11568 Background: Soft tissue sarcoma (STS) are highly heterogenous in both histology and underlying genetic alterations, resulting in heterogenous outcomes of standard therapy. With recent stride of immunotherapies in oncology it is crucial to find the optimal match of the new therapies to disease subtypes. HRDScore and iTME parameters are two top candidate biomarkers to predict clinical outcomes in complex diseases such as STS. Methods: We prospectively profiled 102 Chinese STS patients who had received more than 1 line of prior treatment, mainly chemo-drugs. We then classified the patients according to their HRDScore (calculated from WES data using the Richardson method, ref. Telli M.L. et al. Clin. Cancer Res. 2016) and iTME parameters (calculated from RNAseq data using CIBERSORT method). Lastly, we tested the correlation of HRDScore or iTME with patient survival and clinic outcomes of anti-PD-1/L1 based combination therapy. Results: The histological subtypes of the patients included 32 LMS, 44 LPS, and 26 others (such as UPS, clear cell, myxofibrosarcoma, uterus LMS etc.). The top mutation genes (frequency > 5%) included TP53 29%, KMT2C 16%, NOTCH2 9%, ATRX 8%, NF1 8%, RB1 8% and PTEN 6%. The patients could be classified to HRD-high or HRD-low (HRDScore > or < = 42). Alternatively, unsupervised clustering of iTME revealed three patient subgroups (named iTME-I, -II, and -III). While all the three groups are characterized by a generally suppressed immune environment, iTME-I have high proportion of M0 macrophages and median M2 macrophages; iTME-II have high proportion of M2 macrophages but low M0 cells; and iTME-type III are low in both M0 and M2 cells. No significant M1 cells present in all the three iTME groups. Patient survivals were correlated with the iTME types but not HRDScore in Kaplan-Meyer analysis. The trend of survival time was iTME-III > iTME-II > iTME-I, with an HR = 0.156 for iTME-III over iTME-I (p = 0.008, log-rank test). Treatment response of anti-PD-1/L1 based combination therapies also showed a positive correlation to iTME-III, but not HRDScore although the small sample size prevented a definitive conclusion. Clinical evaluation of the 22 patients who received anti-PD-1/L1 therapy showed 1 PD (20%) and 4 SD/PR (80%) in iTME-III, 4 PD (40%) and 6 SD/PR (60%) in iTME-II, and 1 PD (33%) and 2 SD (67%) in iTME-I. Conclusions: iTME is a better biomarker than HRDScore in STS for survival and treatment outcomes. Differential Infiltration of M0 and M2 macrophages can distinguish patients with different survival and response to anti-PD-1/L1 combination therapy. The iTME subtypes may be used for treatment screen of combination immunotherapy but larger and randomized clinical studies are required to validate the discovery.