Immunohistochemical Determination Of Pd-L1 Expression With Three Different Antibody Clones And Determination Of Sox10 Expression In Triple Negative Breast Cancer.

e15237 Background: In triple negative breast cancer (TBNC), checkpoint inhibitors directed against PD-L1/PD-1 show an improvement of therapy. For the immunohistochemical diagnosis of the predictive biomarker PD-L1, there are various antibodies and kits available. We examined the expression of PD-L1 in TNBC with different antibody clones to compare these results with regard to the staining of tumor cell membranes, staining of immune cells and cytoplasmic staining in order to find out whether the different clones or methods can be exchanged. It was also checked whether PD-L1 or SOX10 expression correlates with the existing clinical parameters. Methods: Breast cancer tissue of 60 patients with TNBC were examined for the expression of PD-L1 and SOX10 by immunohistochemistry. The detection kit used, was the ultraView universal alkaline phosphatase red detection kit for the antibodies anti-human PD-L1 clone 22C3 from Dako, the clone 28-8 from abcam and the SOX10 antibody clone SP267 from Cell Marque. The anti-PD-L1 antibody clone SP142 was detected with OptiView DAB. The cut-offs for the expression of PD-L1 at the tumor cell membrane were < 1%, > 1 to < 50% and > 50%. In the evaluation of SP142 the staining of the immune cells was evaluated with the score: percentage of PD-L1 positive immune cells to the tumor cells that were present. For SOX10, the nuclear staining was evaluated with the score: < 1, > 1 to 50, > 50 to > 100 and 100. The relationship between PD-L1 expression (TC and IC) and SOX10 expression was evaluated with the clinical parameters of the patients from the time of diagnosis until the end of data collection. Results: The antibodies clone 22C3 and clone 28-8 lead to the same results (22,0 % PD-L1 (TC) positive). Clone SP142 showed significantly (p < 0,001) more PD-L1 positive cases(40,7%). For the additional evaluation of the cytoplasmic staining with clone 22C3 and clone 28-8 it could be shown that the PD-L1 Expression is equivalent (40,7%) to the immune cell staining with clone SP142. With exception of Ki67 we were unable to demonstrate any correlation between PD-L1 (membrane and cytoplasmic), SOX10 and other clinical parameters in TBNC. Conclusions: The antibodies clone 22C3 and 28-8 can be used equivalently for PD-L1 determination in TBNC. Clone SP142 showed different results. The cytoplasmatic staining with 22C3 and 28-8 could gain importance in the future because the results were equivalent to the immune cell staining with clone SP142. There is no correlation between PD-L1 and SOX10 in TNBC.