Multi-Omics Tumor Immunogenicity Score Predicts Immunotherapy Outcome and Survival

Area of Science:

• Oncology
• Immunology
• Genomics
• Transcriptomics

Background:

• Tumor immunogenicity is key for predicting response to immunotherapies like immune checkpoint inhibitors (ICIs).
• Current single biomarkers (e.g., tumor mutation burden, PDL1 expression) have limited predictive power for patient survival.
• Machine learning models integrating multiple biomarkers show promise but often lack interpretability.

Purpose of the Study:

• To develop and validate a novel Multi-Omics Tumor Immunogenicity score (MOTIscore) for predicting ICI response.
• To assess the generalizability of MOTIscore across diverse cancer types.
• To compare MOTIscore's performance against existing biomarkers and machine learning models.

Main Methods:

• Integrated multiple immunogenicity biomarkers (genomic and transcriptomic) using a weighted sum scoring scheme.
• Determined biomarker weights using statistical analysis in a melanoma ICI cohort.
• Compared MOTIscore performance against a machine learning model and tumor mutation burden in melanoma, gastric cancer, and pan-cancer datasets.

Main Results:

• MOTIscore demonstrated predictive performance comparable to a machine learning model and superior to tumor mutation burden for ICI outcomes in melanoma and gastric cancer.
• High MOTIscores were significantly associated with extended overall and progression-free survival in melanoma and gastric cancer patients.
• Analysis revealed enrichment of immune-related pathways and identified C-X-C motif chemokine ligands as key indicators of successful ICI therapy and improved survival.

Conclusions:

• MOTIscore offers improved prediction of ICI outcomes compared to single-omic biomarkers.
• Elevated tumor immunogenicity, as measured by MOTIscore, is linked to better survival in specific cancers.
• MOTIscore shows potential for guiding personalized ICI treatment strategies, warranting further investigation in prospective studies.