Background: Despite several second-line options are accessible for the treatment of advanced STS, there is a lack of predictive biomarkers available to support the rational selection of these drugs. Trabectedin specifically targets mononuclear cell lineage (macrophages and monocytes) that ultimately could inhibit tumor angiogenesis. Moreover, trabectedin seems to induce the expression of immune-checkpoint proteins (e.g. PD-L1); however, the predictive value of these factors remains unknown. We present the analyses of immune-checkpoint genes (CD274, CD86, CTLA4, HAVCR2, LAG3 and PDCD1) and CD163, CD4, CD68 and CD8A expression as potential predictive factors of response to trabectedin in a subset of STS patients of the GEIS registry. Methods: Selection criteria included patients with STS, pretreated with at least 2 lines in the advanced setting (one line being trabectedin), with paraffin block available and ethic committee’s approval. Direct transcriptomics was performed using HTG Molecular Oncology Biomarker Pathway panel (HTG Molecular Diagnostics, Inc.; Tucson, AZ, USA), following manufacturers’ instructions. Data analyses were performed taking into account the median Log2 of expression of each gene and by correlating it with progression-free survival (PFS) for trabectedin, and overall survival measured from the starting date of trabectedin treatment (OS). Results: Among 387 registered patients, fitting with the inclusions criteria, 139 cases were used for gene expression analyses, as the discovery set. Patients had median age of 52 years, 54% were females and had a median follow-up from diagnosis of 44 months. High expression of CD274 (PD-L1) was significantly associated with better PFS of trabectedin (5.4 vs. 3.0 months; p= 0.006). Similar results were obtained with high expression of CTLA4 and LAG3: 6.0 vs 3.1 months; p = 0.005 and 5.4 vs 2.7 months; p = 0.042, respectively. Expression of CTLA4 and LAG3 showed no significant impact in OS; whereas CD274 high expression showed a trend towards better OS (17.9 vs 10.2 months; p = 0.077). Also, no significant correlation was achieved for CD163, CD4, CD68, CD8A, CD86 and HAVCR2; PDCD1 (PD-1) showed a trend towards better PFS of trabectedin, p = 0.114. Conclusions: The expression of selected immune-checkpoint genes exhibited a potential predictive value for trabectedin in advanced STS. Validation studies (at the transcriptional and protein level) are currently ongoing to confirm their potential predictive role.