Purpose: A risk-adapted treatment strategy, the ISG/OS-2 trial, evaluated the use of mifamurtide in patients with P-glycoprotein (Pgp)-positive localized osteosarcoma. The primary objective was the identification of prognostic classifiers based on tumor microenvironment (TME) gene profiling in all patients and in those undergoing mifamurtide treatment.
Experimental design: RNA from pretreatment formalin-fixed, paraffin-embedded nondecalcified tissues of 62 patients was analyzed with the PanCancer Immune Profiling Panel (NanoString Technologies). Thirty-three (53%) of 62 Pgp-positive patients underwent chemotherapy + mifamurtide; 29 (47%) of 62 Pgp-negative patients received chemotherapy alone. Univariate Cox regression, ROC curve, and CIBERSORTx algorithm gene deconvolution analyses were performed.
Results: A 21-gene signature able to stratify all patients into high risk (Hi-R) and low risk (Lo-R) was identified: 5-year overall survival (OS) of 47% for Hi-R and 92% for Lo-R (P = 3e-06). TME of Lo-R compared with Hi-R was significantly enriched in CD8 T cells, regulatory T cells, and NK-activated cells and diminished in CD4 T cells. The 21-gene signature was validated in two independent sets: TARGET-OS The Cancer Genome Atlas (n = 62) and GSE33382 (n = 57). For patients treated with chemotherapy + mifamurtide, 31 mifamurtide-related genes able to distinguish Hi-R and Lo-R in terms of OS (P = 1e-09) and event-free survival (P = 3e-06) were also identified. After multivariate analysis, the 21- and 31-gene mifamurtide-related signatures were independently associated with OS (P = 0.00044 and P = 0.000079, respectively).
Conclusions: A validated osteosarcoma TME prognostic gene signature has been identified, regardless of mifamurtide treatment. Importantly, a mifamurtide-related signature was also developed. Tumor-immune interactions possibly implicated in disease progression and treatment response were shown.
©2025 American Association for Cancer Research.