One immunotherapeutic agent for patients with advanced non-small cell lung carcinoma, pembrolizumab, has a companion immunohistochemistry (IHC)-based assay that predicts response by quantifying programmed death-ligand 1 (PD-L1) expression. The current study assessed the feasibility of quantifying PD-L1 expression using cytologic non-small cell lung carcinoma specimens and compared the results with those from small biopsy and surgical resection specimens.
PD-L1 expression was quantified using the IHC-based 22C3 pharmDx assay, with "positivity" defined as staining in ≥50% viable tumor cells; ≥ 100 tumor cells were required for test adequacy. For cytology specimens, IHC was performed on cell block sections.
A total of 214 specimens were collected from 188 patients, 206 of which (96%) were found to be adequately cellular, including 36 of 40 cytology (90%) and 69 of 72 small biopsy (96%) specimens. There was no significant difference noted with regard to the feasibility of PD-L1 IHC on small biopsy specimens compared with surgical resection specimens (P = .99), or between the percentage of PD-L1-positive cytology and histology (including surgical resection and histologic small biopsy) specimens (P = .083). PD-L1 expression was found to be concordant among samples from 21 of 23 patients from whom > 1 specimen was collected (91%). There also was no significant difference observed with regard to rates of PD-L1 positivity when comparing age, sex, diagnosis, and specimen site.
Quantification of PD-L1 expression is feasible on cytology specimens, and the results are comparable to those obtained from surgical resection and small biopsy specimens, including in matched specimens and using a single predictive IHC marker. Future studies will be necessary to determine the comparative value of other antibodies and their ability to predict response to immunotherapy. Cancer Cytopathol 2017;125:896-907. © 2017 American Cancer Society.