ECP 2023 Abstracts

S141 Virchows Archiv (2023) 483 (Suppl 1):S1–S391 13 markers PGP9.5 and S100 were combined with antibodies against tyrosine hydroxylase for the detection of sympathetic nerves, VAChT for parasympathetic nerves, and Substance P and CGRP for sensitive nerves detection. Results: In BPH were sympathetic filaments in 1/10 of nerves, only a few scattered delicate parasympathetic and rare sensitive fibres in larger nerves. The pattern of innervation in prostate cancer was analogous to BPH. However, there was a considerably lower amount of nerves of all types, especially in high-grade cases. In high-grade prostate cancer, there were no sympathetic fibres in the central areas of the tumour, with increasing positivity in areas of perineural spread. Most cancer cases (8/10) were negative for sensitive fibres, with only occasional fibres on the periphery of the tumour nest. Conclusion: The presented results may differ from those of other authors. However, the majority of published works are experimental, with minimal data about human prostate cancer. The decrease in nerve density in prostate cancer does not subsidize the role of nerves in it, which has been proven repeatedly. The effects of nerves can be modu- lated by many factors, including neurotransmitter receptors. Evaluation of nerves distribution in different Gleason patterns is important for further research and assessment of neurotransmitter and nerve-targeted therapy. Funding: VEGA 1/0684/21 PS-26-017 Analysis of renal cell carcinoma in end-stage renal disease of Hungarian patients L. Kuthi*, F. Sánta, B. Pósfai, D. Semjén, D. Dénes, Á. Somorácz, A. Fintha, G. Fórika, A. Jenei, T. Micsik, D. Dobi, K. Eizler, N. Giba, B. Iványi, A. Sejben *University of Szeged, Hungary Background & objectives: End-stage renal disease (ESRD) and acquired cystic kidney disease (ACKD) are known risk factors of renal cell carcinoma (RCC). In this study, the clinical and pathological char- acteristics of RCCs developed in ESRD were investigated. Methods: A database of 2597 nephrectomy cases was built, from 6 pathology departments in Hungary. Then, we identified 42 RCCs in 35 patients that developed in ESRD. Demographic, clinical and pathologi- cal parameters, along with the survival data were investigated. RCCs were reclassified according to the current WHO Classification of Uri- nary and Male Genital Tumours. Results: Twenty-eight tumours developed in men and 13 in women, with a median age of 57 years (27-75 years). The causes of ESRD were glomerulonephritis (n=7), hypertensive kidney disease (n=6), autosomal dominant polycystic kidney disease (n=6), chronic pyelo- nephritis (n=4), diabetic nephropathy (n=3), chemotherapy-induced nephropathy (n=1), and undetermined (n=8). ACKD complicated ESRD in 13 patients. The following histological subtypes were iden- tified: clear cell RCC (n=23), papillary RCC (n=8), clear cell papillary tumour (n=5), ACKD RCC (n=3), and eosinophilic solid and cystic RCC (n=2). The median tumour size was 32 mm (10-80 mm), and 39 tumours were kidney-confined (pT1-pT2). No tumour-specific death occurred during the study period. Progression was registered in one patient. Conclusion: In our cohort, the most common RCC subtype was clear cell RCC (56%), with a frequency that exceeded international data appreciably (14-25%). The incidence of clear cell papillary tumour and ACKD RCC (12.2% and 7.31%) was lower than data reported in the literature (30% and 40%). Our results indicate a favourable outcome of RCC in ESRD. This research was funded by the University of Szeged, Albert Szent- György Medical School Research Fund-Hetényi Géza Grant (Grant No. 5S 340 A202) and the New National Excellence Program (Grant No. ÚNKP-22-4-SZTE-305). PS-26-018 Development and validation of a machine learning model for detec- tion and classification of tertiary lymphoid structures (TLS) in renal cell carcinoma (RCC) to predict clinical outcome I. Laklouk* *University of Wisconsin Madison, USA Background & objectives: The prognostic value of TLS in localized RCC is uncertain, and the evaluation of TLSs is hindered by inter- observer variability. To address this, the study aimed to develop a machine-learning model capable of quantifying TLSs automatically using IHC-multiplex whole-slide images. Methods: TLS maturation was graded using routine H&E slides according to previously published criteria. A HALO platform-based machine learning model was developed and confirmed to automatically identify, tally, and classify TLSs in IHC-multiplex whole-slides. A quantitative scoring scheme for TLSs was proposed and its relationship with survival was investigated in patients with localized RCC. Results: 88 patients with localized RCC ≥7cm, underwent com- plete surgical resection at a single institution. Follow-up median of 31 months (IQR 15-62 months). Tumours were categorized into two groups based on the presence of mature TLSs using Brightfield (H&E) staining. Mature TLS (N=38) and no mature TLS (N=50). Using mul- tivariable Cox regression, the impact of TLS on patients’ survival rates was analysed while adjusting for prognostic factors (such as sarcoma- toid and rhabdoid features, and nuclear grade). The presence of mature TLS was linked with a lower early recurrence likelihood (P < 0.05). Furthermore, tumours lacking TLS had significantly worse overall sur- vival rates than those with mature TLS (P < 0.05). Conclusion: In this study, a machine-learning model was developed that is easy to understand and could potentially identify TLSs on IHC- multiplex whole-slides with high precision. This model could be used alongside other methods to assess the risk stratification of localized RCC. To confirm its accuracy, further studies using (H&E) staining whole-slide are needed. PS-26-019 PD-L1 expression heterogeneity in localized renal cell carcinoma tumours and its prognostic value for predicting metastatic progres- sion: evidence from multi-region sampling I. Laklouk*, M.H. Lee, D. Jarrard, T. Borza, G. Allen, K. Richards, W. Huang, E.J. Abel, D. Shapiro *University of Wisconsin Madison, USA Background & objectives: Prior studies show conflicting evidence regarding the prognostic ability of PD-L1 expression in localized renal cell carcinoma (RCC), likely due to expression heterogeneity through- out RCC tumours. We employed multi-region tumour sampling to evaluate PD-L1 expression heterogeneity and PD-L1 prognostic ability. Methods: Tumours from 46 patients with localized RCC were used to construct tissue microarrays containing samples from at least 12 loca- tions within each tumour. Patients were matched on clinical variables and divided into two cohorts (progression vs. no progression). PD-L1 expression was quantified with immunofluorescence (Vectra platform). Heterogeneity was assessed using Kernel-density plots. PD-L1 expres- sion was compared using nested t-tests. Results: 26 patients developed metastatic progression [median fol- low-up of 7 years (IQR 5-11)], and 20 patients remained disease-free [median follow-up of 11 years (IQR 8-15)]. The median tumour size was 9 cm in both cohorts. There was no statistical difference in age, gender, grade, or stage between cohorts. We observed substantial heterogeneity in PD-L1 expression within all tumours, particularly within tumours that progressed. Additionally, tumours that progressed had higher PD-L1 expression (mean optical density 0.025 vs. 0.011,