ECP 2023 Abstracts

S310 Virchows Archiv (2023) 483 (Suppl 1):S1–S391 13 matching FFPE samples. The DNA yield per mm2 tissue of 3/4 tis- sues was higher for NFPE than FFPE and the 4th was only marginally lower. The ladder PCR for NFPE samples was better than FFPE for all 4 tissues, the NGS library generation capacity per ng DNA was higher for NFPE than FFPE for all 4 tissues and the mean read length in NGS was higher for all 4 NFPE samples compared to the FFPE samples with the same tissue. Conclusion: Formalin is toxic and carcinogenic because it can bind to DNA, leading to mutations that cause cancer, making it a very undesir- able compound to work with. In DNA analysis for diagnostic purposes formalin also is a nuisance because it progressively damages DNA over time, leading to fragmentation and sequencing artefacts mimicking mutations. Using supercritical CO2 and paraffin embedding, without toxic formalin fixation may result in remarkable 5-year conservation of DNA quality, resulting in superior molecular parameters compared to FFPE material. E-PS-15-022 Lack of correlation of MET and PD-L1 expression in non-small cell lung cancer challenges anti-PD-1/PD-L1 and anti-MET therapeutic strategies, per a comparative study of matched biopsies and surgi- cal resection samples M. Ilié*, V. Hofman, C. Bontoux, S. Goffinet, J. Benzaquen, J. Boutros, S. Lassalle, E. Long-Mira, A. Gómez-Caro, C. Cohen, J. Berthet, C. Marquette, P. Hofman *CHU Nice LPCE, Pavillon J, France Background & objectives: Companion diagnostics for the treatment of advanced non-small cell lung carcinoma (aNSCLC) patients include MET expression and PD-L1 tumour proportion score (TPS). To deter- mine the correlation rate between them, we examined matched biopsies and surgically resected specimens from NSCLC patients. Methods: This retrospective analysis assessed the prevalence and cor- relation between MET expression (SP44 clone) and the PD-L1 TPS (22C3 clone) by immunohistochemistry together with molecular altera- tions determined by targeted next-generation sequencing in matched lung biopsy and surgically lung resected specimens from 70 patients with NSCLC. Results: The study discovered a significant correlation between the MET H-score in surgical samples and matched biopsies (P-value<0.0001), as well as between the PD-L1 TPS in paired biop- sies and surgical samples (P-value<0.0001). However, there was no significant correlation found between the MET H-score or expres- sion subgroups and the PD-L1 TPS in both types of paired samples (P-value=0.47 and P-value=0.90). In addition, the MET H-score was significantly higher in adenocarcinoma compared to squamous cell carcinoma (P-value<0.0001). A mutational analysis indicated that the MET H-score was significantly higher in NSCLC cases with targetable molecular alterations (P-value=0.0095), particularly MET amplifica- tions, while no significant correlation was found for the PD-L1 TPS. Conclusion: Our study found no significant correlation between PD-L1 and MET expression in samples from NSCLC patients, highlighting the importance of personalized treatment strategies based on individual expression profiles. These findings provide valuable insight into the development of effective immunotherapy and targeted therapy for NSCLC patients. E-PS-15-023 Creation of contrived plasma samples with RNA variants to use as liquid biopsy reference material M. Jasti*, S. Hernandez, M.D. Chen, K. Bramlett, A. Luchetti *THERMO FISHER SCIENTIFIC, USA Background & objectives: Creating RNA variant containing contrived plasma is challenging due to high concentration of ribonucleases in plasma that degrade exogenous RNA. Current study reports three pro- prietary methods for creating contrived plasma with RNA variants, to be applied for liquid biopsy studies. Methods: The proprietary methods used for creation of contrived plasma samples with RNA variants include three different procedures for introducing the RNA variants into healthy donor plasma. Different RNA variants such as fusions and splicing variants were used for spike in to plasma pool. Following the spike in, nucleic acid was isolated from plasma and sequenced to detect the variants. Results: To evaluate the methods used, nucleic acid from plasma pool with RNA variants spiked in was isolated using a GenexusTM Cell- Free Total Nucleic Acid Purification Kit and GenexusTM Purification System. The isolated nucleic acid was sequenced on an Ion TorrentTM sequencing platform using an AmpliSeqTM HD target amplification assay to detect the presence of variants. Feasibility data for all these proprietary methods showed that exogenously spiked in RNA variants are stabilized by the methods used and increased recovery rates. In addition, varying the ratio of exogenous RNA variant inputs to healthy donor plasma shows a linear recovery trend indicating a good correla- tion between input amount and recovery. Conclusion: This study demonstrates the successful creation of con- trived plasma samples with RNA variants that mimic the biological complexity of human blood samples, which can be used in liquid biopsy tests. The use of both surrogate and contrived samples can foster innovation, when real variant positive samples are difficult to obtain. E-PS-15-024 The Pathology Genomics Imaging Collection: progress on the col- laboration between Genomics England and the National Pathology Imaging Cooperative C. Jennings*, M. Humphries, C. Colquhoun, D. Kaye, G. Stankeviciute, C. Chen, G. Chan, B. Selinsek, R. Chabra, L. Valis, D. Bansal, P. Aru- mugam, D. Westhead, D. Treanor *Leeds Teaching Hospitals NHS Trust and National Pathology Imaging Cooperative, Leeds UK, Genomics England, Queen Mary University of London, University of Leeds, United Kingdom Background & objectives: Datasets containing whole slide images and molecular data have shown potential to advance diagnostic, prog- nostic and predictive understanding of cancer through multimodal deep-learning. Addressing the scarcity of multimodal datasets is a priority and should be informed by understanding data requirements. Methods: A collaboration between Genomics England and the National Pathology Imaging Cooperative will enhance pathology assets of the cancer arm of The 100,000 Genome Project with approximately 250,000 whole slide images and 15,000 COSD compliant, .XML for- mat structured pathology reports. These will accessible with genomic, radiological and clinical data in the new multimodal platform of the Genomics England Trusted Research Environment. Results: Over 70,000 images from 18 contributing NHS sites have been scanned at the National Pathology Imaging Cooperative Scanning facility, generating 100 terabytes of image data so far. Additionally, the results of user discovery research; including expert interviews, a survey of prospective users and focussed user interviews, have been used to inform the data curation, tooling, and architecture decisions in developing the multimodal research environment. Conclusion: As this work nears completion, we share our experiences and lessons learnt for the multimodal research community. We also present the dataset and describe the architecture that has been devel- oped within the Genomics England Research Environment to enable its use alongside the other modes of data available for its participants. Funding: National Pathology Imaging Co-operative, NPIC (project no. 104687), is supported by a £50m investment from the Data to Early Diagnosis and Precision Medicine challenge, managed and delivered by UK Research and Innovation (UKRI).