Changes in the use of commercial and lab-developed IVD assays in the clinical setting for PD-L1 expression testing in NSCLC in the United States.

Abstract

154 Background: In 2016 we observed a lack of standardization in the use of cut-off points to define positivity when testing for PD-L1 expression in NSCLC, despite these being specified by assay manufacturers or recommended based on trial data. One year on we look at how clinical practice has changed in light of new approvals for PD-(L)1 inhibitors, availability of new IVD assays and changes in clinical practice guidelines recommending the use of immunotherapy for stage IV NSCLC. Here we explore how the variety of test brands and cut-off points used in the US has changed since 2016 by examining real-world clinical usage data. Methods: Between June and August 2016 and June and August 2017, a panel of pathologists in the US (n = 21 in 2016 and n = 28 in 2017) was asked to report on their practices relating to PD-L1 expression testing in NSCLC, through the submission of online de-identified record forms (n = 167 and n = 224 PD-L1-tested samples in 2016 and 2017 respectively). Results: Of the 224 samples gathered in 2017, 187 (84%) were tested with the Dako 22C3 pharmDx assay (vs 67% in 2016), 16 (7%) with the Dako 28-8 pharmDx assay (vs 22% in 2016) and 11 (5%) with a lab-developed test (LDT). An increase in the use of 1% staining as the cut-off was observed for both 22C3 and 28-8 pharmDx. The full distribution of cut-offs used is shown in the table below. Conclusions: Following initial fragmentation of clinical practices in 2016, PD-L1 expression testing has seen consolidation towards greater use of the Dako 22C3 assay and higher conformity in testing at the recommended cut-off points. While greater standardization simplifies testing, the choice of assay has potential implications on subsequent treatment: current PD-L1 assays allow physicians to confirm whether a specific PD-(L)1 inhibitor is appropriate for a patient, but there is no single PD-L1 expression test that supports oncologists in making treatment decisions for the PD-(L)1 inhibitor class as a whole.[Table: see text]