Reproduce “ColonScopeX: Leveraging Explainable Expert Systems with Multimodal Data for Improved Early Diagnosis of Colorectal Cancer” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Natalia Sikora, Robert L. Manschke, Alethea M. Tang, Peter Dunstan, Dean A. Harris, Su Yang
Swansea University researchers developed ColonScopeX, a multimodal, explainable AI framework for early detection of colorectal cancer and pre-cancerous polyps. This system integrates non-invasive Raman spectroscopy data from blood samples with patient metadata, achieving up to 0.975 AUC for cancer detection and generating clinician-friendly textual reports for transparent decision-making.
Key insights
- Multimodal data fusion significantly enhances diagnostic accuracy for both polyps and CRC compared to single-modality approaches, with specific fusion architectures outperforming others.
- Explainable AI, coupled with expert-annotated libraries and textual reporting, is crucial for fostering clinician trust and interpretability of complex AI-driven diagnostic decisions.
- Feature importance varies between polyp and CRC detection; spectral data is consistently vital for CRC, while polyp detection also heavily relies on patient metadata like 'previous malignancy' and 'sex'.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.