Machine learning survival prediction using tumor lipid metabolism genes for osteosarcoma

Area of Science:

• Oncology
• Molecular Biology
• Metabolomics

Background:

• Osteosarcoma is a primary bone cancer with poor prognosis, influenced by tumor heterogeneity.
• Lipid metabolism is increasingly recognized as a key driver in cancer progression and heterogeneity.
• Understanding lipid metabolism subtypes is crucial for developing targeted therapies and improving patient outcomes.

Purpose of the Study:

• To identify novel molecular subtypes of osteosarcoma based on lipid metabolism.
• To develop a predictive signature for osteosarcoma patient survival using lipid metabolism-related genes.
• To evaluate the predictive performance of the developed signature against existing prognostic models.

Main Methods:

• Consensus clustering of four independent cohorts (TARGET-OS, GSE21257, GSE39058, GSE16091) to define molecular subtypes.
• Differential gene expression analysis, univariate Cox analysis, and StepAIC for biomarker identification.
• Machine learning algorithms were employed to construct and validate a Lipid Metabolism-Related Signature (LMRS).

Main Results:

• Two distinct lipid metabolism subtypes (C1 and C2) were identified, showing significantly different survival rates.
• C1 subtype exhibits increased cholesterol, fatty acid synthesis, and ketone metabolism.
• C2 subtype is characterized by steroid hormone biosynthesis, arachidonic acid, and glycerolipid/linoleic acid metabolism.
• A 12-gene LMRS was developed, demonstrating robust and superior predictive accuracy for osteosarcoma prognosis across multiple cohorts.
• The LMRS outperformed 12 previously published prognostic signatures.

Conclusions:

• The study successfully delineated two novel lipid metabolism-driven molecular subtypes in osteosarcoma.
• The developed 12-gene LMRS provides a reliable and accurate tool for predicting osteosarcoma patient survival.
• This signature has the potential to significantly improve clinical management and patient outcomes in osteosarcoma.