The RNA signature of immune-related genes and distinct spatial distribution of immune cells are closely associated with the treatment outcomes in the malignant melanoma patients receiving immune checkpoint inhibitors.

e21504 Background: The immune checkpoint inhibitors (ICIs), have improved survival with durable response for advanced malignant melanomas, but, only a minority of patients has achieved. To stay sway failure of triage patients, considerable side effects and costs of ICB therapies, the development of robust signatures predictive of response to ICB therapies is warranted. Methods: Clinical data were collected at Seoul St. Mary’s Hospitals. Among them, we examined RNA expression profile by QuantSeq 3’ mRNA-Sequencing and the composition, and spatial distance of immune cells by multiplex immunofluorescent staining. Results: The gene sets with relatively higher RNA expression in the good response group, were as follows; hematopoietic cell dedifferentiation gene, T-cell related adaptive immunity gene, IFN-related cytokine gene, and macrophage (TAM), NK cell, innate immunity gene set. Meanwhile, in the poor response group, inflammatory response, fibrosis response genes was increased.The cell densities of T cells, cytotoxic T cells, Treg cells, CD8+Treg cells, and exhausted T cells in the good response group tended to be higher, which was clearer in the peritumoral (PT) area than the intratumoral (IT) area. For the good response group, M1/M2 ratio in the IT area was significantly higher compared to that of poor response group (p = 0.045). According to the nearest neighbor analysis for % of melanoma cells within 15um distance present a close distance from M1, M2, and NK cells tended to be higher in the IT area than PT area. For the cases with higher spatial distribution from M1 in the IT area, PFS and OS was significantly extended (PFS, p = 0.0021; OS, p = 0.036). The gene sets with relatively higher mRNA expression in high neighborhood macrophage group, were cell death, apoptosis process gene set. In low neighboring M1 macrophage group, relatively higher mRNA expressions of extracellular component inflammation and fibrosis gene, cell migration, and cell de-differentiation gene set were observed. Conclusions: Taken together, our results suggest 4 key points. First, good and poor responder groups for ICB show marked differences from mRNA profiling to protein expression according to central dogma. Second, immune cells can be distributed sporadically in many places, but the actionable ground of the adaptive immune cells of the T cell family is the PT, and the playground of innate immune cells such as TAM and NK cells is the IT. In particular, the anti-tumor response of M1 macrophage and M2 macrophage was oppositely in IT. Third, high TAMs in IT are interpreted to have a tumoricidal effect because they show a stationery state while coding the killing effect gene set. In addition, topologically, spatial analysis as well as bulk genomics are important for predicting tumor response for ICIs.