The clinical development of checkpoint inhibitor-based immunotherapy has revolutionized cancer treatment, with durable responses observed in melanoma and other malignancies. While PD-1 and PD-L1 inhibitors are generally well-tolerated, combination therapies increase the risk of immune-related adverse events. Predictive biomarkers are crucial for optimizing patient benefit, minimizing toxicity, and guiding treatment combinations. PD-L1 expression is the most studied biomarker, but it is insufficient for most cancers. Emerging biomarkers include tumor-infiltrating lymphocytes, mutational burden, immune gene signatures, and multiplex immunohistochemistry. Future biomarkers will integrate multiple approaches to characterize the immune tumor microenvironment. PD-L1 expression on tumor cells is a logical biomarker for predicting response to anti-PD-1 or anti-PD-L1 therapies. Initial studies showed that PD-L1 positivity correlated with clinical benefit, but its predictive value is suboptimal. PD-L1 expression is regulated by various mechanisms, including genetic and epigenetic factors, and can be transient or heterogeneous. Poor reliability of PD-L1 immunohistochemistry as a biomarker is due to variable antibody use and thresholds. Despite these limitations, PD-L1 testing helps stratify patients in clinical trials. Tumor-infiltrating lymphocytes (TILs) are associated with improved survival in various cancers. Higher TIL density is linked to better clinical outcomes, but baseline TIL density alone is not sufficient. T-cell receptor clonality is also a potential biomarker, with restricted clonality associated with response to anti-PD-1 therapy. Mutational or neoantigen burden is another promising biomarker, with higher mutational burden correlating with better responses to immunotherapy. However, most neoantigens are patient-specific, limiting their utility. Peripheral blood markers, such as lymphocyte counts and regulatory T-cell populations, have been associated with clinical benefit but lack validation as predictive biomarkers. Immune gene signatures, particularly those related to interferon gamma, show promise as predictive biomarkers. Multiplex immunohistochemistry provides detailed information on immune cell phenotypes and spatial relationships, offering complementary data to gene expression profiling. Combined biomarker strategies, integrating multiple factors such as PD-L1 expression, TIL density, mutational burden, and immune gene signatures, may improve prediction of response to immunotherapy. These strategies are essential for optimizing treatment selection and minimizing toxicity. Future research should focus on developing integrated biomarkers that account for the complexity of the tumor microenvironment and individual patient variability.The clinical development of checkpoint inhibitor-based immunotherapy has revolutionized cancer treatment, with durable responses observed in melanoma and other malignancies. While PD-1 and PD-L1 inhibitors are generally well-tolerated, combination therapies increase the risk of immune-related adverse events. Predictive biomarkers are crucial for optimizing patient benefit, minimizing toxicity, and guiding treatment combinations. PD-L1 expression is the most studied biomarker, but it is insufficient for most cancers. Emerging biomarkers include tumor-infiltrating lymphocytes, mutational burden, immune gene signatures, and multiplex immunohistochemistry. Future biomarkers will integrate multiple approaches to characterize the immune tumor microenvironment. PD-L1 expression on tumor cells is a logical biomarker for predicting response to anti-PD-1 or anti-PD-L1 therapies. Initial studies showed that PD-L1 positivity correlated with clinical benefit, but its predictive value is suboptimal. PD-L1 expression is regulated by various mechanisms, including genetic and epigenetic factors, and can be transient or heterogeneous. Poor reliability of PD-L1 immunohistochemistry as a biomarker is due to variable antibody use and thresholds. Despite these limitations, PD-L1 testing helps stratify patients in clinical trials. Tumor-infiltrating lymphocytes (TILs) are associated with improved survival in various cancers. Higher TIL density is linked to better clinical outcomes, but baseline TIL density alone is not sufficient. T-cell receptor clonality is also a potential biomarker, with restricted clonality associated with response to anti-PD-1 therapy. Mutational or neoantigen burden is another promising biomarker, with higher mutational burden correlating with better responses to immunotherapy. However, most neoantigens are patient-specific, limiting their utility. Peripheral blood markers, such as lymphocyte counts and regulatory T-cell populations, have been associated with clinical benefit but lack validation as predictive biomarkers. Immune gene signatures, particularly those related to interferon gamma, show promise as predictive biomarkers. Multiplex immunohistochemistry provides detailed information on immune cell phenotypes and spatial relationships, offering complementary data to gene expression profiling. Combined biomarker strategies, integrating multiple factors such as PD-L1 expression, TIL density, mutational burden, and immune gene signatures, may improve prediction of response to immunotherapy. These strategies are essential for optimizing treatment selection and minimizing toxicity. Future research should focus on developing integrated biomarkers that account for the complexity of the tumor microenvironment and individual patient variability.