Reducing demographic bias in biomedical machine learning for cancer detection using cfDNA methylation - PubMed
5 hours ago
- #Cancer detection
- #Bias correction
- #Machine learning
- Machine learning models in biomedical research often suffer from demographic biases due to imbalanced clinical datasets.
- DeBias is a computational framework designed to mitigate demographic biases in high-dimensional biomedical datasets.
- DeBias identifies and removes bias-associated subspaces using control samples, preserving disease-specific signals.
- Applied to cell-free DNA methylation data for cancer detection, DeBias significantly reduces demographic bias in features.
- DeBias outperforms existing methods in improving cancer detection performance for minority populations.
- The approach is validated in independent cohorts, demonstrating robustness.
- DeBias represents a step toward more equitable machine learning models in biomedical research.
- The study complies with ethical regulations and has received IRB approval.
- Several authors have competing interests related to EarlyDiagnostics, Inc., including patents and stock options.