ECP 2023 Abstracts

S132 Virchows Archiv (2023) 483 (Suppl 1):S1–S391 13 Results: Genexus sequencing reported ten cases with 3’/5’ imbalance values. Six cases were ALK 3’/5’ imbalance and five presented other concomitant driver mutations: EGFR, KRAS, MET exon skipping and ALK-EML4 rear- rangement. These five caseswere either ALKFISHor IHCnegative. One case presented ALK 3’/5’ imbalance with FISH negative and IHQ (3+) positive. Four cases presented RET 3’/5’ imbalance, three with concomitant muta- tions; two with EGFR mutations and one with RET fusion CCDC6. All were RET FISH negative. One case presented no other known mutation and RET-FISH was negative. Conclusion: NGS has brought advantages in multiple genes mutations/ fusions detection. This technology is fast, informative and demands less cellular burden to detect novel mutations/fusions. Possible novel fusion mutations (3’/5’ imbalance) detections requires confirmatory analyses. Tumoural cells that contain a gene fusion are often expected to have elevated expression of the 3’ assay compared to the 5’ assay, and these cases have to be confirmed though the other methods – IHC and FISH, to complete Molecular Pathology Reports for target thera- pies prescription. PS-24-003 Moving away from reliance on formalin fixed paraffin embedded tissue- can we use circulating tumour DNA to identify driver muta- tions in diffuse large B-cell lymphoma? N.H. Cutmore*, I. Moutsopoulos, J. Krupka, C. Trethewey, S. Anand, L. Raso-Barnett, D. Hodson *Department of Haematology, University of Cambridge, United Kingdom Background & objectives: Diffuse large B-cell lymphoma is molecu- larly heterogenous. Genetic profiling is critical to improving future therapeutics. Biopsies have many issues: scant material, artefacts, logis- tics. Circulating tumour DNA offers an alternative. Objective: Compare mutation profile in paired samples using custom bioinformatics pipeline. Methods: A custom bioinformatics pipeline, AULE (Automated Ultrasensitive Lymphoma Evaluation) was built. Pre-processing used unique molecular identifiers (UMIs). Somatic variant calling used four callers (SAGE, Strelka2, LoFreq, Mutect2), and only variants called by 3 or more were included in downstream analysis. Mutations were annotated using Variant Effect Predictor. Driver mutations were annotated as previously described (Lacy et al, PMID: 32187361). Results: Paired samples from 23 patients included 23 plasma (ctDNA) and 24 formalin fixed paraffin embedded (FFPE) biopsies (one patient had two primary biopsies). 2 FFPE and 7 ctDNA samples had no driver mutations. The number of drivers between sample types was similar (FFPE: median = 6, IQR = 3-9, n = 22; ctDNA: median = 7, IQR = 5-9, n = 17), and within published range (6-13). 9/17 patients had 100% concordance between sample types and 13/17 had over 85% concord- ance (IQR = 87.5- 100%, n = 17). Cases with concordance below 1st IQR or no drivers have lower DNA concentration input into libraries (median [2.7 ng/uL] vs [16.3 ng/uL]). Conclusion: Genetic profiling is the driver of future improvements in DLBCL treatment. Access to adequate genetic material is vital. ctDNA is attractive and more convenient for patients. Here we demonstrated that ctDNA can be used to genetically profile patient tumours. Caution must be taken when the input DNA concentration is low- this applies to both sample types. We conclude that ctDNA can be used as an alterna- tive to FFPE for genetic profiling in DLBCL. Funding: Cancer Research UK Clinical Research Fellowship: SEBSTF-2021\100001 PS-24-004 Performance evaluation of different genomic testing workflows for the detection of gene-fusions in non-small cell lung cancer J. Fairley*, M. Gupta, I. Simon, S. Deans *GenQA, United Kingdom Background & objectives: Gene-fusions in RET , NTRK 1/2/3 genes and ROS 1 are actionable biomarkers in non-small cell lung cancer (NSCLC) and challenging for clinical laboratories to detect. Per- formance of six different testing workflows in seven independent laboratories to detect these gene-fusions was evaluated. Methods: FFPE blocks (n=11) from resected NSCLC cases with known fusions [ RET (n=7), NTRK3 (n=1), ROS1 (n=1)] and cases without fusions (n=2) were sourced. These were provided to seven clinical laboratories for testing. Results and various quality metrics were collected. Methods evaluated were Oncomine Precision Assay (OPA), TSO500, custom panels for ArcherDx, Qiagen and Roche DNA hybrid-capture, and Idylla Gene-Fusion assay. Results: From the 11 samples used for workflow evaluation in seven labs, concordance, discordance and failure rates were: OPA= 11, 0, 0; TSO= 11, 0, 0; ArcherDx= 8, 0, 3; Qiagen= 7, 1, 3; Roche= 8, 3, 0; Idylla-Lab-A= 9, 2, 0; Idylla-Lab-B= 8, 2, 1. The failures on NGS tech- nologies were mainly because of the low RNA quality or quantity. Some of the discordant results on Idylla were because of known limitations in the assay design. Characteristics of tissue samples, RNA metrics, workflows, partner genes and assay design that explain these failures and discordant results will be discussed. Conclusion: Choosing a genomic technology is difficult and can be impacted by many variables, including sample characteristics typical of the tumour type, the clinically required turn-around-time, and the skill and experience required for each molecular technology in the laboratory. Clinical testing laboratories may benefit from improved understanding of strengths and limitations of different workflows. It is important to keep these in mind when reporting results. PS-24-005 Development of a novel metric to measure genomic instability using unbalanced copy number changes with fast comprehensive genomic profiling M. Gupta, P. Jermann*, G. Lowman, D. Cyanam, E. Norris, V. Mittal, S. Bandla, Y. Jin, M. Torro, L. Pickle, P. Williams, E. Wong-Ho, S. Sadis, J.L. Costa *Thermo Fisher Scientific, Switzerland Background & objectives: One of the primary causes of Homologous Recombination Deficiency (HRD) is BRCA1/2 pathogenic mutations and the consequence is genomic instability (GI). We developed a novel metric termed genomic instability metric (GIM) to measure GI in FFPE cancer samples. Methods: We developed an amplicon-based enrichment next-genera- tion sequencing panel (OCA Plus) that enables comprehensive genomic profiling (CGP) by interrogating 500+ genes relevant to precision oncology including 46 genes in the HRR pathway including BRCA1/2 using the Ion Torrent Genexus system that automates library prepara- tion, templating, and sequencing with next day results. A large cohort of ovarian cancer FFPE samples was sequenced. Results: Genome segmentation was performed using amplicon copy number (CN) log-ratio profiles and allele frequency log odds for thousands of single nucleotide polymorphisms (SNPs) in OCA Plus panel with high minor allele frequencies and GIM was calculated by summarizing unbalanced copy number segments. As expected, we observed GIM could stratify BRCA 1/2 mutated samples from wild type (WT) in the ovarian cancer FFPE cohort. In addition, we noted high GIM values in samples lacking BRCA1/2 mutations and observed over half of the ovarian cancer samples overall exhibited HRD based on BRCA 1/2 mutations and/or high GIM. CGP allowed us to explore associations between mutations in HRR pathway genes and GIM.