Comparative Effectiveness of Artificial Intelligence Versus Conventional Methods for Detecting Peritoneal Metastasis
Mohamed Elsaigh1, Safa Baqar2, Bakhtawar Awan3
1Emergency Surgery, Northwick Park Hospital, London North West University Healthcare NHS Trust, London, GBR.
Cureus
|November 28, 2025
Summary
Artificial intelligence (AI) significantly improves early detection of colorectal cancer peritoneal metastasis over conventional methods. AI tools offer higher accuracy and efficiency, aiding clinical decisions but require further validation for widespread use.
Area of Science:
- Oncology
- Medical Imaging
- Artificial Intelligence
Background:
- Colorectal cancer (CRC) is a leading cause of cancer death globally.
- Peritoneal metastasis in CRC patients is associated with poor prognosis and challenging early detection.
- Conventional imaging techniques like CT have limited sensitivity for small peritoneal lesions.
Purpose of the Study:
- To systematically assess AI and machine learning's accuracy and efficiency in detecting peritoneal metastasis in CRC.
- To compare AI-driven methods against conventional imaging and clinical assessments.
- To evaluate AI's role in predicting tumor spread patterns.
Main Methods:
- Systematic review following PRISMA guidelines (2015-2025).
- Searched PubMed, Web of Science, Cochrane, Embase, Scopus databases.
- Included 22 studies (over 40,000 patients); assessed quality using QUADAS-2 and modified Radiomics Quality Score.
Main Results:
- AI consistently outperformed traditional methods.
- AI-assisted cytological detection: >95% accuracy, 99% specificity.
- Radiomics models achieved AUCs up to 0.941; ctDNA integration increased risk identification 8.5-fold.
- Computer-assisted laparoscopy improved surgical diagnostic accuracy from 52% to 79%.
Conclusions:
- AI shows promise for enhanced, objective, and faster detection of peritoneal metastasis.
- Human-AI collaboration can improve clinical decision-making.
- Further large-scale prospective and external validation studies are crucial before clinical adoption.
