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)

  • 23245

    +

    P2.02-047 - Comparison of PD-L1 Immunohistochemistry Assays and Response to PD-L1 Inhibitors (ID 10278)

    09:30 - 09:30  |  Author(s): S. Aberrbock
    • Abstract

    Background:
    The FDA has approved different companion and complementary PD-L1 assays for selection of patients for different PD-L1 inhibitors. As a result, laboratories intending to implement FDA approved assays face significant operating and capital expenses. Several studies have demonstrated technical concordance between different PD-L1 assays, but their clinical validity in terms of the response to treatment has not entirely been explored. The aim of our study is to prospectively assess the staining performance of laboratory developed tests (LDT) for Ventana SP263 (nivolumab) and Dako 22C3 (pembrolizumab) clones on formalin fixed paraffin embedded samples, and to investigate the association between PD-L1 assays and response to PD-L1 inhibitors.
    
    Method:
    189 sequential lung tumor samples from patients with advanced NSCLC were prospectively stained with Ventana SP263 and Dako 22C3 clones. Both antibodies were optimized for use on the automated Ventana BenchMark ULTRA platform, and validated against corresponding FDA approved assays. Scoring algorithms for staining of the tumor cells approved by matched FDA assays were applied to all samples. Best overall response (BOR) for 43 patients treated with either nivolumab or pembrolizumab were assessed using RECIST v.1.1 and correlated with PD-L1 expression with SP263 and 22C3 using cut off points of 1%, 5% and 10%, 1-49% and 50%.
    
    Result:
    Ventana SP263 and Dako 22C3 LDTs showed good agreement with a concordance correlation coefficient of 0.86 (95% CI 0.82-0.90). Comparing the assays using the cutoffs of 1%, 5% or 10% for SP263 and the two cutoffs of 1% and 50% for 22C3 showed an association between the two assays as well (p<0.0001). There were no differences in BOR between pembrolizumab and nivolumab (p=0.12). For SP263, BOR was associated with a cut off point of 10% (Fisher’s p=0.030); while the 1% (Fisher’s p=0.09)and 5 % (Fisher’s p=0.054) cut off points were not associated with response. In contrast, for 22C3, BOR was associated with a cut off point of 1% (Fisher’s exact test p-value = 0.006), 5% (p 0.006) and 10% (0.006)
    
    Conclusion:
    Similar to FDA approved assays, Ventana SP263 and Dako 22C3 LDT, if properly validated, demonstrate good concordance, but are not interchangeable. Dako 22C3 is more sensitive assay and its expression shows better association with therapeutic response. Larger studies evaluating associations of alternative staining assays and a response to specific therapy are needed.