(P 396) COMPREHENSIVE WHOLE-GENOME MOLECULAR PROFILING IN BONE AND SOFT TISSUE SARCOMA: MULTI-OMICS ANALYSIS OF 627 CASES ACROSS HISTOLOGIC SUBTYPES

Objective: There exist no comprehensive studies based on whole-genome analysis for bone and soft tissue sarcomas, and genomic abnormalities in sarcomas are still largely unknown.

Methods: We included 627 bone and soft tissue sarcomas collected from the Japan Sarcoma Genome Consortium across subtypes which are classified into 2 categories: translocation-related sarcomas (TRS) with specific gene fusions and complex-karyotype sarcomas (CKS) without known fusion. Whole-genome sequencing, RNA sequencing, and methylation array analysis for frozen tumors and paired normal tissues were conducted.

Results: The number of somatic short nucleotide mutations (SNV) and structural variants (SV) was higher in CKS than in TRS. Somatic mutations in cancer-related genes, such as TP53, KMT2C, NF1, IDH1, RB1, and ATRX, were recurrently detected, most of which were more frequent in CKS than in TRS. Non-coding driver mutational analysis revealed frequent SNVs and SVs in the cis-regulatory element of TERT, including a known promoter mutation hotspot (NM_198253.2(TERT):c.-124C＞T). SVs located within 2kb upstream from the transcription start site lead to higher expression of TERT. SV clusters with copy number oscillating pattern which is known as chromothripsis were frequent in all the CKS subtypes (33.3 - 98.2%), while it was rarely seen in most of the TRS subtypes. Chromothripsis underlay 76.6% of focal amplification events in oncogenes and 27.7% of copy number loss in tumor suppressor genes, which were higher than the reported proportions in a study of pan-cancer analysis of whole genomes (PCAWG). We detected a varied range of telomere length (TL) not only among tumor subtypes but among samples of a specific subtype in high- and intermediate-grade sarcomas, whereas a little variation of TL was seen in low-grade and benign tumors. TL was longer in cases with loss-of-function aberrations in ATRX, while gain-of-function alterations in TERT were not associated with TL gain. A clustering analysis based on methylome data for CKS did not make isolated clusters according to the subtypes. An integrated analysis of methylome and expression analysis divided cases of CKS into 4 clusters.

Conclusion: We elucidated the molecular characteristics of bone and soft tissue sarcomas across histologic subtypes by a multi-omics study.