Related Experiment Video
Updated: Aug 13, 2025

Machine Learning Algorithms for Early Detection of Bone Metastases in an Experimental Rat Model
Published on: August 16, 2020
Development of personalized classifier based on metastasis and the immune microenvironment to predict the prognosis
Zhifeng Zhang1, Jianyong Yuan2, Yi Wang1
1Department of Orthopaedic Surgery, Third Affiliated Hospital of Naval Military Medical University, Shanghai, China.
Background:
Osteosarcoma is a common malignant bone tumor with a poor prognosis. The progression and metastasis of osteosarcoma are significantly influenced by the tumor microenvironment (TME). This study aimed to develop a personalized classifier based on metastasis and immune cells in the TME to achieve better prognostic prediction in osteosarcoma.
Methods:
Firstly, osteosarcoma metastasis-related differentially expressed genes (DEGs) and infiltrating immune cells in the TME were analyzed using a series of bioinformatics methods. The metastasis-related gene signature (MRS) and TME score of osteosarcoma patients were then developed. On this basis, a personalized MRS-TME classifier was constructed and validated in other clinical cohorts and different subgroups. In addition, the relationship between the MRS-related genes and the immune microenvironment was also clarified. Finally, the signaling pathways and immune response genes in osteosarcoma patients among different MRS-TME subgroups were analyzed to explore the underlying molecular mechanism.
Results:
We first identified the metastasis-related DEGs in osteosarcoma, which were primarily involved in the muscle system process, calcium ion homeostasis, cell chemotaxis, and leukocyte migration. A personalized MRS-TME classifier was then constructed by integrating the MRS (10 genes) and TME (six immune cells) scores. The MRS-TME classifier demonstrated a potent capacity of predicting the survival prognosis in diverse osteosarcoma cohorts as well as in the clinical feature subgroups. The MRS score was negatively associated with the TME score, and patients in the MRSlow/TMEhigh subgroup exhibited a better prognosis compared to all other subgroups. Significant differences existed between the cellular signaling pathways and immune response profiles among the different MRS-TME subgroups, especially in relation to the metabolism-related biological processes and the inflammatory response.
Conclusions:
The MRS-TME classifier might be a beneficial tool to aid in the prognostic evaluation and risk stratification of osteosarcoma patients.
More Related Videos
07:44Practical Considerations in Studying Metastatic Lung Colonization in Osteosarcoma Using the Pulmonary Metastasis Assay
Published on: March 12, 2018
11:15A Preclinical Mouse Model of Osteosarcoma to Define the Extracellular Vesicle-mediated Communication Between Tumor and Mesenchymal Stem Cells
Published on: May 6, 2018