Diagnostic Utility of Whole Genome Sequencing in Adults with B-Other Acute Lymphoblastic Leukaemia

Abstract

Background B-other ALL represents a working definition for patients with B cell precursor (BCP) ALL without a known primary chromosomal abnormality at diagnosis. In this study we use whole genome sequencing (WGS) to characterize the molecular pathogenesis of adult B-other cases (age ≥ 25yrs) from the UKALL14 trial (NCT01085617). Methods Of 652 patients aged 25-65yrs enrolled onto UKALL14, 333 (51%) had B-other ALL. Sufficient material was available to screen 156/333 B-other cases for recurrent Ph-like fusion events (CRLF2, JAK2, ABL1, ABL2 & PDGFRB) using FISH and MLPA (kit P335). This identified 28 (18%) "Ph-like" fusion events (21 CRLF2, 5 ABL-class and 2 JAK2). Of the remaining 128 B-other cases 57 had available samples for tumor normal paired WGS (read depth 60x and 30x respectively). Diagnostic RNA was available for 33 of the 57 cases Bioinformatic analysis was performed to determine small somatic mutations; single nucleotide variants (SNVs) and insertion/deletions (INDELs) as well as copy number aberrations (CNA) and structural variants (SV). We developed WGS approaches to identify IGH enhancer hijack and DUX4 rearrangements, utilizing gGnome and GRIDSS respectively. For 33 cases RNA-seq data was analyzed to validate fusion genes and genetic classification into one of 18 ALL subtypes using the ALLSorts v2 classifier. Results 52/57 cases passed QC. In 5 (10%) cases an established WHO2016 genetic subtype was identified that was missed by standard of care including three high-risk (2 low hypodiploid and a near haploid) and 2 standard risk (high hyperdiploid and TCF3::PBX1) cases. Cytogenetic analysis for these 5 cases either failed (n=2) or produced a normal karyotype (n=3). The WGS inferred tumor purity for these 5 cases was variable with a median of 46% and a range of 33% to 96% compared to 89% (range 13%-99%) for the remaining cases (n=47) in the dataset. Among the remaining 47 B-other cases, WGS analysis identified abnormalities enabling 41 (87%) cases to be classified into a genetic ALL subtype: DUX4 rearrangements (DUX4r, n=8), PAX5alt (n=7), PAX5 P80R (n=4), ZNF384 rearrangements (ZNF384r, n=5), ZEB2/CEBP (n=4), MEF2D rearrangements (MEF2Dr, n=3), UBTF::ATXN7L3 fusions (n=2), IDH1/2 mutations (n=2), JAK-STAT abnormalities (n=2), along with single cases of IGK::BCL2, IGH::CEBPA, IGH::ID4 and IGH::MIR125B1. WGS resolves 8 cases with complex karyotype into one of 6 different genomic subtypes: 2xDUX4r, 2xMEF2Dr, ZEB2/CEBP, UBTF::ATXN7L3, IGK::BCL2 & IGH::ID4. The MEF2Dr subtype (n=3), is enriched in chromothripsis (2/3), and complex SVs targeting the MEF2D gene. Lastly, in all cases with PAX5 mutations (PAX5 P80R (n=4) and PAX5alt (n=4)) we observe bi-allelic targeting of locus by mutations, deletions or LOH. The remaining 6 (13%) B-other cases could not be classified into a recognized B-other subtype. A MYO18A::FGFR1 fusion (t(8;17)(p11;q23)), which most likely maps to the WHO2016 "Lymphoid neoplasms with FGFR1 rearrangement" classification. Two cases had clonal frameshifts in genes associated with T-ALL (PTPN2 p.A108fs*5 and WT1 p.V371fs*14 respectively). A case with a somatic ETV6 p.R399C previously reported in germline risk with a ZNF292::KDM6A fusion. Lastly a case with a clonal mutation typically associated with myeloid disease ASXL1 p.R417*. Among the 52 cases studied, WGS only failed to identify a driver or putative driver in one case, which had the lowest burden of SNVs (861) and an estimated tumor purity, 28%. For the 31 cases with both DNA and RNA, WGS and RNA analysis were concordant for 100% of cases, with RNA identifying an additional KDM6A::CGA fusion. In 3 cases WGS provides evidence that would have challenged diagnosis and/or treatment; IGK::BCL2 with BCL2 and MYC overexpression, IGH::DUX4 with hypermutation (MSH6 loss) and IGH::ID4 with homozygous loss of CD58 and concurrent LOH of HLA-B. Conclusions WGS assigned 88% (46/52) of cases called B-other to an established genetic subtype of ALL. This includes 5 cases with WHO2016 defined subtypes, 3 were high-risk ploidy subtypes that would have changed their UKALL14 risk group and post-induction treatment in the absence of other risk factors. Among the true B-other cases, 87% (41/47) were assigned to one of the newly described genetic subtypes of ALL. Taken together our findings demonstrate the clinical utility of WGS as a diagnostic assay to inform and improve the management of adult B-Other ALL patients in the future.