Multidimensional Single-Cell Transcriptomic Profiling of Uterine Leiomyosarcomas Identifies Molecular Subtypes with Distinct Therapeutic Vulnerabilities

Uterine leiomyosarcoma (ULMS) is an orphan disease that frequently recurs and metastasizes, with patients undergoing multiple lines of chemotherapy due to lack of effective therapeutic targets. To address this gap, we used single-cell RNA sequencing and spatial transcriptomic analysis to comprehensively profile ULMS. We uncovered eight subtypes of tumor cells, including stem cell-like hormone receptor-positive cells, tumor cells with mesenchyme-like features, ischemic tumor cells defined by a MYC program, and inflammatory tumor cells with active interferon signaling. The spatial correlates of these tumor cell subtypes demonstrated unique localization patterns. By correlating these signatures to bulk RNA sequencing data, we demonstrate the scalability of these findings to clinical outcomes. Finally, using the single-cell integration and drug response prediction algorithm, we derive previously undescribed drug predictions targeting specific tumor subtypes that may be more efficacious than existing adjuvant regimens. Our findings highlight new individualized and multifaceted therapeutic avenues to treat ULMS.