Abstract
The selection of patients for immunotherapy remains challenging given the lack of highly specific and highly sensitive biomarkers. Nevertheless, it is essential that testing laboratories are able to fulfil licencing criteria by providing the tests which have been validated as providing useful predictive information. Programmed cell death protein 1 (PD-1) expression assessment is now established in routine practice, although the situation regarding the selection of a particular assay remains complex, and testing protocols are likely to change in future. It is probable that programmed death-ligand 1 (PD-L1) expression assessment will be supplemented in the near future by tumour mutation burden (TMB), but this will require novel solutions to allow testing to be completed using small tissue samples and within narrow timeframes. While DNA mismatch repair (MMR) testing and CD8 T-cell density may also have a role to play in predicting immunotherapy response, their roles are not well-defined at present. Above all, the main challenge facing laboratories will be to perform this multitude of tests alongside the molecular markers already established in clinical practice [e.g., epidermal growth factor receptor (EGFR) mutation, anaplastic lymphoma kinase (ALK) translocation, ROS proto-oncogene 1 (ROS1) translocation]; the challenge for pathologists and clinicians will be to develop algorithms which will integrate the complex set of results from these tests and provide clinically-useful management regimens.