ECP 2023 Abstracts

S109 Virchows Archiv (2023) 483 (Suppl 1):S1–S391 13 suggest particular HDAC inhibitors as a promising therapeutic target for CRC management by reactivating CDX2 tumour suppressor expres- sion to counteract metastasis. Funding: 1. Swiss Government Excellence Scholarship: https://www. sbfi.admin.ch/sbfi/en/home/education/scholarships-and-grants/swiss- government-excellence-scholarships.html 2. Swiss National Science Foundation: https://www.snf.ch/en PS-17-023 Gastric dysplasia: clinicopathologic features and mucosal charac- teristics according to morphologic subtypes G. Menetlioğlu*, K. Kolçak, Ç. Ataizi Çelikel *Marmara University, Departmant of Pathology, Turkey Background & objectives: Gastric dysplasia (GD) can be morpholog- ically categorized as adenomatous dysplasia (AD), foveolar dysplasia (FD), hybrid dysplasia (HD), basal crypt dysplasia (BCD), and ser- rated dysplasia (SD). This study investigates GD subtypes in terms of clinicopathological features and background mucosal characteristics. Methods: A total of 63 endoscopic biopsies and 65 resection speci- mens diagnosed as GD between 2018-2023 were retrieved from pathol- ogy archives and re-evaluated. GD cases were subtyped (AD, FD, HD, BCD, SD) and degree of dysplasia (low and high grade) was noted. Site of dysplasia, type of gastritis, presence of H Pylori (HP), and intestinal metaplasia (IM) status were assessed. Results: Cases were grouped as AD (64.84%), FD (17.96%), BCD (10.15%), and HD (7.05%). No significant difference was found between groups in terms of demographic data. Most common site was antrum in AD (57.62%), FD (60.86%), and HD (62.5%), while 55.8% of BCD was located in corpus/fundus. There was no difference between AD and FD in respect to the background mucosa (type of gas- tritis, presence of HP and IM). Although chronic gastritis (-/+IM) was detected in most AD and FD cases, BCD and HD were associated with multifocal atrophic gastritis and chronic atrophic gastritis, respectively. AD, FD, and BCD had similar rates of complete/incomplete IM, while IM was rare in HD. Conclusion: Our study revealed no significant difference between AD and FD. Among GD subtypes HD showed distinct background mucosal characteristics, suggesting that it may develop along a differ- ent neoplastic pathway. Contrary to the literature, BCD was not rare in our study. Our observation of BCD in the proximal stomach in the background of multifocal atrophic gastritis may implicate a difference in its pathogenesis. This study contributes to a better understanding of GD subtypes and emphasizes the need for more detailed research. PS-17-024 Digitopathological evaluation of gastroenteropancreatic neuroen- docrine tumours T. Micsik*, L. Csellar, A. Jakab, V. Jonas, B. Molnar *Department of Pathology and Experimental Cancer Research, Hungary Background & objectives: Gastro-Entero-Pancreatic NeuroEndocrine Tumours (GEPNETs) have a grading (Grade1 to Grade3) system rely- ing on mitotic (MI) and proliferation indices (PI) defined by WHO. Manual counting is quite laborious, thus we aimed to use a digital pathology approach in grading GEPNETs. Methods: We scanned 60 GEPNET cases from the archive of the Department of Pathology and Experimental Cancer Research (Sem- melweis University, Budapest) with 3DHistech`s (Budapest, Hungary) digital platform and defined exact PIs with manual (MarkerCounter- MC) and automatic (NuclearQuant-NQ) counting on Ki67-stained immunoslides. Tumour-recognition application (PatternQuant-PQ) was applied for fully automated PI-assessment and detailed cellular morphometrical analysis (cell size, perimeter, shape-factor) was also performed. Results: We found excellent correlations between the PIs, defined by either method. Clinical-PI: defined by the pathologist with visual esti- mation on glass-slides; MarkerCounter-PI: manually counted on digital slides; NuclearQuant-PI: automatically assessed PI with NuclearQuant on automatically selected tumourous regions by PatternQuant. Correla- tions: Clinical vs. MarkerCounter =0,897, Clinical vs. PatternQuant =0,913, MarkerCounter vs. PatternQuant =0,963. Furthermore, we found significant differences in the cellular parameters of the various grades of GEPNETs defined by either method (Clinical, MarkerCoun- ter, PatternQuant). The averages and standard deviations of the tumour cell-areas, tumour cell-perimeters and shape factors of the tumour cells significantly differed in more aspects in relation G1/G2, and in all aspects of G1/G3, and G2/G3 groups. Conclusion: Grading GEPNETs is a laborious work following WHO- guidelines of cell-counting. An exact manual-PI counting on digitized slides is very reliable, but also laborious. Our results showed that manual counting can be replaced by automatic tumour recognition and proliferation index counting, significantly decreasing time and workload and delivering robust data. Furthermore, detailed cellular morphometrical analysis can further increase our accuracy in grading of GEPNETs. Altogether, digital pathological evaluation of GEPNETs is a reliable and robust alternative to ease our everyday routine. PS-17-025 A tertiary centre experience of reporting extramural vascular invasion in rectal cancers: pathology reporting frequencies and correlations with radiology A. Mukherjee*, C. Clarke, C. Santos *University of Nottingham BDI3, University Park, United Kingdom Background & objectives: Extramural vascular invasion (EMVI), a known prognostic marker in colorectal cancer (CRC), remains under- reported histologically. Radiological techniques, especially MRI, diag- nose EMVI quite accurately. This study aimed to explore reporting frequencies and histopathology versus radiology correlations for EMVI in rectal adenocarcinomas. Methods: Rectal adenocarcinoma patients’ records at a tertiary institu- tion were interrogated over a five-year span for EMVI status on histo- pathology and MRI reports. A direct comparison of these investigative modalities was performed, and discrepancies noted. Neo-adjuvant treat- ment status was noted for confounding effects. Resections from acute presentations with unavailable radiology were excluded from analyses. Results: Of 107 cases, 83 (~78%) were matched for EMVI status per both modalities [67 negative; 16 positive]. 32 cases (~30%) were positive for EMVI on pathology; 24 cases (~22%) were positive on radiology. 16 cases (~15%), negative on radiology, were positive on histology. In 3 such, pathology matched pre-chemotherapy rather than pre-operative EMVI on MRI; and 2 cases recorded tumour deposits on MRI but not EMVI. 8 cases (~7.5%), reported positive/probable on imaging, were reported negative by pathology. In 1, the MRI report was downgraded to negative on review. Confounding fibrosis (1), movement artefact (1) on radiology, and fibrosis (1) /tumour deposit (1) on pathol- ogy, was noted in discrepant reports. Conclusion: Histopathology reporting frequency for EMVI in this series of rectal cancers achieved the standards recommended by the Royal College of Pathologists’ minimum dataset. Strong correlations were observed with radiology detected EMVI status. Comparison across modalities could further fine-tune holistic service provision. Particularly, node negative cancers with EMVI positivity on radiol- ogy, but negative on histology, could be investigated further with special stains. Time lapse between the modalities, history of neo- adjuvant treatment and fibrosis may be confounding factors causing discrepancies.