ECP 2022 Abstract Book

Virchows Archiv (2022) 481 (Suppl 1):S1–S364 13 Results: Lymph node metastasis was significantly associated with HER2-positive and luminal B breast cancers, higher histo- logic grade, as well as higher Ki67 proliferation index (p<0.01). Independent predictors of nodal positivity included tumour size and vascular invasion (p<0.0001). There were significant dif- ferences between subtypes and nodal positivity, with luminal A cancers (23.3%) had the lowest and Her2-positive (50%) the highest rates (p=0.001). There was no difference in lymph node positivity between PR+ and PR- tumours amongst all subtypes (p=.1). Conclusion: Tumour size and vascular invasion are independ- ent predictors of nodal positivity in this study. HER2-positive breast cancers, higher histologic grade and Ki67 proliferation index are significantly associated with lymph node metastasis. These findings may play a role in guiding regional management considerations. PS-01-027 GATA3 expression in human tumours: a tissue microarray study on 13,204 tumours V. Reiswich, M. Lennartz, F. Büscheck, K. Möller, C. Bern- reuther, G. Sauter, R. Simon, S. Minner*, E. Burandt, A. Marx, T. Krech, A. Hinsch, F. Jacobsen *University Medical Center Hamburg, Germany Background & objectives: GATA3 is a transcription factor involved in epithelial cell differentiation. Positive GATA3 immu- nostaining is used as a diagnostic marker for breast and urothelial cancer but can also occur in other neoplasms. Methods: A set of tissue microarrays containing 16,611 samples from 133 different tumour types and subtypes and 608 samples of 76 different normal tissue types were analysed by immunohistochemistry to comprehensively evaluate GATA3 expression in normal and tumour tissues. Results: GATA3 positivity occurred in 70 different tumour types including 24 (18%) with ≥1 strongly positive tumour. Highest positivity rates occurred in non-invasive papillary urothelial carcinoma (92-99%), lobular (98%) and NST (92%) carcinoma of the breast, cutaneous basal cell carcinoma (97%), invasive urothelial carcinoma (73%), T-cell lymphoma (23%), adenocarcinoma of the salivary gland (16%), and cutaneous squamous cell carcinoma (16%). In breast cancer, low GATA3 staining was linked to advanced pT stage (p=0.03), high BRE grade (p<0.0001), HER2 overexpression (p=0.0085), oestrogen and progesterone receptor negativity (p<0.0001 each) and reduced survival (p=0.03). In urothelial neoplasms, low GATA3 was linked to poor grade within pTa tumours (p=0.01) and with advanced stage (p<0.0001). Conclusion: GATA3 positivity can be seen in various tumour entities. Particularly high frequency and levels of GATA3 expres- sion occur in breast and urothelial carcinoma. A reduced level of GATA3 reflects cancer progression and poor patient prognosis in these tumour entities. PS-01-029 Analysis of intraoperative frozen sections of sentinel lymph nodes - a 10-year long series in a tertiary Portuguese hospital J. Boavida*, R. Moiteiro da Cruz, L. Correia *Department of Pathology - Hospital de Santa Maria, CHULN; Faculdade de Medicina da Universidade de Lisboa, Portugal Background & objectives: Intraoperative frozen sections of sentinel lymph node (IFSSLN) allowed better management of breast cancer (BC) particularly in the decision to perform axillary lymphadenectomy (AL). This study reviews results of IFSSLN, their definitive evaluation and correlation with AL status when performed. Methods: Through our department’s information system, we com- piled every case of BC where IFSSLN was performed along with data pertaining diagnosis and AL status, from 2010 to 2020; 929 out of 1056 cases were selected. Only cases with a diagnosis of invasive disease made in surgical specimens were chosen and AL was solely considered when referred as such by the surgeon. Results: Median age of diagnosis was 59,4 years (range 27-95 years), with a female predominance (n=923). 695 were IFSSNL- negative with 88,9%(n=618) being true-negatives; the remaining showed signs of disease in definitive evaluation (mostly isolated tumour cells[ITC] and 2 cases with metastasis). AL was performed in 7 false-negative cases which revealed metastasis in 2 cases and micrometastasis in 1; 42 false-negatives had nodal evaluation not by AL (either satellite lymph node or axillary inspection), showing metastatic disease in 6 of these (4 metastasis, 1 micrometastasis and 1 ITC). There were no false-positive cases. Two cases were deferred. Invasive breast carcinoma of no special type was the most common diagnosis (n=725). Conclusion: IFSSNL continues to be a useful assessment method in BC staging, showing in this retrospective study highly favourable statistical values when compared with posterior definitive evalua- tion (75,1% sensibility; 100% specificity; 100% positive predictive value; 88,9% negative predictive value); this results are vastly sup- ported by our use of immunohistochemical stains during the proce- dure. 34,5% of true-positive cases benefited from this evaluation, since AL revealed the presence of metastatic disease; furthermore, it spared 571 patients of AL-associated complications. PS-01-030 Determination of inter-observer agreement in the immunohistochemical interpretation of PD-L1 clones 22C3 and SP142 in triple-negative breast cancer (TNBC) R. Abreu, R. Peixoto, M. Corassa, W. Nunes, T. Neotti, T. Rodrigues, C. Toledo, T. Domingos, D. Carraro, H. Gobbi, M. de Brot, R. de Paula* *Ac Camargo Cancer Center, Brazil Background & objectives: Advanced triple-negative breast cancer can be treated with immune checkpoint inhibitors, and the expression of PD-L1 is one of the biomarkers used to select patients with clinical benefit. This study analyses the interobserver agreement in the immunohistochemical report of PD-L1. Methods: 168 cases of TNBC were previously tested and diagnosed between 1987-2016 in 3 institutions. Medical records were reviewed for clinicopathological data collection. The study was performed on 2 TMAs, using two clones. The samples were analysed by 2 breast pathology subspecialists and 2 general pathologists without training. The Kappa value was calculated to assess interobserver agreement, with p < 0.001. Results: For clone 22C3, (κ) among the 4 evaluators was 0.387, with κ = 0.625 among subspecialist examiners and κ = 0.264 among general pathologists, with 73% concordant cases. Of the discordant cases, a subspecialist evaluator considered 1.7% samples as positive, while the others classified them as negative. For clone SP142, (κ) among the 4 evaluators was κ = 0.331, with κ = 0.420 among subspecialists and κ = 0.285 among general pathologists, with 109 (64.8%) cases in agreement. Of the discordant cases, in 10.1% one evaluator was inconsistent and interpreted the sample as positive, with the discordant interpretation given by subspecialists in 14 cases and by general pathologists in 3. Conclusion: We demonstrated different rates of interobserver variability between 4 evaluators and different PD-L1 clones, S68