PD-L1 expression in NSCLC: Role of cell blocks and concordance between samples

A. Ambrosini-Spaltro, A. Dubini, F. Pieri, C. Ravaglia, A. Delmonte, V. Poletti

Research output: Contribution to journalArticlepeer-review


Background: PD-L1 immunohistochemistry is currently performed in patients with advanced non-small cell lung carcinoma (NSCLC) to identify responders to immune checkpoint inhibitors. Cell blocks from fine needle aspiration of NSCLC are frequently used for diagnostic purposes. The aims of the study are to analyze: (a) the distribution of PD-L1 in cell blocks, in comparison to biopsies and surgical specimens; (b) the concordance of PD-L1 in specimens of the same patients. Methods: PD-L1 analyses conducted in NSCLCs were retrieved. Cell blocks were prepared with the self-clotting method. PD-L1 was performed with Dako 22C3 on the Ventana BenchMark ULTRA platform. Results were divided by tumor proportion score (TPS) in 3 categories: <1%; 1% to 49%; ≥50%. Results: A total of 483 samples from 456 patients was collected: 120 cell blocks (24.8%), 307 endoscopic or transthoracic biopsies (63.6%), 56 surgical specimens (11.6%). TPS was: <1% in 230 samples (47.8%), 1% to -49% in 136 (28.3%) and ≥ 50% in 115 (23.9%); in two samples material was insufficient. Statistics did not reveal significant differences in PD-L1 expression among the various materials (χ2 = 2.905; P =.574). In 50 samples from 25 patients, PD-L1 was carried out in two samples of the same patients, with moderate agreement (concordance rate: 68.0%, k = 0.469). Conclusion: (a) PD-L1 is similarly distributed in the different materials; (b) PD-L1 shows moderate concordance in different samples of the same patients. PD-L1 may be routinely tested in cell blocks, but interpreted with caution and repeated whenever possible.

Original languageItalian
Pages (from-to)303-310
Number of pages8
JournalDiagnostic Cytopathology
Issue number2
Publication statusPublished - 2021

Cite this