Virtual Library

Start Your Search

Frédérique Penault-Llorca



Author of

  • +

    P2.09 - Pathology (ID 174)

    • Event: WCLC 2019
    • Type: Poster Viewing in the Exhibit Hall
    • Track: Pathology
    • Presentations: 1
    • Moderators:
    • Coordinates: 9/09/2019, 10:15 - 18:15, Exhibit Hall
    • +

      P2.09-17 - Real-World Concordance Across Pathologists for PD-L1 Scoring in Non-Small Cell Lung Cancer: Results from a Large Nationwide Initiative (ID 898)

      10:15 - 18:15  |  Author(s): Frédérique Penault-Llorca

      • Abstract

      Background

      PD-L1 immunohistochemistry (IHC) is an important routine biomarker in patients with metastatic and locally advanced non resectable non-small cell lung cancer (NSCLC). Currently, the thresholds of ≥1% and ≥50% of tumor cells stained are clinically relevant. Scoring concordance across pathologists was reported only in small groups of pathologists or across thoracic pathology experts. Here, we provide real-world concordance data in a large group of pathologists (n=161) with various experience of PD-L1 testing and practice in thoracic pathology.

      Method

      Twenty-nine NSCLC samples, mostly biopsies, stained in routine clinical pathology practice with PD-L1 IHC standardized assays (22C3, 28-8 and SP263), were selected to represent various PD-L1 expression levels. Slides were digitalized and scored for the percentage of tumor cells with membranous staining by 161 pathologists using an online digital platform. A consensus score was defined for each case by a group 15 expert pathologists. Data regarding experience, training and practice of PD-L1 testing were also collected for each pathologist.

      Result

      Consensus score determined by the expert group highly correlated with the median of scores for each case (correlation coefficient=0.992). Overall concordance across pathologists was moderate, higher for the ≥50% cutoff (K=0.64) than the ≥1% cutoff (K=0.58). A higher concordance was achieved in the expert group (15 pathologists) as compared to the other pathologists (146 pathologists), in particular for the ≥1% cutoff. Concordance across pathologists correlated with training to PD-L1 scoring as well as the number of PD-L1 tests evaluated weekly. No correlation was found with the number of years of thoracic pathology practice or the type of pathology practice (private laboratory, community hospital, university hospital). The issues observed in the most discrepant cases were evaluated and described.

      Conclusion

      Concordance across pathologists for PD-L1 scoring in NSCLC was higher in the expert group of pathologists as compared to other pathologists, in particular for the ≥1% cutoff. Training to PD-L1 scoring and experience in routine pathology practice correlated with higher concordance. These data emphasize the importance of training to achieve a high concordance across pathologists in the real-world setting.