Dynamic peripheral immune monitoring during checkpoint blockade: a framework for interpreting response, toxicity and resistance.

Immune checkpoint inhibitors (ICIs) have reshaped cancer treatment, but their clinical benefit remains variable. Many patients show primary or acquired resistance, and immune-related adverse events (irAEs) can limit treatment or complicate clinical decisions. Tissue-based biomarkers have improved patient selection in selected settings. These include programmed death-ligand 1 (PD-L1) expression, tumor mutational burden, microsatellite instability/mismatch repair deficiency, tumor-infiltrating lymphocytes and T-cell-inflamed gene signatures. Yet these markers are mostly static and tumor-centered. They do not fully capture host immune competence, systemic inflammation, treatment-induced immune changes or toxicity-prone immune activation. Peripheral immune biomarkers may help address this gap. Blood-based markers are repeatable and can be assessed across treatment, making them useful for longitudinal monitoring during checkpoint blockade. In this review, we discuss peripheral biomarkers as dynamic readouts rather than isolated predictors of response, toxicity or resistance. For narrative synthesis, we group these signals into four heuristic interpretive patterns developed for this review rather than established biological states or clinically validated patient categories: immune-cold resistance, effective anti-tumor activation, toxicity-prone inflammatory amplification and inflamed but ineffective resistance. We review several layers of peripheral monitoring, including immune-cell dynamics, T-cell receptor repertoire changes, cytokine and chemokine networks, humoral signals, eosinophilic inflammation and circulating tumor burden markers. We discuss how these readouts can be interpreted across key clinical windows. These include baseline, early on-treatment assessment, first imaging, irAE onset, progression and rechallenge. Peripheral monitoring should complement tissue biomarkers, imaging and clinical judgment. Its clinical translation will require standardized sampling, assay harmonization, integrated interpretation and prospective validation.