.7%, p = 0.15, respectively). Conclusion TAL1-SE and STIL-TAL1 showed a relatively good prognosis, though high relapse rate in TAL1-SE and a poor prednisolone response in STIL-TAL1 were observed. STIL-TAL1 cases were biologically divided into 2 groups based on the expression profile (with TAL1-SE and STIL-TAL1 only), and these 2 groups showed different clinical and genetic features such as 3y-OS or frequency of NOTCH1/FBXW7 mutations. Our findings illustrate the clinicopathological differences between STIL-TAL1 and TAL1-SE cases, and thus, it might be helpful for development a new therapeutic strategy for these patients with TAL1 overexpression T-ALL. Disclosures Kataoka: Boehringer Ingelheim: Honoraria; Kyowa Hakko Kirin: Honoraria; Yakult: Honoraria. Ogawa:Takeda Pharmaceuticals: Consultancy, Research Funding; Kan research institute: Consultancy, Research Funding; Sumitomo Dainippon Pharma: Research Funding.

Identifications of Highly Aggressive Phenotype with SPI1 Overexpression in Pediatric T Cell Acute Lymphoblastic Leukemia/Lymphoma

T-cell acute lymphoblastic leukemia/lymphoma (T-ALL/LBL) accounts for 10% to 15% of newly diagnosed cases of childhood acute lymphoblastic leukemia (ALL), arising from the malignant transformation of hematopoietic progenitors primed toward T cell development, as result of a multistep oncogenic process. However, since the prognostic significance of these genetic alterations in pediatric T-ALL is not clear, genetic basis which contributes aggressive phenotype or progression of pediatric T-ALL is still to be elucidated. Therefore, to discover driver genetic events, which involved in the aggressive phenotype of pediatric T-ALL and to identify it's novel prognostic markers, we performed integrated genetic analysis in a large cohort of T-ALL case.Our cohorts included samples from Tokyo Children's Cancer Study Group (TCCSG) and Japan Association of Childhood Leukemia Study (JACLS). Whole transcriptome sequencing (WTS) was performed in 123 cases. Representative recurrent fusion genes were as follows, SIL-TAL1 (n=25), MLL-ENL (n=5), PICALM-MLLT10 (n=5), and NUP214-ABL1 (n=2). Intriguingly, novel recurrent in-frame SPI1 fusions (STMN1-SPI1 n=2; TCF7-SPI1 n=5) were detected, and RT-PCR analysis in additional 60 cases revealed other 2 TCF7-SPI1 fusions. Thus, SPI1 fusions accounted for 4% of pediatric T-ALL/LBL. Expression data of WTS revealed cases with SPI1 fusion showed significantly higher expression of SPI1 compared to cases without SPI1 fusion, implicating that aberrant high expression of SPI1 involved in leukemogenesis.To address the functional activities of SPI1 fusions, we performed luciferase assay using the reporter vector contained the CSF1 promoter region with SPI1 binding site. Transient transfection of Hela cells with the SPI1 fusions expression vectors as well as the wild type SPI1 expression vector showed strikingly high levels of transcription of the reporter genes, as compared to transfection with the empty expression vector, indicating that both SPI1 fusions have transcriptional activities.Next, to analyze the leukemogenic potential of SPI1 fusions in vitro, we transduced fusions cDNA into mouse double negative T-cells. Since p16(CDKN2A) is frequently silenced in T-ALL, we also used p16 null double negative T-cells. Both wild-type and p16 null double negative T-cells expressing SPI1 fusions showed increased cell proliferation compared to the MOCK cells.We further evaluated the impact of SPI1 fusions on T cells differentiation. TCF7-SPI1 or MOCK vector was transduced mononuclear cells isolated from mouse bone marrow. These cells were cultured under stimulating factors, such as IL6 and TPO for 3 days, and then transplanted into the irradiated mouse. Subsequently, 6 week after transplantation, FACS analysis was performed. Of interest, significantly high population of cells expressing TCF7-SPI1 was observed in the immature single positive stage, implicating that their differentiation was impaired at the pre-T cell stage. These results indicate that novel SPI1 fusions have a potential leukemogenic effect in pediatric T-ALL. We defined SPI1 overexpression cases as outliers of SPI1 expression, resulting in extremely poor prognosis (log-rank p = 1.9 ×10-6). Of note, significant poor outcome was confirmed by univariate and multivariate analysis in cases with SPI1 overexpression cases (log-rank p = 9.3 ×10-6, and p = 2.0 ×10-6, respectively).In conclusion, SPI1 fusions expressing cells expanded and they remained at an immature stage, implicating a potential leukemogenic activity of these fusions. Not only the cases with SPI1 fusions, but also the cases with high SPI1 expression without fusions showed extremely poor prognosis, suggesting the prognostic value of aberrant SPI1 expression in pediatric T-ALL. Although it remains unclear, why cases with SPI1 fusions/high SPI1 expression have a poor prognosis, our results indicate that these cases are genetically distinct subgroup from other pediatric T-ALL. DisclosuresKataoka:Kyowa Hakko Kirin: Honoraria; Yakult: Honoraria; Boehringer Ingelheim: Honoraria. Ogawa:Kan research institute: Consultancy, Research Funding; Takeda Pharmaceuticals: Consultancy, Research Funding; Sumitomo Dainippon Pharma: Research Funding.

The Association Between L-Asparaginase Hypersensitivity and Genetic Variants in Japanese Childhood ALL Patients

BackgroundL-asparaginase is important in successfully treating childhood acute lymphoblastic leukemia (ALL). However, some patients experience hypersensitivity, mechanisms for which have not been fully elucidated. Previous studies, predominately based on populations of European ancestry, have described L-asparaginase hypersensitivity associations with genetic variant tagging the NFATC2, GRIA1, ASNS, and HLA-DRB1genes. The aim of this study was to evaluate the association between L-asparaginase hypersensitivity and genetic variants in Japanese childhood ALL patients.MethodsThis study included 472 Japanese children with ALL who received E. coli derived L-asparaginase comprising 6000 U/m2 intravenously or 10000 U/m2 subcutaneously according to the Tokyo Children's Cancer Study Group (TCCSG) L84-11, L89-12, L92-13, L95-14, L99-15, L04-16, and more recent protocols. L-asparaginase hypersensitivity was defined as experience of fever, rash, and other allergic reaction after L-asparaginase infusion. Through an ongoing genome-wide association study, patient DNA from saliva were genotyped using the Illumina platform and genome-wide single nucleotide polymorphism (SNP) imputation was performed using the 1000 Genomes Project Phase I Version 3 as the reference population. Tests of association between childhood ALL and SNP genotypes across the candidate loci, NFATC2, GRIA1, and ASNS, were performed using logistic regression and assuming a log-additive model of inheritance.ResultsWe observed 47 (10%) hypersensitivity patients, of which 32 patients could not be infused again. These L-asparaginase hypersensitivities were not associated with patients' gender and age in this population. SNPs previously reported in NFATC2 rs6021191, GRIA1 rs4958351, and ANSN rs3832526 were not significantly associated with L-asparaginase hypersensitivity (P = 0.76, 0.99, and 0.53, respectively). For confirmation, the NFATC2 rs6021191 SNP was directly genotyped in a large subset of the ALL patients, but no association was observed (P = 0.44; 377 patients). Evaluation of other variants within those genes and flanking regions showed evidence of association with rs11482430 (NFATC2, P = 0.01), rs558550699 (GRIA1, P = 0.02), and rs7803055 (ANSN, P= 0.04), although not significant after correction for multiple testing.ConclusionThis lack of association with variants reported in populations of European ancestry may be influenced by ethnic specific differences in genetic structure surrounding these variants as exemplified by differences in observed minor allele frequency of the reported NFATC2 rs6021191 and GRIA1 rs4958351 SNPs in Japanese (0.11 and 0.01) and Europeans (< 0.01 and 0.36). Moreover, the type of L-asparaginase, and/or differences in therapeutic protocol may also contribute to inconsistencies with the previous reports. There is indication of an involvement of genetic variation of NFATC2 in L-asparaginase hypersensitivity, but whether the signal is representing the same regions as the previous reports needs to be further examined. DisclosuresNo relevant conflicts of interest to declare.

Genetic Abnormalities and Prognosis in Pediatric B-Precursor Acute Lymphoblastic Leukemia without Conventional Genetic Abnormalities; Tokyo Children's Cancer Study Group

Introduction; Pediatric B-cell precursor acute lymphoblastic leukemia (BCP-ALL) is a heterogeneous disease that can be subdivided according to primary recurrent genetic abnormalities that are strongly associated with characteristic biological and clinical features. In approximately two-thirds of pediatric patients with BCP-ALL, well-characterized genetic abnormalities can be detected by standard genetic analyses. In the remaining BCP-ALL patients, major pathogenic or driver cytogenetic abnormalities have yet to be clarified and they are called B-others. Recent studies using advanced analytical approaches have stratified a variety of subgroups harboring novel genetic abnormalities. For example, Ph-like ALL, iAMP21, deletion of ERG and rearrangements of ZNF384 and MEF2D have been identified in B-others. However, a certain portion of B-others still remains as genetically unclassified.Patients and Methods; 291 children with de novo BCP-ALL were enrolled on Tokyo Children Cancer Study Group (TCCSG) ALL L0416 and L0616 trials (between 2004-2007). Conventional genetic abnormalities, including BCR-ABL1, MLL-AF4, MLL-AF9, TCF3-PBX1, ETV6-RUNX1, hyperdiploidy and hypodiploidy were detected in 16 (5.5%), 3 (1.0%), 2 (0.7%), 18 (6.2%), 54 (18.5%), 65 (22.3%) and 3 (1.0%) cases by G-banding, fluorescent in situ hybridization and molecular analyses. 131 cases (44.9%) were classified B-others. We studied 121 (93.1%) out of 131 B-others ALL of the L0416/L0616 cohort. We performed whole transcriptome sequencing (WTS) of 35 pediatric B-others cases. For the confirmation and screening of the fusion transcripts we performed RT-PCR in a total of 121 B-others cases. Alterations of ERG and iAMP21 were examined by Multiplex Ligation-dependent Probe Amplification (MLPA) kit P327. We also analyzed the known genetic mutations or copy number alterations associated with these patients using the data from WTS and MLPA kit P335 and P200.Results; A total of 47 candidate in-frame gene rearrangements were identified in 29/35 (82.8%) B-others samples. The fusion genes including ZNF384 (n=12, 4.1%) or MEF2D (n=6, 2.1%) were frequently found in this cohort. All of patients with ZNF384-related fusion genes were found to have a weak or negative expression of CD10. In addition, 7 and 17 patients out of 22 patients with ZNF384-related fusion genes exhibited more than 20% expression of CD13 and CD33. We identified 6 recurrently mutated genes, including NRAS, KRAS, PTPN11, EZH2, MLL2 and ASH1L. Especially, activating mutations in RAS signaling pathway genes, including NRAS, KRAS and PTPN11 were detected in 4 out of 10 ZNF384-translocated patients. Mutations in B-cell developmental genes such as PAX5, BTG1 and IKZF1 were not detected at all in DNA samples with ZNF384-related fusion genes. We identified a total of 6 (1.8%) patients with the fusion genes related with Ph-like ALL, including IGHM-CRLF2, P2RY8-CRLF2, IGH@-EPOR and PAX5-JAK2. Three of 6 patients with Ph-like ALL related fusion gene relapsed and two patients died. We identified iAMP21 in 6 (2.1%) patients and deletion of ERG in 7 (2.4%). One of 6 patients with iAMP21 and 3 of 8 patients with ERG deletion had IKZF1 deletion. No mutations of CRLF2 and JAK2 were identified in patients with iAMP21 and ERG deletion. Among 6 patients with iAMP21, three patients relapsed, and two patients died. No patients with ERG deletion relapsed or died.Conclusions; WTS analysis unmasked a complexity of gene rearrangements in pediatric BCP-ALL without conventional genetic abnormalities. In this study, 10.6% of genetic alterations were revealed in Japanese pediatric B-others cohort. Apart from increasing our understanding of the pathogenesis of pediatric BCP-ALL, this may help improve risk stratification and eventually increase the therapeutic option. DisclosuresNo relevant conflicts of interest to declare.

Analysis of gene expression and DNA methylation patterns in childhood acute lymphoblastic leukemia

The 5-year survival rate of patients with childhood B-cell precursor acute lymphoblastic leukemia (BCP-ALL) now exceeds 90%, though there are still patients who fail to achieve remission or relapse early. To improve the outcomes of these cases, new diagnostic markers for stratification of those with unfavorable outcomes and novel targets for treatment have been investigated based on data from the OMICs analysis. We performed gene expression analysis of leukemic cells from 91 near-diploid BCP-ALL cases without specific fusion genes enrolled in Tokyo Children's Cancer Study Group (TCCSG)-L0416 & L0616 clinical trials employing the Affymetrix Human Genome U133 Plus 2.0 Array. Among them, DNA methylation status was analyzed in 24 cases by using the Infinium HumanMethylation450 BeadChip. Herein, initially, the current situations of gene expression analysis and DNA methylation analysis of childhood BCP-ALL are reviewed. Then, our analyses of gene expressions and DNA methylation related to the prognosis of childhood ALL without fusion genes are presented.

Genetic Abnormalities and Prognosis in Pediatric B–precursor Acute Lymphoblastic Leukemia Treated on Tokyo Children′s Cancer Study Group Protocol

Introduction; Acute lymphoblastic leukemia (ALL) is the most common cancer in children. Most childhood cases of B-precursor ALL (BCP-ALL) may be subclassified by the presence of either gross or submicroscopic genetic alteration. Recent genomewide analyses have identified genetic or cytogenetic subtypes, such as Ph-like ALL, deletion of ERG, intrachromosomal amplification of chromosome 21 (iAMP21) in BCP-ALL. Deletion of ERG and iAMP21 were associated with frequent IKZF1 deletions, which were found in Ph-like ALL associated with poor prognosis. However, the frequency and prognostic implications of these genomic abnormalities have not been defined in Japan.Patients and Methods; We studied 208 children with de novo BCP-ALL enrolled on Tokyo Children Cancer Study Group (TCCSG) ALL L0416 and L0616 trials. All patients had a diagnosis of BCP-ALL without BCR-ABL1 phenotype confirmed by morphology, immunophenotyping and mRNA Real-time PCR. Alterations of ERG and iAMP21 were examined by Multiplex Ligation-dependent Probe Amplification (MLPA) kit P327-B1. To identify Ph-like ALL and BCR-ABL1 like ALL, We investigated gene expression of 235 BCP-ALL cases enrolled on TCCSG L0416 and L0616 trials by microarray GeneChip Human Genome U133 Plus 2.0 and analyzed using clustering analysis as described previously. To assess whether these genomic alterations overlap with known gene abnormalities, such as mutations of CRLF2,JAK1, JAK2 and IL7R, rearrangement of P2RY8-CRLF2, PAX5-JAK2 and PDGFRB-EBF1 genes and copy number alteration of IKZF1, CDKN2A/2B, PAX5, EBF1 and SH2B3 genes, mutational analyses of these genes were also performed using direct sequencing, RT-PCR or MLPA kit P335-A1 and P200.Results; We identified Ph-like signature in 9 (3.8%) patients and BCR-ABL1 like ALL signature in 11 (4.7%) patients. Five (55.6%) of 9 Ph-like ALL patients and 3 (27.3%) of 11 BCR-ABL1 like ALL patients have IKZF1 deletions. Five patients were classified into both Ph-like ALL and BCR-ABL1 like ALL. Among these 5 patients, one patient had CRLF2 high expression, JAK2 R683S/R683G mutation, CDKN2A/2B deletion and EBF1 deletion, and another patient have PAX5-JAK2 rearrangement. These two patients relapsed, and the latter patient died. Three of 9 Ph-like ALL patients relapsed and 2 died, while 7 of 11 BCR-ABL1 like ALL patients relapsed and 2 died. We identified iAMP21 in 5 (2.4%) patients and deletion of ERG in 8 (3.8%). One of 5 patients with iAMP21 and 3 of 8 patients with ERG deletion had IKZF1 deletion. One patient with iAMP21 had BCR-ABL1 like ALL signature, but not Ph-like ALL signature. CRLF2 high expression was identified in two of 5 patients with iAMP21, and P2RY8-CRLF2 rearrangement and SH2B3 deletion was identified in one of them. No mutations of CRLF2 and JAK2 were identified in patients with iAMP21 and ERG deletion. Among 5 iAMP21 patients, patient with BCR-ALB1 signature and IKZF1 deletion relapsed, and one patient died. No patients with ERG deletion relapsed or died.Conclusions; Using GeneChip microarray analysis, low frequency of Ph-like ALL signature and BCR-ABL1 like ALL signature was found. Using MLPA as a novel screenings technique, frequency of iAMP21 and ERG deletion in pediatric BCP-ALL was similar to that of previous studies. BCR-ABL1 like ALL and iAMP21 was associated poor prognosis, while ERG deletion was associated good prognosis in pediatric ALL. These data suggest that screening for these genetic abnormalities in pediatric patients with BCP-ALL is critical not only for outcome prediction but also for risk-adapted therapy. [Display omitted] DisclosuresNo relevant conflicts of interest to declare.

Long-Term Outcome of Six Months Maintenance Chemotherapy for ALL in Children: An Extended Follow up Study of TCCSG L92-13

▪Maintenance therapy is a key component of ALL treatment. Although maintenance with 6-mercaptopurine (6-MP) and methotrexate (MTX) is generally less hemato-toxic, it occasionally causes severe complications such as infections, and too long maintenance is possibly associated with lower adherence. Thus, the duration of maintenance should be as short as possible, but excess shortening of maintenance therapy leads to high relapse incidence, as shown in our previous clinical trial, the Tokyo Children’s Cancer Study Group (TCCSG) L92-13 (1992 - 1995, n = 347). In this study, the shortened maintenance therapy to 6 months resulted in EFS of 59.5% at 5.5 years although OS was as good as 81.5%. Especially, higher relapse rate in standard-risk (SR) resulted in EFS of as low as 60.2% (Toyoda Y, et al. J Clin Oncol 2000). However, it should be noted that about 60% of ALL achieved continuous complete remission (CCR) with short maintenance therapy. Thus, to confirm long-term outcome of ALL with this short maintenance therapy, we conducted an extended follow up study of patients enrolled on the TCCSG L92-13.In the L92-13 trial, previously untreated children (15 years or younger) with ALL were enrolled, and were assigned to three risk groups, standard-risk (SR), high-risk (HR) and extremely high-risk (HEX), according to age at diagnosis, immunephenotype, sentinel cytogenetics and initial leukocyte count. All patients received four-drug induction, followed by consolidation therapy including re-induction. Maintenance therapy with daily 6-MP and weekly MTX was shortened to 6 months, and all treatment was discontinued at 12 months after diagnosis.The data was analyzed as of 2014 July. A median of follow up period of this extended study was 14.8 years, and the survival curve is shown in Figure 1A. Relapse was reported in 128 patients, with a median time of relapse was 1.8 years from diagnosis. EFS at 12 years for all patients was 58.7 +- 2.7% (59.2% for SR [n = 127], 57.8% for intermediate-risk [IR, n = 122], and 59.2% for high-risk [HR, n = 101]) and OS was 78.7 +- 2.2% (85.4% for SR, 80.0% for IR, and 69.0% for HR). Incidences of relapse and non-relapse mortality were 37.2 +- 2.6% and 3.2 +- 0.9%, respectively. Five patients (0 in SR group, 3 in HR group, and 2 in HEX group) developed secondary malignant neoplasms before relapse, which resulted in cumulative incidence of secondary malignancy of 0.9 +- 0.5% at 12 years.Patient gender was associated with the outcome, and female had better EFS (65.3 +- 3.7%) compared to male (52.2 +- 3.8%) (p = 0.04, Figure 1B). High-hyperdiploid (HHD) patients had relatively worse prognosis, with EFS of 50.0 +- 11.8%, although it is generally considered as good prognostic factor, and most of the relapsed HHD patients were salvaged and OS was 94.4 +- 5.4%.It is confirmed that 6 months maintenance therapy is insufficient for male, HHD, and standard risk patients. In contrast, most of patients who had been in first CR at previous analysis maintained CCR status. Our results demonstrate that short maintenance therapy can cure a substantial portion of ALL with extremely low non-relapse mortality, and that it may be possible to shorten the duration of maintenance therapy for female and non-HHD patients. This study provides precise information of leukemia biology and highlights the role of maintenance therapy. Ongoing molecular characterization of the cured patients will clarify the subset of patients who can be cured with short maintenance therapy. [Display omitted] DisclosuresNo relevant conflicts of interest to declare.

Outcome of Adolescent and Young Adults with Acute Myeloid Leukemia Treated with Pediatric Protocols: A Report from the 3 Japanese Cooperative Studies

BACKGROUND: There have been conflicting results from several large pediatric AML collaborative groups comparing survival of adolescent and young adults (AYAs) to that of younger patients with AML. We conducted a retrospective analysis using data from 3 Japanese pediatric AML studies; AML99 (2000-02), AML9805 (1998-2002), and AML-05 (2006-10), conducted by the Japanese Childhood AML Cooperative Study (consisted of Tokyo Children’s Cancer Study Group (TCCSG), Japan Association of Childhood Leukemia Study (JACLS), and Kyushu Yamaguchi Children’s Cancer Study Group (KYCCSG)), the Japanese Childhood Cancer and Leukemia Study Group (CCLSG), and the Japanese Pediatric Leukemia/Lymphoma Study Group (JPLSG, the first Japanese national study group), respectively.PATIENTS & METHODS: Among the 782 patients with de novo AML (excluding acute promyelocytic leukemia and Down syndrome), 44 were AYAs (age ≥15 years), 574 were age 2 to 14 years, and 164 were infants (age <2 years) at diagnosis. Because of the underlying differences in biological characteristics, further analysis on event-free survival (EFS), overall survival (OS), cumulative incidence of relapse (RR) and treatment-related mortality (TRM) were compared between AYAs and patients 2 to 14 years old.RESULTS: 29.5% (13/44) of AYAs were t(8;21) (vs. 37.1% of 2 to 14 years old, P=0.315), 38.6% (17/44) were normal karyotype (vs. 22.0%, P=0.011), and 16.2% of the tested cases (6/37) were FLT3-ITD positive (vs. 9.4%, P=0.554). Complete remission (CR) rate after 2 induction courses was similar between AYAs and patients 2 to 14 years old: 88.6% vs. 88.8% (P=0.965). Five-year EFS rate was also similar between the two cohorts: 55.2% vs. 59.2% (P=0.578). However, 5-year OS rate was significantly poorer in AYAs: 54.7% vs 73.8% (P=0.005). Total 19 events were observed among the 44 AYA patients; 4 failures to achieve CR, 12 relapses, and 3 deaths. Fourteen out of the 16 patients with non-CR or relapse eventually died and 8 deaths were treatment-related. In fact, 5-year RR was similar between the 2 age groups (32.7% vs. 30.1%, P=0.819), but TRM was significantly higher in AYAs (33.2% vs. 12.4%, P=0.001). In multivariate analysis, using Cox proportional hazard regression model, age ≥15 years old at diagnosis, absence of low risk cytogenetics (t(8;21) or inv(16)), and no remission after initial induction course were associated with both poorer OS and higher TRM, while high white cell count at diagnosis (≥ 100,000/μL) and presence of high risk cytogenetics (-7, 5q-, t(16;21)(p11:q22)/FUS-ERG, Philadelphia chromosome, and/or FLT3-ITD) were not siginificant.CONCLUSIONS: AYAs with AML had inferior OS, but not EFS, due to higher incidence of TRM especially after failure of initial treatment. DisclosuresNo relevant conflicts of interest to declare.

Outcome of pediatric acute lymphoblastic leukemia with very late relapse: a retrospective analysis by the Tokyo Children's Cancer Study Group (TCCSG).

Relapse period is strongly associated with second relapse risk in relapsed acute lymphoblastic leukemia (ALL) in children. In this context, the treatment outcome of very late relapse should be better; however, data regarding very late relapse is limited. We retrospectively analyzed the outcomes of two consecutive Tokyo Children's Cancer Study Group (TCCSG) ALL trials (1995-2004) with a focus on late relapse, which was divided into two categories: late relapse (6-24 months from the end of therapy, n = 48) and very late relapse (>24 months from the end of therapy, n = 57). Forty-three patients (29 late relapse and 14 very late relapse) received allogeneic hematopoietic stem cell transplantation (HSCT) at second remission. The event-free survival (EFS) probabilities of late relapse and very late relapse were 54.5 ± 7.3 and 64.8 ± 6.8 % at 7 years, respectively (P = 0.36), and were not significantly different. However, the second relapse incidence of late relapse (34.7 ± 7.1 %) was higher than that of very late relapse (15.5 ± 5.1 %, P = 0.03). The second relapse risk was low for very late relapse ALL, which suggests that these patients should be treated without allogeneic HSCT.

Treatment outcomes of adolescent acute lymphoblastic leukemia treated on Tokyo Children’s Cancer Study Group (TCCSG) clinical trials

There is no standard treatment for adolescents aged 15 years or older with acute lymphoblastic leukemia (ALL), although this age group has been reported as having a poorer prognosis compared to younger patients. We retrospectively analyzed the outcomes of three consecutive Tokyo Children's Cancer Study Group ALL trials (1995-2006) of 373 patients aged 10 years or older, with particular focus on adolescents aged 15-18 years (older-adolescents n = 41), compared to those aged 10-14 years (younger-adolescents n = 332). The probability of event-free survival at 8 years was 67.5 ± 7.4 % for the older-adolescents and 66.5 ± 2.6 % for the younger-adolescents (p = 0.95). Overall survival was 70.7 ± 7.1 % for the older-adolescents and 74.3 ± 2.4 % for the younger-adolescents (p = 0.48). The differences between groups in relapse incidence, non-relapse mortality, and death rate during induction were not statistically significant, although the older-adolescents trended towards a higher frequency of having stem-cell transplantation during the first remission. In conclusion, our treatment strategy, which consists of intensive induction and block-type consolidation, provided improved outcomes for patients aged 15-18 years, comparable to those for patients aged 10-14 years.

An Analysis Of Ph-Like ALL In Japanese Patients

▪ Introduction“Ph-like acute lymphoblastic leukemia (ALL)” or “BCR-ABL1-like ALL” is a recently identified subtype of high-risk B-cell precursor (BCP)-ALL with poor outcome and high frequency of IKZF1 deletion. They exhibit a gene expression profile similar to BCR-ABL1-positive (Ph1-)ALL but do not have BCR-ABL1 and other common chimeric genes. Recent study from others has identified rearrangements involving tyrosine kinase genes, such as ABL1, PDGFRB, JAK2, and CRLF2, by employing transcriptome sequencing using next generation sequencer. In the present study, we intended to explore the biologic characteristics and clinical outcomes of Ph-like ALL cases in Japanese patients. MethodWe investigated gene expression of 235 BCP-ALL cases enrolled on Tokyo Children Cancer Study Group (TCCSG) L0416 and L0616 trials by microarray GeneChip Human Genome U133 Plus 2.0 and analyzed using both clustering analysis and Gene Set Enrichment Analysis (GSEA), a computational method that ascertains whether a given gene set is significantly enriched in a list of genes ranked by their correlation with a phenotype of interest. Four-year event-free survival (EFS) rate of the entire cases was 83.2%. ResultsReal-time PCR analysis revealed that 155 cases did not have common chimeric genes (“chimera-”), including BCR-ABL1, ETV6-RUNX1, E2A-PBX1, MLL-AF4, MLL-AF9, and MLL-ENL. Among “chimera-” cases, 11 (4.7% of the entire BCP-ALL cases) patients were identified as ALL with gene expression profile similar to that of Ph1-ALL by clustering analysis. Their 4-year EFS rate and the frequency of IKZF1 deletion were 66.7% and 80.0%, respectively.On the other hand, GSEA detected 26 ALL cases with Ph-like signature (11.1%), including 9 cases those identified to have Ph-like phenotype by clustering analysis, and they exhibited 4-year EFS rate of 58.3% and IKZF1 deletion frequency of 38.1%. Interestingly, GSEA also revealed that 7 cases simultaneously possess ETV6-RUNX1-like signature. When we excluded 7 cases having ETV6-RUNX1-like signature and 5 cases with higher NOM p-value than 0.05, 14 cases were remained and they represented a highly unfavorable outcome (event-free survival 38.1%), while the frequency of IKZF1 deletion was 41.7%.We performed transcriptome sequencing of ALL cases with Ph-like phenotype using next generation sequencer and identified rare and novel chimeric fusions involving tyrosine kinases, such as ABL1, as well as CRLF2-related chimeric fusions. ConclusionsWe identified the presence of ALL subset having Ph-like phenotype in Japanese patients. They represent unfavorable outcome. However, the diagnosis of “Ph-like ALL” varies according to the method to identify Ph-like phenotype. Therefore, the definition of “Ph-like ALL” needs more consideration. Disclosures:No relevant conflicts of interest to declare.

Secondary cancers among children with acute lymphoblastic leukaemia treated by the Tokyo Children's Cancer Study Group protocols: a retrospective cohort study

With improvement in survival, it is important to evaluate the impact of treatment on secondary cancers in acute lymphoblastic leukaemia (ALL) survivors. A retrospective cohort study comprising 2918 children diagnosed with ALL and enrolled on Tokyo Children's Cancer Study Group (TCCSG) protocols between 1984 and 2005 was conducted to evaluate the incidence of secondary cancers and associated factors including treatment protocol, cranial irradiation and other characteristics of the primary ALL. Thirty-seven patients developed secondary cancers, including acute myeloid leukaemia (n=11), myelodysplastic syndrome (n=5), non-Hodgkin lymphoma (n=2), brain tumours (n=13) and other solid carcinomas (n=6) within a median follow-up duration of 9·5years. The cumulative incidence of any secondary cancers was 1·0% (95% confidence interval (CI), 0·7-1·4%) at 10years and 2·4% (95% CI, 1·5-3·7%) at 20years, respectively. Standardized incidence rate ratio of secondary cancers was 9·3 (95% CI, 6·5-12·8). Multivariate analyses showed an increased risk of secondary cancers associated with the recent treatment protocol and cranial irradiation. There was no evidence of a reduction in secondary cancer incidence despite marked decreases in cranial irradiation use in the recent protocols.

A Brief Use of Imatinib Immediately Before Hematopoietic Stem Cell Transplantation (HSCT) in Children with Philadelphia Chromosome-Positive Acute Lymphoblastic Leukemia (Ph+ALL). Results of the Japanese Pediatric Leukemia/Lymphoma Study Group (JPLSG) Study Ph+ALL04

Abstract▪2554▪This icon denotes a clinically relevant abstract Aims:Children with Ph+ALL generally have a poor prognosis when treated with chemotherapy alone. The timing and duration of the use of imatinib has not been determined. We investigated a role of imatinib immediately before HSCT. Methods:All the patients with ALL were screened for diagnosis of Ph+ALL using RT-PCR. Children with Ph+ALL were enrolled on JPLSG Ph+ALL04 Study within 1 week of initiation of treatment for ALL. Treatment regimen consisted of 5 therapeutic phases: Induction phase (5-drug induction), Intensification phase (high-dose cytarabine and BFM Ib), Re-induction phase (4-drug re-induction), 2 weeks of Imatinib monotherapy phase (23 weeks after diagnosis), and HSCT phase (Etoposide+CY+TBI conditioning). Before and after each phase, minimal residual disease (MRD), the amount of BCR-ABL transcripts, was measured with the real-time PCR method (cut-off 50 copies/microgram RNA). The study was registered in UMIN-CTR (Medical Information, University hospital Medical Information Network - Clinical Trials Registry): UMIN ID C000000290. Results:During the period 2004–08, 42 patients were registered in the Ph+ALL04 study. Out of 42 patients, 37 patients (88%) achieved CR and 7 of 37 patients also achieved MRD-negative after induction phase. There were 13 patients who had no MRD at the beginning of imatinib monotherapy phase, and 14 patients were MRD-negative after imatinib phase, consequently, 14 patients were MRD-negative at the time of HSCT. Six patients relapsed before HSCT. In total, 31 patients received HSCT in 1st CR. All the patients had engraftment and no patients died because of complications of HSCT. Five patients relapsed after HSCT and 4 of the 5 patients were MRD-negative before HSCT and the other patient had detectable MRD although it was less than 50 copies. Twenty-six patients continue to be in 1st CR and MRD-negative for median of 3 years after diagnosis. The 3-year event-free survival rate and over-all survival rate for all the patients was 57% and 80%, respectively (figure 1). Five patients did not achieve CR after induction phase and they were treated with imatinib-contained chemotherapy. Four of the 5 patients achieved CR. All of the 4 patients received cord blood transplantation and remains in continued CR. Interpretation:The chemotherapy we employed was based on the previous high-risk regimen of TCCSG (Tokyo Children’s Cancer Study Group) L-99-15 Study. The chemotherapy was intensive enough to induce MRD-negative in 13 at the time of imatinib phase and 31 of 42 patients were in CR at the time of HSCT (around 25–28 weeks after diagnosis). We planned to assess the efficacy of imatinib immediately before HSCT but it was not possible because of the low amount of MRD in most patients at the beginning of imatinib phase. Conclusion:Although EFS and OS was excellent in this study, 88% of induction rate appeared unsatisfactory and relapse occurred before HSCT in 6 out of 37 patients who achieved CR after induction phase. Earlier and longer use of imatinib may improve EFS in children with Ph+ALL and HSCT may be omitted in a subset of patients who achieve an early and deep remission status. [Display omitted] Disclosures:No relevant conflicts of interest to declare.