Author of

• 18972

  +

  P2.02 - Biology/Pathology (ID 616)

  • Event: WCLC 2017
  • Type: Poster Session with Presenters Present
  • Track: Biology/Pathology
  • Presentations: 1
  • Moderators:
  • Coordinates: 10/17/2017, 09:30 - 16:00, Exhibit Hall (Hall B + C)

  • 23233

    +

    P2.02-035 - PD-L1 IHC Test on Cytological Cell Block Specimen; Potential Utility and Practical Issues (ID 9018)

    09:30 - 09:30  |  Presenting Author(s): Michiko Sugiyama
    • Abstract

    Background:
    PD-L1 IHC test is an important biomarker for predicting the response of the immune checkpoint inhibitor against the PD1/PD-L1 axis. The FFPE tissue sample is an only validated specimen used in the clinical study, although it is sometimes difficult to obtain an enough tissue sample in advanced stage patients. Cytology specimen is an expected candidate. In this study, we evaluated the PD-L1 IHC expression on cytology cell block specimen (CB) and compared to the corresponding formalin-fixed-paraffin-embedded tumor tissue sample (FFPE-T).
    
    Method:
    Nine primary lung cancer patients who have both surgical resected FFPE-T and pleural effusion CB were recruited. CB was prepared as following; pleural fluid was centrifuged to collect the cell pellet, then fixed in formalin and embedded in paraffin. PD-L1 expression was evaluated using two clones (DAKO PharmDx kit, 22C3 and 28-8). Three pathologists (two certified, one path-trainee) and one cytotechnologist reviewed the slides independently. The proportion score of tumor cell (TPS) was evaluated and divided into 2-tier (positive, negative for 28-8) and 3-tier (no, low, high expression for 22C3) categories, according to the manufactural protocols. The correlation between CB and FFPE-T and the inter-observer agreement (kappa value) were calculated.
    
    Result:
    All samples were acceptable for PD-L1 evaluation. FFPE-T resulted in 2 positive, 7 negative (28-8); 3 low and 6 no expression (22C3), respectively. CB resulted in 5 positive, 2 negative (28-8); 3 low and 6 no expression (22C3), respectively. The TPS and tiered-category of CB did not correlate to those of FFPE-T, statistically. The concordant rate of tiered-category between FFPE-T and CB resulted in 4/9 (45.4%) for both clones. It can be explained by the heterogeneity of PD-L1 expression. The TPS and category judgment of two tests (28-8 and 22C3) within each observer were statistically correlated (R=0.588-0.951, p-value <0.001). The kappa value of the inter-observer agreement varied from 0.18 to 1.0, depending on the experience and education. Two certified pathologists reached moderate (kappa=0.59 for 28-8) to high (1.0 for 22C3) agreement on CB, but low (0.05 and 0.14) on FFPE-T. The kappa value between certified pathologist and path-trainee/ cytotechnologist was 0.6/ <0.01 for FFPE-T, and 0.18/0.57 for CB, respectively. These results seem to be influenced by the recognition of appropriate target tumor cells.
    
    Conclusion:
    Our study suggested that the properly processed cytology sample has a potential clinical utility for PD-L1 evaluation. The difficulty of target cell recognition on cytology specimen seems to be one of the critical issues of standardization.