Cell proliferation, immune cell infiltration, and survival outcomes in sarcomas with high cancer-associated fibroblast expression.

O Olivia Martin (1Division of Hematology, Department of Internal Medicine, The Ohio State University, Columbus, OH) K Kei Kawashima (Roswell Park Comprehensive Cancer Center, Buffalo, NY) M Masanori Oshi (Yokohama City University Hospital, Yokohama, Japan) R Rongrong Wu K Konstantinos Chouliaras (Baptist MD Anderson Cancer Center, Jacksonville, FL) A Akimitsu Yamada K Kazutaka Narui I Itaru Endo K Kazuaki Takabe A Ankit Patel (Department of Computer Science, Rice University)

Abstract

11541 Background: Cancer-associated fibroblasts (CAFs) play an integral role in the tumor microenvironment and have been linked to tumor aggravation and promotion of metastatic potential. CAFs are often abundant in sarcomas, but their role is understudied partly because both share a similar mesenchymal origin. We aimed to investigate the relationship between CAFs and clinical characteristics in sarcomas using transcriptomics. Methods: Genomic and clinicopathologic data was acquired from The Cancer Genome Atlas Sarcoma cohort. Cases were divided into high and low CAF groups using the median CAF level. Immune cell compositions including CAFs were determined using the xCell algorithm. Institutional Review Board waiver was deemed applicable as the data is de-identified from publicly available databases. Results: A total of 258 patients including leiomyosarcoma (LMS, n=83), myxofibrosarcoma/undifferentiated pleomorphic sarcoma (MFS/UPS, n=80), dedifferentiated liposarcoma (DDLPS, n=46), and other pathologies (n=20), were analyzed. DDLPS had significantly greater CAFs than the other subtypes (p=0.0025). CAF expression varied significantly by tumor grade, with grade 3 tumors having the lowest CAF expression and grade 1 & 2 tumors having significantly higher CAF expression (p<0.001). Ki67 was significantly lower in the high CAF tumors (p<0.001). On GSEA, high CAF sarcomas showed significantly less enrichment in myc targets v1 (NES= -2.67, FDR<0.001), myc targets v2 (NES=-2.25, FDR<0.001), G2M checkpoint (NES=2.39, NES<0.001), E2F targets (NES=-2.66, FDR<0.001), mitotic spindle (NES= -1.33, FDR= 0.015), and mTORC1 signaling (NES=-2.33, FDR <0.001); all consistently showing that high CAF sarcomas are less proliferative. In terms of genomic instability, high CAF sarcomas were significantly associated with less homologous recombination deficiency (p<0.001), intratumor heterogeneity (p<0.001), aneuploidy (p=0.003), and silent mutation rates (p=0.036). Analysis of immune cell composition revealed significantly less infiltration of B-cells (p=0.004), CD8 T-cells (p=0.013), Th1 cells (p<0.001), Th2 cells (p=0.008), and macrophages (p=0.027) in the high CAF group. Differential gene expression analysis demonstrated that C7, SESN1, FAM198A, were the most, and WNT7B, ENO1, TM4SF19 were the least expressed among 20511 genes analyzed in high CAF sarcomas. High CAF sarcoma were associated with significantly better disease-free survival (HR= 0.48 (95% CI 0.34-0.68), p <0.001) as well as overall survival compared with low CAF sarcomas (HR= 0.5 (95% CI 0.34-0.76), p< 0.001). Conclusions: CAFs are associated with less cell proliferation, less immune cell infiltration, and better survival. Further studies are warranted to develop CAFs as a prognostic biomarker for sarcoma patients.

Article Details

Volume / Issue Vol. 44, Issue 16_suppl
Published June 01, 2026
Pages 11541-11541
ISSN 0732-183X
Publisher Lippincott Williams & Wilkins

Journal Info

Journal of Clinical Oncology

Lippincott Williams & Wilkins

ISSN: 0732-183X Health Sciences

Authors (10)

O

Olivia Martin

1Division of Hematology, Department of Internal Medicine, The Ohio State University, Columbus, OH

K

Kei Kawashima

Roswell Park Comprehensive Cancer Center, Buffalo, NY

M

Masanori Oshi

Yokohama City University Hospital, Yokohama, Japan

R

Rongrong Wu

K

Konstantinos Chouliaras

Baptist MD Anderson Cancer Center, Jacksonville, FL

A

Akimitsu Yamada

K

Kazutaka Narui

I

Itaru Endo

K

Kazuaki Takabe

A

Ankit Patel

Department of Computer Science, Rice University