Objective: Interest in immune therapies has exploded since the 2014 approval of first-generation programmed cell death 1 blocking antibodies for use in advanced melanoma. Clinical trials have focused primarily on histological material as the gold standard for evaluating programmed death ligand 1 (PD-L1) by immunoperoxidase (IPOX) studies. Studies validating the use of cytological specimens in the assessment of PD-L1 by IPOX staining are needed to optimise tissue utilisation in complementary diagnostic testing.
Methods: Twenty-three melanoma surgical biopsies (SBx) with an IPOX stain for PD-L1 clone 28-8, and a corresponding cytological specimen from the same patient, adequate for PD-L1 evaluation, were selected. Cell-transfer cell blocks (CBs) and conventional CBs were used to perform PD-L1 testing. Tumour proportion scores (TPS) were generated and the results were correlated with the corresponding SBx.
Results: Overall agreement (OA) using a ≥1% TPS cut-off for SBx compared to CB was 88.9%, positive percent agreement (PPA) was 87.5%, and negative percent agreement (NPA) was 100%, OA using a ≥5% TPS cut-off was 55.6%, PPA was 42.9%, and NPA was 100%. SBx compared to cell-transfer CB using a ≥1% TPS cut-off had an OA of 65.2%, a PPA of 55.6%, and a NPA of 100%, while a ≥5% TPS cut-off generated an OA of 52.2%, a PPA of 35.7%, and a NPA of 77.8%.
Conclusion: Our results demonstrate that cytological material, particularly conventional CB, is a viable alternative for evaluating PD-L1 in melanoma cases and suggest that a lower threshold (≥1%) may be beneficial when evaluating cytological material.
Keywords: cell block; cytopathology; immunocytochemistry; immunohistochemistry; melanoma; programmed death ligand 1.
© 2018 John Wiley & Sons Ltd.