ECP 2022 Abstract Book

Virchows Archiv (2022) 481 (Suppl 1):S1–S364 13 Background & objectives: Deregulated germinal centre (GC) reactiv- ity is a main determinant for the SLE pathogenesis. Emerging evidence suggest that TFH cells may play an important role in this process. We sought to compare GC immune cell subsets between SLE and control GCs. Methods: Multiplex imaging was applied for the in situ quantita- tive analysis of GC cell populations in control, selected to express active GCs, (n=4) and SLE (n=4) lymph nodes (LNs). The lat- est Vectra Polaris system and inForm/PhenoptReports software (Akoya) were used. Luminex analysis was performed in matched serum samples from the SLE individuals. Results: No difference was found for GC B and CD8 cell subsets analysed. A strong correlation between CD4hiPD1hi TFH and GC B cells was found only in control LNs. Significantly higher num- bers of less differentiated (CD4hiPD1hiCD57lo) TFH cells were found in SLE compared to control LNs, a profile associated with significantly higher numbers of CD4hiPD1hiCD57lowGATA3low and lower numbers of CD4hiPD1hiCD57loBcl6hiKi67loGATA3hi TFH cells. A clear trend for larger FDC network area and CXCX- L13hi cells within GCs was observed in SLE. Significantly higher extra-follicular IFNa expression, a supporter of initial TFH dif- ferentiation, was found in SLE. Among circulating biofactors ana- lysed, a significant correlation between CXCL13 and CXCL13hi GC cells was found in SLE. Conclusion: We provide evidence for an altered differentia- tion of TFH cells in SLE, characterized by reduced expression of highly differentiated (CD57hi), potential providers of IL4 (GATA3hi) TFH cells as well as a dissociation between TFH and GC B cells. Whether this reflects an aberrant, not affinity-based development of TFH needs further investigation. E-PS-14-011 Digital Pathology reporting of breast core biopsies in Malaysia T. Pasupati Meenakshi* *Gribbles Pathology M Sdn Bhd, Malaysia Background & objectives: Digital pathology has become an integrated component of primary reporting in Gribbles Pathology for the past two years. During the perpetual COVID 19 pandemic in 2021, most of the core biopsies of breast were reported digitally, away from the office. Methods: A highly advanced Aperio AT2 slide scanner (Leica Bio- systems), approved by FDA for primary reporting, is in usage for routine digital pathology reporting in Gribbles Pathology. A total number of 283 breast core biopsy cases were reported, digitally. Reporting was done as per latest international guidelines along with IHC reporting of hormonal status as per CAP protocol. Results: Digital pathology reporting of breast core biopsies is as accurate as conventional microscope reporting, and in many instances, more advantageous and precise. With an integrated Dragon Speech software, all cases were reported with ease, without resorting to the conventional microscope for a review. The response and acceptance from the surgeons are rewarding. Measurements of small malignant foci, and interpretation of grade, coupled with digital photography has made the routine reporting more acceptable. Constant and continuous practice has made the digital reporting more dependable, and accurate than conventional microscope reporting. IHC analysis with comparison of different hormonal receptors in a single view is a futuristic outcome of digital reporting. Conclusion: Digital pathology reporting of core biopsies of breast is simple, dependable, accurate and should be incorporated as a rou- tine primary reporting procedure, where digital pathology is avail- able. IHC reporting is much easier and precise. The accuracy and effectiveness are very advantageous for the current demand of oncol- ogy management in many hospitals with tertiary care. This is the initial step taken by Gribbles Pathology for future diagnostic digital pathology and artificial intelligence (AI) in Malaysia. E-PS-14-012 Classification of histological features based on IHC staining heterogeneity within tumours H. Smits*, L. Ruiter, G. Breimer, S. Willems, M. Philippens *UMC Utrecht, The Netherlands Background & objectives: Haralick texture features are used to quan- tify image heterogeneity. In this study, the heterogeneity of proliferation (Ki67 staining) and immune cells (CD45 staining) within tumours was used to classify histological characteristics of laryngectomy specimens. Methods: 85 whole mount tumour slides of 22 laryngeal or hypopharyngeal carcinomas were immunohistochemically stained for Ki67 and CD45 and scored on histological characteristics. The tumour area was annotated in QuPath. Haralick features independ- ent of DAB-intensity were extracted from the isolated DAB-signal and used as input for a principal component analysis (PCA). A support vector machine was fitted for classification. Results: The PCA included 16 Haralick features (8 from Ki67 and 8 from CD45 staining). The first four principal components (97.2% explained variance) were used to fit a linear classifier. Based on Ki67 and CD45 heterogeneity, only laryngeal vs. hypopharyngeal tumours and cohesive vs. non-cohesive growth showed promising results for classification. The linear classifier resulted in a classification accuracy of 85.9% for laryngeal vs. hypopharyngeal tumours and 76.5% for cohesive vs. non-cohesive growth. A leave- one-patient-out cross validation resulted in an error rate of 0.25 and 0.34 for both classifiers respectively. Conclusion: This study shows the feasibility of tumour classifica- tion by histological characteristics based on the heterogeneity of DAB-stained biomarkers within the tumour. It is a matter of finding the right (combination of) biomarkers for classification by different characteristics and classification of different tumour types. The classifiers created in this study are a proof of concept, since more data is needed to create robust classifiers, but the method shows great potential for automated tumour classification. Funding: Dutch Cancer Society research grant 10978 E-PS-14-013 A thyroid cytology case report using Google Lens for the diagnosis F. Pereira*, C. Padrão, J. Tinoco, A.M. Gonçalves Pereira, A. Alves *Hospital Prof. Doutor Fernando Fonseca, Portugal Background & objectives: Pathologists use manual methods to exam- ine samples through a microscope, but new imaging technologies have been designed to assist in the pathology diagnosis. An example is the use of digital cameras from smartphones in combination with artificial intelligence applications. Methods: We present a case report of a 40-year-old woman, with a personal history of asthma and rheumatoid arthritis, who under- went ultrasound-guided fine-needle aspiration cytology (FNAC) due to increase of the right thyroid lobe. The patient had no asso- ciated symptoms and no analytical changes. Ultrasound revealed slight enlargement of the right thyroid lobe, with two nodules measuring 12x9x6mm and 14x13x11mm. Results: A FNAC was performed, without complications. Cytol- ogy revealed a sparsely cellular sediment, with rare follicular cells S301