840 Unsupervised learning reveals different degrees of heterogeneity as well as cell involvement in cutaneous lupus erythematous antimalarial treatment response subgroups

First-line treatment for cutaneous lupus erythematosus involves the use of antimalarials. Treatment response is highly variable with some patients responding well to hydroxychloroquine (HCQ), necessitating addition quinacrine (QC) and others refractory from both (NR). Here we used unsupervised learning highlight differences between immune infiltrate composition three groups. Infiltrate was characterized using imaging mass cytometry, where stained biopsies (12 HCQ, 11 QC, 20 NR) two separate panels 37-metal conjugated antibodies, which enabled measurement different cell types, cytokines pathway proteins within tissue sample. Correlation matrices cytokines/pathways were constructed each groups visualize correlation strength all markers simultaneously. Hierarchical clustering performed on top matrices, revealed a single large cluster in medium clusters QC many NR. This indicates relatively homogenous cytokine/pathway activity mechanisms significant heterogeneity Cell measurements scaled type then subjected K-means clustering. Clustering results visualized heatmaps demonstrated heavy CD4/CD8 dominance noteworthy subgroups, one being dominated by CD8 cells (61% vs. 7%, p=0.1) other cDC (62% 15%, p=0.1), NR subgroups CD8-dominant (48% 9%, p<0.00001), cDC-dominant (38% 10%, p<0.001) CD14+CD16+ dominant (45% 4%, p<0.00001). analysis repeated SCLE DLE clinical subtypes remarkably similar population