Nowadays, the major limit to the immunohistochemical (IHC) analysis of tissue PD-L1 is the high variability of the monoclonal antibodies commercially available. Aims of the present paper are to assess the best clone and the most suitable scoring for PD-L1 IHC determination on human hepatocellular carcinoma (HCC) among three commercially available clones, and to evaluate which PD-L1 clone is the best in predicting HCC aggressiveness in vivo. We built a tissue microarray (TMA) with 60 retrospective HCC cases, including the correspondent non-tumoral tissue. IHC was automatically performed using the following anti-PD-L1 clones: 28.8, SP142, and SP263. As results, we did not find any immunoreactivity for PD-L1 in both neoplastic and normal hepatocytes included in the TMA using the three antibodies. Positivity for PD-L1 was exclusively seen in inflammatory cells within the HCC tissue and in cirrhotic parenchyma. When a gold standard was assessed, the sensitivity of SP142, 28.8 and SP263 was 46 %, 54 % and 85 % respectively. Using the SP263 clone, the absolute number of PD-L1-positive inflammatory cells in the HCC cores was paired with the number of PD-L1-positive inflammatory cells in the corresponding non-tumoral tissue (P = 0.001). Finally, using SP263, the mean number of PD-L1-positive cells was 11.3 ± 12.6 in HCC from deceased patients, versus 4.7 ± 5.2 in alive patients (p = 0.039). SP263 is the most sensitive clone for PD-L1 IHC tissue determination in HCC, as well as the best antibody for the assessment of its biological behavior.
Keywords: Hepatocellular carcinoma; Immunohistochemistry; PD-L1; Tissue micro-array.
Copyright © 2019 Elsevier GmbH. All rights reserved.