Inhibitory receptor states, immune homeostasis and the benefit-risk boundary of cancer checkpoint blockade

Immune checkpoint blockade has transformed cancer treatment, but its clinical success depends on releasing antitumor immunity without collapsing peripheral tolerance. Biomarkers such as PD-L1 expression, mismatch-repair deficiency and tumor mutational burden capture only selected aspects of this balance and provide limited guidance on immune-related adverse events (irAEs). This Review examines inhibitory receptor states as dynamic readouts of the therapeutic window in cancer immunotherapy. Rather than treating PD-1, CTLA-4, LAG-3, TIM-3, TIGIT, VISTA and NKG2A as isolated abundance markers, we discuss how their signaling intersects with CD3, CD28 and cytokine-driven activation, immune-cell differentiation, tissue localization and homeostatic restraint. We further consider how checkpoint blockade converts these circuits into either tumor control or organ-specific immune injury. Evidence from tissue pathology, blood immunomonitoring, single-cell and spatial multi-omics, routine laboratory markers, imaging and emerging host-immune surrogates is integrated into a longitudinal, regimen-specific framework. We argue that clinically useful models should jointly estimate response and toxicity risk, distinguishing reinvigoratable antitumor states from fixed dysfunction and autoreactive vulnerability. Such an approach could support patient selection, regimen tailoring, early toxicity surveillance and more precise use of combination immunotherapy.