Integration of an OS-Based Machine Learning Score (AS Score) and Immunoscore as Ancillary Tools for Predicting
Isidro Machado1,2,3,4, Raquel López-Reig5,6, Eduardo Giner7
1Pathology Department, Instituto Valenciano de Oncología, Calle Gregorio Gea 31, 4to Piso, 46009 Valencia, Spain.
Cancers
|August 14, 2025
Summary
A new AS score model predicts angiosarcoma outcomes, identifying aggressive tumors with high immune infiltration. This score, combined with Immunoscore, refines prognosis and may guide immunotherapy selection for patients with poor outcomes.
Area of Science:
- Oncology
- Immunology
- Genomics
Background:
- Angiosarcomas (ASs) are aggressive tumors with poor response to systemic treatments and short survival.
- Developing prognostic models is crucial for improving patient outcomes.
Purpose of the Study:
- To develop and validate an immune-related prognostic model, the AS score, for angiosarcomas.
- To assess the prognostic value of the AS score and its correlation with the tumor immune microenvironment.
Main Methods:
- A prognostic model was developed using gene expression data from 25 angiosarcoma samples.
- Candidate genes were identified using Maxstat and log-rank testing.
- The AS score was computed using Cox regression coefficients and validated in the TCGA Sarcoma cohort (n=253).
- The Immunoscore was inferred using the ESTIMATE package.
Main Results:
- Four genes (IGF1R, MAP2K1, SERPINE1, TCF12) were selected for the AS score, classifying cases into distinct prognostic groups (p=0.00012).
- High AS scores correlated with poorer outcomes and were associated with higher Immunoscore values (p=0.025).
- Validation in the TCGA cohort confirmed the prognostic value of both scores and their strong correlation (p=2.9 × 10^-8).
- The AS score effectively stratified outcomes in tumors with high Immunoscore, identifying a subgroup with poor prognosis potentially sensitive to immunotherapy.
Conclusions:
- The combined AS score and Immunoscore classification offers prognostic relevance in sarcoma.
- Angiosarcomas with high Immunoscore and high AS score may benefit from immunotherapy.
- Prospective validation of this approach is warranted.
