ECP 2022 Abstract Book

Virchows Archiv (2022) 481 (Suppl 1):S1–S364 13 with a minimum overall level of agreement. However, in com- parison with general pathologists, the agreement rate between subspecialist evaluators was higher, reaching a moderate level for clone 22C3 and pointing to the possibility of improvement with tutorials. It is noteworthy that even among trained examin- ers, the agreement rate did not reach optimal levels, indicating the existence of challenges in the analysis of PD-L1 expression. PS-02 | Poster Session Head & Neck Pathology PS-02-001 DNA methylation analysis in oropharyngeal squamous cell carcinoma – unraveling novel prognostic biomarkers in a clinico-pathological and molecular genetic study of 51 cases J. Laco*, N. Birknerová, H. Kovaříková, I. Baranová, A. Přikrylová, H. Vošmiková, B. Gajdošová, M. Hodek, M. Vošmik, M. Chmelařová, A. Ryška *The Fingerland Department of Pathology, Charles University, Faculty of Medicine and University Hospital Hradec Kralove, Czech Republic Background & objectives: Hypermethylation of tumour suppres- sor genes leads eventually to malignant transformation. The aim of our study was to determine DNA methylation status of selected tumour suppressor genes in oropharyngeal squamous cell carci- noma (OPSCC) and to find correlation with clinico-pathological characteristics. Methods: A total of 101 samples were analysed in the study (31 primary tumours with 31 corresponding metastases, 20 non- metastasizing primary tumours, and 19 control samples). In every patient, classical clinico-pathological parameters were recorded. For methylation analysis, methylation-specific multiplex ligation- dependent probe amplification (MS-MLPA) was performed for a set of 25 tumour suppressor genes (Probe mix ME002-C1). Results: The study sample comprised 37 males and 14 females, aged 45–80 years (median 58 years). A total of 80% of tumours were HPV-positive. During the follow-up period (range 3–180 months; median 82 months), 10% of tumours recurred and 10% of patients died due to tumour. We observed significantly higher methylation of WT1, PAX6, and CADM1 genes in primary tumours compared to controls (p < 0.05). WT1 and CADM1 genes were significantly hypermethylated in HPV-positive OPSCC compared to HPV-negative OPSCC (p < 0.01). Kaplan-Meier survival curve showed that patients with higher methylation levels of PAX5 gene had impaired overal survival compared to patients with PAX5- unmethylated tumours (p = 0.04). Conclusion: In summary, significant correlation was observed between methylation status of selected tumour suppressor genes and clinico-pathological parameters in our OPSCC study sample. Our promising results unravel novel potential biomarkers which, if confirmed by further studies, could be used as prognostic markers in the sense of tailored therapy and treatment individualization of patients with OPSCC. Funding: This study was supported by the Ministry of Health Czech Republic conceptual development of research organization (UHHK, 00179906), by the Specific University Research Program (SVV 260396) from Charles University, Faculty of Medicine in Hradec Kralove, by the program PROGRES Q40/11, and by the project BBMRI-CZ LM2018125. PS-02-002 Predictive gene expression model for detection of SDHx muta- tion in carotid paragangliomas V. Pavlov*, A. Kobelyatskaya, M. Fedorova, D. Kalinin, A. Golovyuk, G. Krasnov, A. Kudryavtseva, A. Snezhkina *Engelhardt Institute of Molecular Biology, Russian Academy of Sciences, Moscow, Russia Background & objectives: Mutations in SDHx genes occur in more than 40% of carotid paragangliomas (CPGLs). SDHB and SDHD variants are associated with the high risk of metastasis and multifocality, respectively. Identification of SDHx mutation is important for management of patients with CPGLs. Methods: Whole-transcriptome sequencing on an Illumina plat- form and bioinformatics analysis were performed for 71 CPGLs. Based on gene expression data (CPMs), a fully connected neural network (FCNN) was constructed using Keras, Tensorflow, and KerasR libraries. The predictive model was trained and tested on a studied cohort, as well as additionally tested with RNA-Seq data for pheochromocytomas (PCCs) from TCGA. Results: We created a two-step predictive gene expression model for identification of deleterious variants in SDHx genes. At first step, the model defines variants in SDHx genes overall based on the expression of four genes (CEP104, ERP29, EYA3, and NFRKB) and a lncRNA (STXBP5-AS1). The first gene set has the following metrics: sensitivity–0.85/specificity–1/accuracy–0.92/AUC–0.92 for the CPGL cohort and sensitivity–0.52/specificity–0.69/accu- racy–0.61/AUC–0.61 for the PCC cohort. In the second round, the model predicts SDHB and SDHD mutations using data on CEP104, ERP29, and EYA3 gene expression and had the following metrics: sensitivity–1/specificity–1/accuracy–1/AUC–1 for CPGLs, and sensitivity–1/specificity–0.7/accuracy–0.77/AUC–0.85 for PCCs. This work was financially supported by the grant from the Russian Science Foundation (no.21-14-00353). Conclusion: Clinical genetic testing of patients with CPGLs requires target sequencing of four genes ( SDHA , SDHB , SDHC , and SDHD ) or whole-exome sequencing, which is often long and expensive. Immunohistochemistry of SDHB subunit and measure- ment of succinate-to-fumarate ratio were recently proposed as alternative methods for primary prediction of SDHx mutations. However, these approaches do not detect the specific mutated gene. This model allows identifying not only SDHx -mutated tumours but also detect mutated genes ( SDHB and SDHD ) that are essential for tumour management. Funding: This work was financially supported by a grant from the Russian Science Foundation (no. 21-14-00353) and performed using the equipment of the EIMB RAS “Genome” center (http:// www.eimb.ru/rus/ckp/ccu_genome_c.php) . PS-02-003 Molecular pathways associated with SDHx mutations in vagal paragangliomas V. Pavlov*, M. Fedorova, E. Pudova, A. Kobelyatskaya, D. Kalinin, A. Golovyuk, G. Krasnov, A. Snezhkina, A. Kudryavtseva *Engelhardt Institute of Molecular Biology, Russian Academy of Sciences, Moscow, Russia Background & objectives: Vagal paraganglioma (VPGL) is a rare neuroendocrine tumour occurring along the vagus nerve. Approxi- mately half of head and neck paragangliomas are associated with mutations in SDHx genes. However, molecular changes associated with these mutations have not been fully understood. Methods: Illumina whole-transcriptome libraries were prepared for 33 VPGLs with known SDHx status and subsequently sequenced on a NextSeq 500 at 76bp, single-end mode. Raw sequencing data were subjected to the standard bioinformatics analysis. Analysis of differential gene expression and pathways enrichment were per- formed using the RTrans pipeline and KEGG and GO databases. S69