ECP 2022 Abstract Book

Virchows Archiv (2022) 481 (Suppl 1):S1–S364 13 G3,22% NEC). G3 showed poor prognosis than G1-2, but better than NEC. Conclusion: 1)The poorly differentiated group has more frequent genetic alterations and poor prognosis.Br-NET G3 show genetic alterations and a better prognosis distinct from Br-NEC. Whether the Br-NET G3 concept is different from the Br-NEN G3 needs to be explored by more data. 2)The expression pattern of SSTR1-5, DLL3, ISl-1 and INSM-1 in Br-NEN was similar to that in gastroenteropancreatic NEN, so based on these biomarkers, it may be necessary to explore the potential therapeutic significance in Br-NEN. OFP-10-010 Prediction of the efficacy of neoadjuvant therapy for breast cancer with immunologic and genome characteristics Y. Liu*, M. Han *The Fourth Hospital, China Background & objectives: The effect of different immune cells and genomic alterations in tumour microenvironment before neoadjuvant therapy (NAT) in breast cancer patients on efficacy is currently uncertain. This study aimed to explore its impact on efficacy of NAT in breast cancer patients. Methods: 81 cases of breast invasive cancer diagnosed by core needle biopsy before NAT were selected. Multiple Immunohis- tochemical/Immunofluorescence (mIHC/IF) detected CD3, CD8, PD-L1 and PD-1. Next generation sequencing (NGS) technology analysed the genome and detect somatic cell variation. Logistic regression analysis was used to draw a nomogram to predict the pCR rate of breast cancer patients. Results: Patients were divided into two groups by NAT efficacy: pCR and non-pCR. The mIHC/IF results showed expression level of stromaCD8+cells, parenchyma and stroma PD-L1+cells in pCR was higher, with statistically significant differences (P < 0.05).NGS testing results showed the highest mutation rate was TP53(81%), followed by ERBB2(49%), PIK3CA(47%), CDK12(46%) and LRP1B (8%), et al. The mutation types include missense mutation, copy number amplification, synonymous mutation, et al. Only the difference in LRP1B mutation rate between two groups was statistically significant (P=0.011).Cox regression analysis showed expression level of stromal CD8+ cells, stromal and parenchymal PD-L1+ cells and LRP1B mutation rate influenced pCR (P < 0.05). The calibration chart(AUC:0.78) showed that the nomogram performs well and has high pCR predictionability. Conclusion: There is a difference in immune microenvironment and gene expression profiles between pCR and non pCR patients. And the expression levels of CD8 + cells in tumour stroma, PD-L1+ cells in tumour parenchyma and tumour stroma are high, and LRP1B gene is prone to mutationin pCR patients. The nom- ogram of predicting pCR rate has good consistency, which can provide a certain predictive value for breast cancer patients with neoadjuvant therapy. OFP-10-011 Histological assessment of large fresh breast surgical speci- mens with ultra-fast slide-free confocal microscopy: a feasibil- ity study M. Mathieu*, M. Ragazzi, M. Ferchiou, P. Van Diest, O. Casiraghi, M. Nap, A. Ben Lakhdar, N. Labaied, A. Conversano, M. Abbaci *Gustave Roussy Cancer Campus, France Background & objectives: A new generation of ultra-fast confo- cal microscope (UFCM) with a large field of view allows ex-vivo analysis of surgical specimens. We present the standardization of UFCM protocol for breast lumpectomy and the original online training program in breast UFCM images. Methods: Fresh lumpectomies from 55 patients with breast conservative surgery were cut in two, stained with acridine and imaged with Histolog-Scanner-field of view of 20 cm² (Samantree Medical). The images were colored in purple to simu- late frozen sections and annotated by three pathologists expert in breast and confocal pathology. The images were used for a training program followed by two pathologists. Results: All surgical specimens were successfully imaged. Among these 55 cases, 49 were carcinomas (31 invasive no special type, 11 invasive lobular, 7 ductal carcinoma in situ) and in 6 cases no tumour was present at definitive histological examination. The UFCM protocol was completed in 8-10mn. The global architecture of the tissue was evaluated at low magnification and the cellular details at zoom (x40). The 30μm of axial resolution resulted hypercellularity compared to final histology. Training program was composed of 135 reading sheets including tumoural and non-tumoural features. Pathologists who completed the training program performed a diagnosis on the 88 self-assessment sheets getting 100% accuracy. Conclusion: UFCM allowed to quickly image a 20 cm² area of an entire section of fresh breast lumpectomy. The training of patholo- gists in reading UFCM images was successful, probably since the images resemble traditional frozen sections. However a training is mandatory because the magnification of the images is lower than with microscope and the optical section is 30 μm thick. The pilot study points the feasibility of UFCM for fast and extensive intra- operative margin assessment during breast surgery. Funding: SamanTree medical, European Union’s Horizon 2020 research and innovation program under grant agreement No 823284 OFP-10-012 PIK3CA mutations in endocrine-resistant breast cancer C. Schagerholm*, S. Robertson, B. Holm, H. Awier, H. Toosi, E. Sifakis, J. Hartman *Department of Oncology-Pathology, Karolinska Institutet, Sweden Background & objectives: The majority of breast cancer patients’ tumours express oestrogen receptor alpha (ER) but despite endo- crine therapy, one-third progress or relapse. Studies investigate the phosphoinositide-3-kinase pathway, introducing therapies targeting PIK3CA. This study aims to examine the mutational prevalence in endocrine-resistant tumours. Methods: In a retrospectively collected cohort of verified endocrine-resistant breast tumours diagnosed in 2008-2012, primary ER+ and human epidermal growth factor receptor 2 (HER2) negative breast cancers with an ER+/HER2- relapse during ongoing endocrine therapy were included. Targeted gene panel sequencing was performed on formalin-fixed paraffin-embedded tumour tissue and paired analysis was carried out between the relapse and primary tumours. Results: The overall preliminary analysis showed PIK3CA muta- tions in 48.9% of all patients (n/N=23/47), of which 65.2% (n=15) had mutations in the 11 known hotspot regions. 65.2% of the muta- tions were similar between the primary and relapse tumours, 73.3% of these were found in hotspots, whilst 34.8% seem to develop or disappear during ongoing endocrine therapy. In primary tumours of patients that eventually relapsed during endocrine therapy, 60.9% had mutations in hotspots and 30.4% outside of these regions. In their relapse tumours, 56.5% had hotspot mutations and 26.1% elsewhere. Conclusion: Our study of this unique cohort suggests that endo- crine-resistant breast cancers are associated with high PIK3CA S42