PF235 COMPREHENSIVE GENETIC TEST TO DETECT ALL GENE FUSIONS IN HEMATOLOGICAL MALIGNANCIES?

Abstract

Background:Patients with (acute) leukemias carry a wide range of chromosomal abnormalities, which affect their prognosis and treatment options. Currently, over 500 different translocations are reported to be involved in the disease progression. Traditional methods to detect chromosomal abnormalities involve a combination of techniques such as karyotyping, FISH, array and RT‐PCR. However, these methods are laborious and at times inadequate. A diagnostic test using RNA (cancer transcriptome) might lead to designing a first tier screening tool to detect genetic aberrations in hematological malignancies.Aims:The primary goal is to develop a diagnostic test to detect multiple gene fusions using a RNA sequencing platform. Here we present a diagnostic assay using the TruSight RNA Pan‐Cancer Panel (Illumina) targeting 1385 cancer genes for gene expression, variant, and fusion detection to detect all relevant fusion genes (both known and novel gene fusion partners) in at least 5% of leukemia's.Methods:RNA taken from 135 bone marrows or blood samples from mainly patients suspected of (acute) leukemia and taken for routine genetic diagnosis were analyzed. Sample analysis was performed randomized and blinded. We performed targeted RNA sequencing using TruSight RNA Pan‐Cancer Panel containing 1358 genes according to standard protocol on a NextSeq and MiSeq platform (Illumina). Data analysis was done using STAR for alignment and MANTA for fusion gene detection (RNA‐Seq Alignment BaseSpace App by llumina).For true gene fusions the following decision tree was used:1. >3 or more paired and split reads necessary; if <3 or only split reads: probably an artifact or not reliable;2. No reads evaluated if in a pseudogene or paralog;3. If both sides called (forward and reverse) and >10 paired and/or split reads: evaluable4. Compare findings with COSMIC data base: 1 fusion match than (probably) true fusion or 1 gene match than check UCSC;5. No match check UCSC BLAT for unknown fusion: if perfect match in UCSC BLAT than interesting fusion.Outcome was compared with results from routine genetic testing (karyotyping, FISH, RT‐PCR or qPCR. Novel fusions were confirmed by PCR and/or FISH.Results:In total we detected 21+6(8) gene fusions in 135 samples; 2 (ARHGAP‐NR3C1) were false positive due to quality limits and interpretation criteria; one sample failed due to insufficient input material in concordance with results from routine genetic testing. We found no fusions in 105 samples. We found 6(8) extra fusions whereof 2 different gene fusions in 3 samples: NUP98‐SET, NUP214‐ABL, EBF1‐PDGFRB, PAX5‐AUTS2, PSIP1‐NUP98, and TGFRBR2‐GPR128. These gene fusions where confirmed by PCR and/or FISH. Not detected were inv(3) or t(3;3), t(14;18) and t(11;14) because these are no fusion genes.Summary/Conclusion:Our Pan Cancer panel showed concordant results for all successful sequenced samples. No false positives were found, while 6 additional gene fusions were detected. Therefore, the Pan Cancer test is suited as a first tier screening tool in (acute) leukemia regarding gene fusions with a 100% sensitivity and specificity. A prospective study, comparing the diagnostic yield of the Pan Cancer panel with current tests and in combination with analyzing expression data, can establish whether this is applicable as a routine, automated procedure with a TAT of one week.