(P 057) SARCOMA MICROENVIRONMENT CHARACTERIZATION: CLINICOPATHOLOGICAL ANALYSIS OF ONE-HUNDRED TWENTY-SIX CASES

Objective: Sarcomas are rare and heterogeneous tumors. They comprise more than 80 histotypes, with different clinical behaviors. Given their rarity and heterogeneity, they are often under-recognized by clinicians and diagnosed at advanced stages as locally aggressive large masses and/or with distant metastasis. Complete surgical resection remains the only potentially curative treatment, and current available adjuvant therapies do not substantially improve survival. The interplay between tumor and its microenvironment influences cancer progression and sensitivity to therapies. The aim of this study was to characterize the interplay between sarcoma and its microenvironment, focusing on the immune infiltrate and antigen presenting machinery (APM) molecules of different sarcoma histotypes.

Methods: We selected 8 sarcoma histotypes: chordoma (CH), chondrosarcoma (CHS), gastrointestinal stromal tumor (GIST), retroperitoneal dedifferentiated liposarcoma (RP-DDLPS), leiomyosarcoma (LMS), myxoid liposarcoma (M-LPS), myxofibrosarcoma (MFS), undifferentiated pleomorphic sarcoma (UPS). Primary, non-metastatic sarcomas were included. Metastasis and local recurrencies were excluded. Tissues from patients (15 for each histotype and 21 for GIST) who underwent surgery at Humanitas Research Hospital between 2001 and 2021 were assembled in tissue microarray (TMA). To account for heterogeneity, we sampled for each patient 11 cores representing areas with different characteristics: higher grade, lower grade, high lymphocytic infiltrate, neoplastic areas abutting the non-neoplastic tissue and vice versa, in order to represent tumor heterogenicity. Clinical and pathological variables of the patients were retrieved from a prospectively collected database. We performed a comprehensive analysis of the immune contexture including lymphocytic infiltrate (T and B cells), innate immune system (neutrophils and macrophages), and APM components, exploiting immunohistochemistry (IHC) and digital image analysis. Data derived from clinics, pathology and IHC were analyzed with probabilistic programming (Bayesian data analysis).

Results: In total, specimen from 126 sarcomas were available for evaluation. Patient and tumor characteristics are described in Table 1. At the time of last follow-up 59.5% of patients were alive with no evidence of disease (NED), 10.4% were alive with disease and 30.1% were dead, of which 9.5% died for primary disease, 7.9% died for DM, 6.3% died for LR and 6.3% died for other causes. Thirty-one patients (24.6%) experience local recurrence (LR), and 28 patients (22.2%) had distant metastasis (DM).
UPS was the most immunogenic neoplasm, whereas CHS had the lowest immune infiltrate. Macrophages were the most abundant leucocytic subtype. Neutrophils and cytotoxic T cells were scarce, but interestingly they both showed increased levels in CHS. Considering different TMA cores, we detected increased immune infiltrate in neoplastic areas with respect to the non-neoplastic tissue abutting the tumor (Figure 1).
Regarding APM, we observed the absence of some components, which determines a defect in APM function. TAP2 was the most defective molecule, followed by the transporter LMP10. Interestingly, defects of APM were especially concentrated in high grade regions of the tumors with respect to the other areas (Figure 2).
Moreover, considering every patient separately (figures not shown), there were some cases in which APM was defective across all the tumoral areas, suggesting in these cases immune-escape through APM disruption to be an early event in tumorigenesis.

Conclusion: Our results revealed the heterogenicity of the immune microenvironment among different sarcoma histotypes (inter-tumor heterogeneity) and different tumoral areas (intra-tumor heterogeneity). These results may represent the foundation for further characterization of the relationship between sarcoma and its microenvironment, in order to understand its prognostic and predictive role, and for modelling this relationship to provide insights on possible actionable mechanisms.