Comparative study of two PD-L1 expression assays in patients with non-small cell lung cancer (NSCLC).

Abstract

e20581 Background: Response to anti–PD-L1/PD-1 therapy in patients with NSCLC has been associated with tumor PD-L1 expression. Avelumab is a fully human anti‒PD-L1 IgG1 antibody that has shown antitumor activity in advanced NSCLC, with a trend for greater efficacy in patients with PD-L1+ vs PD-L1– NSCLC tumors, as assessed using a proprietary PD-L1 assay in development (Dako, Carpinteria, CA; based on antibody clone 73-10, Merck KGaA, Darmstadt, Germany). This study compares the analytical performance of the 73-10 assay with the FDA-approved PD-L1 IHC 22C3 pharmDx diagnostic assay (Dako) in NSCLC tumors. Methods: Formalin-fixed paraffin-embedded NSCLC tumor samples were obtained from a commercial source and from a study of first-line (1L) avelumab in patients with advanced NSCLC (NCT01772004). Tumor membrane PD-L1 expression was assessed with the 73-10 and 22C3 assays by immunohistochemistry. Correlation between assays was determined at different PD-L1 cut-off levels. Results: In initial staining of commercial NSCLC samples, the 73-10 assay showed a broad dynamic range. Of 148 commercial samples, the 73-10 assay with ≥1%, ≥50% and ≥80% cut-offs classified 64.2%, 36.5% and 23.6% of samples as PD-L1+, respectively, whereas 20.3% were PD-L1+ with the 22C3 assay at a pre-specified ≥50% cut-off. The overall percentage agreement between assays was 83.8% using the ≥50% cut-off for both assays, and 93.9% using the ≥80% cut-off for 73-10 and ≥50% for 22C3. In follow-up studies, 83 study samples from the 1L NSCLC cohort were evaluable with both assays. With the 73-10 assay at ≥1%, ≥50%, and ≥80% cut-offs, 79.5%, 45.8%, and 31.3% of study tumors, respectively, were PD-L1+; with the 22C3 assay at ≥1% and ≥50% cut-offs, 59.0% and 21.7% were PD-L1+. Conclusions: Using a high tumor cell PD-L1 staining cut-off, the 73-10 and 22C3 assays showed highly concordant staining of NSCLC samples, with similar sensitivity observed with an ≥80% cut-off for 73-10 and ≥50% for 22C3. Using a low frequency of tumor PD-L1 expression, data suggested that the 73-10 assay had greater sensitivity than the 22C3 assay. These results provide a rationale for additional studies using the 73-10 assay at various cut-offs within the avelumab trial program. Clinical trial information: NCT01772004.