RNAseqCNV: analysis of large-scale copy number variations from RNA-seq data.

Transcriptome sequencing (RNA-seq) is widely used to detect gene rearrangements and quantitate gene expression in acute lymphoblastic leukemia (ALL), but its utility and accuracy in identifying copy number variations (CNVs) has not been well described. CNV information inferred from RNA-seq can be highly informative to guide disease classification and risk stratification in ALL due to the high incidence of aneuploid subtypes within this disease. Here we describe RNAseqCNV, a method to detect large scale CNVs from RNA-seq data. We used models based on normalized gene expression and minor allele frequency to classify arm level CNVs with high accuracy in ALL (99.1% overall and 98.3% for non-diploid chromosome arms, respectively), and the models were further validated with excellent performance in acute myeloid leukemia (accuracy 99.8% overall and 99.4% for non-diploid chromosome arms). RNAseqCNV outperforms alternative RNA-seq based algorithms in calling CNVs in the ALL dataset, especially in samples with a high proportion of CNVs. The CNV calls were highly concordant with DNA-based CNV results and more reliable than conventional cytogenetic-based karyotypes. RNAseqCNV provides a method to robustly identify copy number alterations in the absence of DNA-based analyses, further enhancing the utility of RNA-seq to classify ALL subtype.

Molecular classification improves risk assessment in adult BCR-ABL1-negative B-ALL.

Genomic classification has improved risk assignment of pediatric, but not adult B-lineage acute lymphoblastic leukemia (B-ALL). The international UKALLXII/ECOG-ACRIN E2993 (#NCT00002514) trial accrued 1229 adolescent/adult patients with BCR-ABL1- B-ALL (aged 14 to 65 years). Although 93% of patients achieved remission, 41% relapsed at a median of 13 months (range, 28 days to 12 years). Five-year overall survival (OS) was 42% (95% confidence interval, 39, 44). Transcriptome sequencing, gene expression profiling, cytogenetics, and fusion polymerase chain reaction enabled genomic subtyping of 282 patient samples, of which 264 were eligible for trial, accounting for 64.5% of E2993 patients. Among patients with outcome data, 29.5% with favorable outcomes (5-year OS 65% to 80%) were deemed standard risk (DUX4-rearranged [9.2%], ETV6-RUNX1/-like [2.3%], TCF3-PBX1 [6.9%], PAX5 P80R [4.1%], high-hyperdiploid [6.9%]); 50.2% had high-risk genotypes with 5-year OS of 0% to 27% (Ph-like [21.2%], KMT2A-AFF1 [12%], low-hypodiploid/near-haploid [14.3%], BCL2/MYC-rearranged [2.8%]); 20.3% had intermediate-risk genotypes with 5-year OS of 33% to 45% (PAX5alt [12.4%], ZNF384/-like [5.1%], MEF2D-rearranged [2.8%]). IKZF1 alterations occurred in 86% of Ph-like, and TP53 mutations in patients who were low-hypodiploid (54%) and BCL2/MYC-rearranged (33%) but were not independently associated with outcome. Of patients considered high risk based on presenting age and white blood cell count, 40% harbored subtype-defining genetic alterations associated with standard- or intermediate-risk outcomes. We identified distinct immunophenotypic features for DUX4-rearranged, PAX5 P80R, ZNF384-R/-like, and Ph-like genotypes. These data in a large adult B-ALL cohort treated with a non-risk-adapted approach on a single trial show the prognostic importance of genomic analyses, which may translate into future therapeutic benefits.