S. Foersch (Mainz, Germany)
Johannes Gutenberg University MainzAuthor Of 1 Presentation
13P - Comparison study of different programmed death-ligand 1 (PD-L1) assays, readers and scoring methods in triple-negative breast cancer (TNBC)
Abstract
Background
Different immunohistochemical programmed death-ligand 1 (PD-L1) assays and scoring methods in triple-negative breast cancer (TNBC) have been reported to yield variable results. We compared the analytical concordance and interobserver variability of four clinically developed PD-L1 assays assessing immune cell (IC) score and combined positive score (CPS) in TNBC.
Methods
Archival primary TNBC resection specimens (n = 99) were stained for PD-L1 using VENTANA SP142, VENTANA SP263, DAKO 22C3 and DAKO 28-8. PD-L1 expression was scored by four trained readers according to guidelines for IC-score and CPS on whole slide images by virtual microscopy.
Results
The mean PD-L1 positivity ranged between 53%-74% for IC-score ≥1% and CPS ≥1 with highest levels for SP263. The concordance between IC-score ≥1% and CPS ≥1 across all readers for each assay was >93%. Intra-class correlation coefficients (ICCs) revealed poor-to-good inter-reader agreement for each assay for IC-score (0.489-0.793, highest value for SP142) and moderate-to-good agreement for CPS (0.653-0.794, highest value for 22C3). ICCs for each reader (ranging from 0.226-0.595) uncovered poor-to-moderate inter-assay agreements on PD-L1-positivity for both scores. The concordance of all readers across the four assays was 86.8% for CPS ≥1 and 88.4% for IC-score ≥1%. Kappa scores for inter-reader agreement for each assay at IC-score ≥1% were 0.728-0.777 and for CPS ≥1 0.680-0.735. Evaluation of inter-assay agreement for each reader revealed kappa scores 0.446-0.596 at IC-score ≥1% and 0.492-0.587 at CPS ≥1.
Conclusions
We demonstrate a certain degree of concordance between IC-score and CPS for the assessment of PD-L1-positivity across different assays and readers. However, the four PD-L1 assays are analytically not fully concordant.
Legal entity responsible for the study
The authors.
Funding
MSD.
Disclosure
A. Noske: Travel/Accommodation/Expenses: Roche Pharma AG; Advisory/Consultancy: Myriad. D-C. Wagner: Travel/Accommodation/Expenses: Roche Pharma AG; Research grant/Funding (institution): MSD. K. Schwamborn: Research grant/Funding (institution): Roche; Research grant/Funding (institution): MSD. S. Foersch: Research grant/Funding (institution): MSD. K. Steiger: Research grant/Funding (institution): Roche Pharma AG; Research grant/Funding (institution): MSD. M. Kiechle: Honoraria (self): Celgene; Honoraria (self): AstraZeneca; Honoraria (self): Myriad Genetics; Honoraria (self): Teva; Shareholder/Stockholder/Stock options: Therawis Diagnostics GmbH; Shareholder/Stockholder/Stock options: Busenfreundin GmbH. D. Oettler: Full/Part-time employment: MSD. A. Hapfelmeier: Honoraria (institution): MSD. W. Roth: Honoraria (self): Roche Pharma AG; Honoraria (self), Research grant/Funding (institution): MSD. W. Weichert: Advisory/Consultancy, Research grant/Funding (institution): Roche; Advisory/Consultancy, Research grant/Funding (institution): BMS; Advisory/Consultancy, Research grant/Funding (institution): MSD; Advisory/Consultancy: AstraZeneca; Advisory/Consultancy: Pfizer; Advisory/Consultancy: Novartis. All other authors have declared no conflicts of interest.