Radiomics Improve Colorectal Cancer Recurrence Risk Prediction
This meta-analysis, with literature searched through January 1, 2025, pooled 17 studies including 4,600 colorectal cancer patients to evaluate radiomics-based ML models for recurrence prediction. Radiomics models achieved pooled c-index 0.80 versus 0.73 for clinical features, while integrated radiomics+clinical models reached 0.83 in validation sets. Despite promising discrimination, overall study quality was low (mean RQS 13.23/36) with heterogeneity and limited prospective validation.
Key Points
- 1Report pooled radiomics models achieved c-index 0.80 versus 0.73 for clinical features.
- 2Note combined radiomics+clinical models reached c-index 0.83, suggesting additive predictive value.
- 3Advise cautious interpretation due to low Radiomics Quality Scores and heterogeneous, small studies.
Scoring Rationale
Meta-analytic synthesis provides robust pooled estimates, but moderate study quality and heterogeneity limit clinical generalizability.
Sources
Public references used for this report.
Practice with real Ad Tech data
90 SQL & Python problems · 15 industry datasets
250 free problems · No credit card
See all Ad Tech problems
