CD19-directed immunotherapies have revolutionized the treatment of advanced B-cell acute lymphoblastic leukemia (B-ALL). Despite initial impressive rates of complete remission (CR) many patients ultimately relapse. Patients with B-ALL successfully treated with CD19-directed T cells eventually relapse, which, coupled with the early onset of CD22 expression during B-cell development, suggests that preexisting CD34+CD22+CD19− (pre)-leukemic cells represent an “early progenitor origin-related” mechanism underlying phenotypic escape to CD19-directed immunotherapies. We demonstrate that CD22 expression precedes CD19 expression during B-cell development. CD34+CD19−CD22+ cells are found in diagnostic and relapsed bone marrow samples of ∼70% of patients with B-ALL, and their frequency increases twofold in patients with B-ALL in CR after CD19 CAR T-cell therapy. The median of CD34+CD19−CD22+ cells before treatment was threefold higher in patients in whom B-ALL relapsed after CD19-directed immunotherapy (median follow-up, 24 months). Fluorescence in situ hybridization analysis in flow-sorted cell populations and xenograft modeling revealed that CD34+CD19−CD22+ cells harbor the genetic abnormalities present at diagnosis and initiate leukemogenesis in vivo. Our data suggest that preleukemic CD34+CD19−CD22+ progenitors underlie phenotypic escape after CD19-directed immunotherapies and reinforce ongoing clinical studies aimed at CD19/CD22 dual targeting as a strategy for reducing CD19− relapses. The implementation of CD34/CD19/CD22 immunophenotyping in clinical laboratories for initial diagnosis and subsequent monitoring of patients with B-ALL during CD19-targeted therapy is encouraged.

Development of next-generation sequencing (NGS) has provided useful genetic information to redefine diagnostic, prognostic, and therapeutic strategies for the management of acute leukemia (AL). However, the application in the clinical setting is still challenging. Our aim was to validate the AmpliSeq™ for Illumina® Childhood Cancer Panel, a pediatric pan-cancer targeted NGS panel that includes the most common genes associated with childhood cancer, and assess its utility in the daily routine of AL diagnostics. In terms of sequencing metrics, the assay reached all the expected values. We obtained a mean read depth greater than 1000×. The panel demonstrated a high sensitivity for DNA (98.5% for variants with 5% variant allele frequency (VAF)) and RNA (94.4%), 100% of specificity and reproducibility for DNA and 89% of reproducibility for RNA. Regarding clinical utility, 49% of mutations and 97% of the fusions identified were demonstrated to have clinical impact. Forty-one percent of mutations refined diagnosis, while 49% of them were considered targetable. Regarding RNA, fusion genes were more clinically impactful in terms of refining diagnostic (97%). Overall, the panel found clinically relevant results in the 43% of patients tested in this cohort. To sum up, we validated a reliable and reproducible method to refine pediatric AL diagnosis, prognosis, and treatment, and demonstrated the feasibility of incorporating a targeted NGS panel into pediatric hematology practice.

RNA sequencing (RNA-seq) is a reliable tool for detecting gene fusions in acute leukemia. Multiple bioinformatics pipelines have been developed to analyze RNA-seq data, but an agreed gold standard has not been established. This study aimed to compare the applicability of 5 fusion calling pipelines (Arriba, deFuse, CICERO, FusionCatcher, and STAR-Fusion), as well as to define and develop an integrative bioinformatics pipeline (Fusion InPipe) to detect clinically relevant gene fusions in acute pediatric leukemia. We analyzed RNA-seq data by each pipeline individually and by Fusion InPipe. Each algorithm individually called most of the fusions with similar sensitivity and precision. However, not all rearrangements were called, suggesting that choosing a single pipeline might cause missing important fusions. To improve this, we integrated the results of the five algorithms in just one pipeline, Fusion InPipe, comparing the output from the agreement of 5/5, 4/5, and 3/5 algorithms. The maximum sensitivity was achieved with the agreement of 3/5 algorithms, with a global sensitivity of 95%, achieving a 100% in patients’ data. Furthermore, we showed the necessity of filtering steps to reduce the false positive detection rate. Here, we demonstrate that Fusion InPipe is an excellent tool for fusion detection in pediatric acute leukemia with the best performance when selecting those fusions called by at least 3/5 pipelines.

Cellular ontogeny and MLL breakpoint site influence the capacity of MLL-edited CD34+ hematopoietic cells to initiate and recapitulate infant patients' features in pro–B-cell acute lymphoblastic leukemia (B-ALL). We provide key insights into the leukemogenic determinants of MLL-AF4+ infant B-ALL.

Background: B-acute lymphoblastic leukemia (B-ALL) is a hematological neoplasm of the stem lymphoid cell of the B lineage, characterized by the presence of genetic alterations closely related to the course of the disease. The number of alterations identified in these patients grows as studies of the disease progress, but in clinical practice, the conventional techniques frequently used are only capable of detecting the most common alterations. However, techniques, such as next-generation sequencing (NGS), are being implemented to detect a wide spectrum of new alterations that also include point mutations. Methods: In this study, we designed and validated a comprehensive custom NGS panel to detect the main genetic alterations present in the disease in a single step. For this purpose, 75 B-ALL diagnosis samples from patients previously characterized by standard-of-care diagnostic techniques were sequenced. Results: The use of the custom NGS panel allowed the correct detection of the main genetic alterations present in B-ALL patients, including the presence of an aneuploid clone in 14 of the samples and some of the recurrent fusion genes in 35 of the samples. The panel was also able to successfully detect a number of secondary alterations, such as single nucleotide variants (SNVs) and copy number variations (CNVs) in 66 and 46 of the samples analyzed, respectively, allowing for further refinement of the stratification of patients. The custom NGS panel could also detect alterations with a high level of sensitivity and reproducibility when the findings obtained by NGS were compared with those obtained from other conventional techniques. Conclusions: The use of this custom NGS panel allows us to quickly and efficiently detect the main genetic alterations present in B-ALL patients in a single assay (SNVs and insertions/deletions (INDELs), recurrent fusion genes, CNVs, aneuploidies, and single nucleotide polymorphisms (SNPs) associated with pharmacogenetics). The application of this panel would thus allow us to speed up and simplify the molecular diagnosis of patients, helping patient stratification and management.