scResponse: A Rank-Based Method for Identifying Cell States That Contribute to Immunotherapy Response by Single-Cell Data

Immunotherapy has shown great promise in cancer treatment, yet many patients fail to achieve long-term efficacy. Single-cell sequencing is a powerful technique for understanding how cellular diversity affects treatment outcomes; however, there remains a lack of robust methods for evaluating immunotherapy efficacy at single-cell resolution. In this study, we present scResponse, an efficient algorithm that quantifies single-cell responses to immunotherapy. scResponse captures subtle influences of diverse cellular subtypes and states on immunotherapy efficacy, including the effects of vascular normalization states of vascular cells, polarization of macrophages, terminal exhaustion of CD8 + T cells, and the opposing impacts from distinct fibroblast subtypes. By linking therapeutic response to single‑cell states, scResponse enables to delineate how distinct biological processes shape cellular states to drive divergent therapeutic outcomes. Integrated analysis across cancers identified key metabolic pathways associated with responsive states of tumor cells, including glyoxylate and dicarboxylate metabolism as immunotherapy-sensitizing and retinol metabolism as suppressing it. As a proof-of-concept, the opposing roles of glyoxylate and retinol metabolism was validated using tumor‑CD8 + T cell co-culture assays and in vivo mouse models. Thus, scResponse is an effective tool for mapping immunotherapy efficacy to single-cell states, enabling mechanistic insights and target discovery for developing novel immunotherapy sensitization strategies.