Despite the success of immunotherapy in a number of different malignancies, biomarkers to predict patient response to treatment are not clearly defined. There remains a clear need for biomarkers that identify those patients most likely to achieve a clinical response and/or be at increased risk of immune-related adverse effects. The natural history of cancer involves interactions between the tumour and the immune system and immune infiltration at the tumour site and the tumour microenvironment may be indicative of host response. This can have a predictive role, with clinical activity of ipilimumab shown to be correlated with high baseline expression of the immune-related genes FoxP3 and indoleamine 2,3-dioxygenase and an increase in tumour-infiltrating lymphocytes. Several peripheral blood biomarkers have also been shown to correlate with improved response to immunotherapy, including absolute monocyte counts, lactate dehydrogenase (LDH), myeloid-derived suppressor cells (MDSCS), and circulating T-regulatory cells. Other potential biomarker approaches are emerging, including the immunoscore, immunoprofiling and immune gene signatures. For anti-PD-1 therapies, the focus has been the expression of anti-PD-L1 on tumour cells. Although an association between high PD-L1 expression and clinical response has been widely reported, the use of PD-L1 expression to select patients for anti-PD-1 antibody treatment is problematic, since clinical activity is also observed in patients with low or negative PD-L1-expressing tumours. Further research into biomarker strategies is needed before their use can translate into widespread adoption in clinical practice.