ECP 2023 Abstracts

S21 Virchows Archiv (2023) 483 (Suppl 1):S1–S391 13 *County Clinical Emergency Hospital of Oradea, Romania Background & objectives: Only half of lung squamous cell carcinoma in situ (SCIS) lesions progress to squamous cell carcinoma (SCC), while a third undergo spontaneous regression. The objective of this study was to predict the evolution of SCIS lesions using deep learning. Methods: The dataset consisted of 112 H&E stained WSI from patients that underwent biopsies of SCIS lesions and were subsequently fol- lowed up for progression to SCC. Tiles of 256 x 256 pixels were extracted from the WSI and a deep learning model was trained to obtain tile-level predictions. A probability-weighted aggregation module was created in order to generate WSI-level predictions. Results: The model achieved a per-tile AUC of 0.78 on the test set, an F1 score of 0.84 and a sensitivity of 0.94. After applying the probability-weighted tile aggregation module, the model achieved an accuracy of 0.75, an F1 score of 0.8 and a sensitivity of 0.90 on the WSI-classification task. Conclusion: To our knowledge, this study is the first to demonstrate that deep learning has the ability to predict the evolution of SCIS from H&E stained WSI. Deep learning has the potential to be used as a low- cost method that could provide prognostic information for patients with preinvasive lesions. OFP-05-009 The role and importance of BAP1, MTAP and 5-HMC immunohis- tochemical markers in diagnosis and prognosis in pleural malig- nant mesothelioma B.N. Ateş*, E. Kilic Bagir, I.C. Karacaoglu, D. Gümürdülü *Department of Pathology, Faculty of Medicine, Cukurova University, Adana, Turkey Background & objectives: Mesotheliomas are the most common primary pleural tumours. They should be distinguished from benign mesothelial lesions, especially. The aim of this study is to investigate the role of the current markers BAP1, MTAP and 5-hmC, in the diag- nosis and prognosis of mesothelioma. Methods: 65 mesothelioma and 28 reactive mesothelial lesions diag- nosed in Çukurova University Faculty of Medicine, Department of Medical Pathology between 2010-2017 were included in the study. BAP1, MTAP and 5-hmC antibodies were immunohistochemically applied to the formalin-fixed paraffin embedded block of the cases, the loss of markers was evaluated. The effects of these markers on survival in mesothelioma cases were also examined. Results: BAP1, MTAP and 5-hmC were 100% specific in distin- guishing pleural mesotheliomas from the benign group; the sensi- tivity of the markers was determined as 67.7%, 61.5% and 52.3%, respectively. When all three markers are used together, the sensitiv- ity reached 95.4%. BAP1, MTAP and 5-hmC were not found to be statistically significantly associated with survival in all or subtypes of mesothelioma. Conclusion: Most of the mesothelioma cases have to be diagnosed from small biopsy materials. It is difficult to make the differential diagnosis of reactive lesion and mesothelioma, especially where the surrounding adipose tissue cannot be observed. Current immuno- histochemical markers BAP1, MTAP and 5-hmC are highly specific for mesothelioma. The combined use of these markers has higher sensitivity than their individual use. Therefore, the algorithmic application of these markers to use BAP1 in the first line will enable the diagnosis of the majority of mesothelioma cases. OFP-05-010 Small samples, big insights: evaluating PD-L1 expression and inter- observer variability in NSCLC N. Tezcan*, L.S. Gucer, C. Aydin Mericoz, P. Bulutay, P. Firat, I. Kulac *Koç University Faculty of Medicine, Department of Pathology, Turkey Background & objectives: Immunohistochemical PD-L1 expression evaluation poses challenges with adequacy of specimens and inter- observer variability. We aim to evaluate the efficiency of small samples and inter-observer variability with the changes in time over assessment in non-small cell lung cancer (NSCLC) cases. Methods: All PD-L1 studies were retrieved from Koç University Hos- pital, Pathology Laboratory System. PD-L1 tests that were performed on thoracic samples with the diagnosis of NSCLC were included to this study. Type of procedure and tumour, PD-L1 tumour proportion score (TPS), and assessor’s name were recorded. In total, 494 tests were included in the study. Positive TPS corresponded to TPS>0. Results: Of 494 specimens, 152 (30.8%) were resections and 342 (69.2%) were small samples, including tru-cuts and cell-blocks. Tumour types comprised adenocarcinoma (61.3%; n=303), squamous-cell carci- noma (18.6%; n=92), large-cell neuroendocrine carcinoma (2.6%; n=13), and NSCLC-NOS (17.4%; n=86). A total of 275 positive TPS (55%) were recorded. TPS results were categorized into three groups: <1% (44.3%), 1-49% (33%), and >49% (22.7%). There was no difference in the distribution of TPS between small samples and resections, assessors or 1-year periods. Overall positive TPS was 30% (median). Assessors’ positive TPS was ranging between 15-35% (median). There was no sig- nificant difference in median positive TPS values between small/large samples, tumour types, assessors or 1-year periods. Conclusion: This study demonstrates the adequacy of small samples (including cell blocks) for immunohistochemical PD-L1 expression evaluation in NSCLC, with no significant difference in median TPS between tumour types, small samples, and resections. Furthermore, TPS distribution showed no statistically significant difference across small samples-resections, years or assessors, with the latter indicating interobserver consistency. These findings support the reliability of the PD-L1 test in small samples and reveal that inter-observer variability does not make a significant difference in the PD-L1 assessment. OFP-05-011 Surgically resected adenocarcinoma in situ and minimally invasive adenocarcinoma of the lung – a case series with follow-up V. Almeida*, L. Veloso, L. Carvalho *Institute of Anatomical and Molecular Pathology, Faculty of Medicine of the University of Coimbra, Portugal, CIMAGO – Research Center for Environment, Genetics and Oncobiology, Faculty of Medicine, Uni- versity of Coimbra, Portugal, Pathology Service, University Hospital Anatomical Pathology Coimbra, Portugal Background & objectives: Pulmonary adenocarcinoma in situ (AIS) and minimally invasive adenocarcinoma (MIA) are the pre-invasive forms of pulmonary adenocarcinoma, with an excellent prognosis. We aim to report our Hospital´s casuistic on surgically treated AIS and MIAs, with clinicopathological and follow-up data. Methods: We reviewed the clinicopathologic findings of surgically treated AIS and MIAs diagnosed between 2015 and 2019 at Centro Hospitalar e Universitário de Coimbra. The median postoperative follow-up period was 58 months (range: 22 - 98 months) and the recurrence-free survival (RFS) and lung cancer- specific survival (LCSS) are the outcomes of interest. Results: We identified 14 cases - 6 AIS and 8 MIAs. Most patients were female (8/14), and the majority were non-smokers (10/14). The patient´s median age was 65 yo (55-76). Seven were previously biopsied and were subsequently subjected to lobectomy, while the oth- ers (7/14) had an intraoperative frozen section study, completed by lobectomy in 5 cases. All were R0 surgeries. The median tumour size was 11 mm (6-29). For MIAs, the invasive component was acinar (5/8), papillary (2/8), and micropapillary (1/8).