ECP 2023 Abstracts

S86 Virchows Archiv (2023) 483 (Suppl 1):S1–S391 13 PS-10-004 Tumour budding characterisation of endometrial carcinoma in molecular groups A. Alpsoy*, G. Koca Yilmaz, G. Erdogan, H.E. Pestereli *Akdeniz University Department of Pathology, Turkey Background & objectives: Endometrial carcinomas, being the most fre- quent gynaecologic neoplasms, are now classified into molecular groups. Tumour budding can be a helpful indicator of epithelial-mesenchymal transition. We evaluated tumour budding to understand the relationship between tumour budding and endometrial carcinoma molecular groups. Methods: We used 55 endometrial carcinoma cases previously molecu- larly classified. We examined the H&E slides of the 55 cases with an Olympus Bx53 microscope and counted the intratumoral and peritu- moral buds. We later converted the intratumoral and peritumoral bud counts into the International Tumour Budding Consensus Conference (ITBCC) tumour budding score and the ITBCC tumour budding score with the ”0 score”. Results: Our cases consisted of 13 FIGO Grade 2 and 42 FIGO Grade 3 cases. Molecular characterization of endometrial carcinoma cases revealed 11 POLE mutant, 14 MSI, 15 no specific molecular profile (NSMP), and 15 P53 mutant cases. Average peritumoral tumour bud- ding was found to be higher in the NSMP group compared with other molecular groups. Average tumour budding was also higher in cases with cervix invasion (p=0,031). The average ITBCC tumour budding score was significantly higher in 15 NSMP cases compared to the other 40 cases’ average ITBCC score (p=0,036). Conclusion: In conclusion, the successful application of the ITBCC tumour scoring system to the endometrial carcinoma cases uncovered that NSMP cases had higher ITBCC peritumoral budding scores than the other molecular groups. PS-10-005 Potential immunohistochemical markers to find pole mutant endo- metrial carcinomas: AMF and AMFR A. Alpsoy*, G. Koca Yilmaz, C. Karadag, O. Birge, T. Simsek, G. Erdogan, H.E. Pestereli *Akdeniz University Department of Pathology, Turkey Background & objectives: Molecular techniques are the sole way of assessing the POLE status of endometrial carcinomas, making molecu- lar classification challenging. Previous study found higher AMF/PGI and AMFR/gp78 in POLE mutant endometrial carcinomas. We utilized AMF/PGI and AMFR/gp78 immunohistochemistry to address this issue. Methods: We used 55 molecularly classified endometrial carcinomas in our institution. We applied AMF/PGI and AMFR/gp78 immuno- histochemistry. The staining intensity was scored as 0 (negative), 1 (weak), 2 (medium), or 3 (strong) and the extent was scored as 0 (0%), 1 (1–25%), 2 (26–50%), 3 (51–75%), or 4 (76–100%). The final score is the sum of intensity and extent scores. Results: The average AMF/PGI score (p=0,003) and average AMFR/ gp78 score (p=0,043) of the POLE mutant group were found to be significantly higher than the POLE wild-type group. AMF/PGI anti- body reached a score of 6 and more on 9 out of 11 POLE mutant cases. Furthermore, the AMF/PGI antibody showed a sensitivity of 81.8% and a specificity of 61.4%. Lower AMF/PGI and AMFR/gp78 scores were observed in cases with malignant peritoneal cytology (p=0,042, p=0,026). The analysis of the AMFR/gp78 score revealed a statistically significant decrease in recurrent cases (p=0.029). Furthermore, AMFR/ gp78 antibody had a low expression in cases with a high peritumoral tumour budding score (p=0,043). Conclusion: In conclusion, our findings point to the possibility of using the immunohistochemical application of AMF/PGI to prioritize high-potential POLE mutant cases in practice. Our project is funded by Akdeniz University BAP(Scientific Research Projects Coordination) Unit with the TTU-2022-5893 project number. PS-10-008 Next-generation sequencing analysis of mixed epithelial carcinomas of the endometrium and ovary J. Betenbuk*, I. Vereczkey, Z.B. Melegh, E. Tóth, A. Simon, A. Kohánka *National Institute of Oncology, Hungary Background & objectives: Mixed carcinoma of the female genital tract is a tumour that comprises two or more distinct histotypes. These tumours can develop along three different pathways, but most of them represent divergent differentiation or transdifferentiation. Collision tumours are rare. Methods: Between 2017 and 2023, 1208 endometrial and 757 ovarian tumours were investigated. After the exclusion of carcinosarcomas, 16 endometrial and 10 ovarian mixed epithelial carcinomas were identi- fied. Following a morphological examination, ER, PR, p53, mismatch repair (MMR) immunohistochemistry was performed, supplemented with PAX8 for ovarian carcinomas. From both localisations, 8 cases were analysed by next-generation sequencing (Oncomime Comprehen- sive Assay v3). Results: 1.32% of both endometrial and ovarian carcinomas were found to be mixed epithelial carcinomas. Of the 8 endometrial tumours we examined, 4 were dedifferentiated carcinomas (DDE), 3 mixed clear cell and low-grade endometrioid carcinomas (CCC/LGEC), 1 mixed CCC and undifferentiated carcinoma (CCC/UC). All of these tumours were MMR deficient in at least one component. Of the 8 mixed ovar- ian tumours, three were endometriosis-associated, all of which showed PIK3CA mutation. The remaining tumours were 1 mixed borderline mucinous tumour and low-grade serous carcinoma (BMT/LGSC), 2 mixed LGSC and high-grade serous carcinoma (LGSC/HGSC). Two contained mixed high grade carcinoma components. The components had shared pathogenic mutations in all 16 examined tumours. Conclusion: In summary, our results show that mixed epithelial endo- metrial and ovarian carcinomas are rare, but thorough morphological and immunohistochemical examination is a reliable tool for their detec- tion. In the samples we examined, synchronous, biologically unrelated, so-called collision tumours did not occur. The distinction of tumours showing transdifferentiation of one histological type to another, from those arising through divergence of two or more histological types from a common progenitor is not always clear. PS-10-009 Improving the diagnosis of serous tubal intraepithelial carcinoma (STIC) using deep learning J. Bogaerts*, J. Bokhorst, M. van Bommel, M. Steenbeek, J. de Hullu, M. Simons, J. van der Laak *Radboud University Medical Center, The Netherlands Background & objectives: Diagnosing Serous Tubal Intraepithelial Carcinoma (STIC), a precursor of high-grade serous carcinoma, is important for individual patient care and investigating new risk-reduc- ing strategies. However, reproducibility of STIC among pathologists is suboptimal, and many pathologists have limited experience with the diagnosis. Methods: We collected, digitalized, and annotated 496 cases. The ref- erence standard was set by a panel of five gynaecologic pathologists. An automated deep-learning algorithm was developed to detect regions that potentially contained STIC. Discrimination of STIC from normal was assessed by analysis of area under the curve (AUC) in ROC curve analyses. Mapped areas of STIC were highlighted for visual review. Results: The deep learning model was evaluated on two independent test sets, one which was acquired from the same data sources as the