Artificial Intelligence and Rectal Cancer: Beyond Images
View abstract on PubMed
Summary
This summary is machine-generated.Artificial intelligence (AI) models analyzing non-image data, like text or numerical data, show significant promise for understanding cancer variability. This review highlights the underappreciated importance of non-image AI inputs in rectal cancer research and clinical practice.
Area Of Science
- Artificial Intelligence in Medicine
- Oncology Data Science
Background
- Medical big data and cancer variability present challenges.
- Artificial intelligence (AI) models can process diverse data types, including images, numerical data, categories, and free text.
- Existing literature often overemphasizes image-based AI models, neglecting other crucial data sources.
Purpose Of The Study
- To review artificial intelligence models, focusing on non-image data inputs.
- To evaluate the performance and representation of non-image and combined AI models in medical research, using rectal cancer as a case study.
Main Methods
- A comprehensive literature search was performed using PubMed and Scopus.
- Searches were conducted without temporal limits and in English.
- Filters were applied to secondary literature for relevant studies.
Main Results
- AI models were categorized into image-based, non-image-based, and combined (hybrid) types.
- Non-image AI models demonstrated significant performance, challenging the focus on image-based approaches.
- Combined models often outperformed unimodal models, though multicenter validation studies for non-image and combined models are under-represented.
Conclusions
- This review is the first to focus on non-image AI inputs in medical research, alone or combined with images.
- Non-image data components warrant greater attention in research and clinical applications.
- Multimodality, extending beyond imaging, is crucial for rectal cancer and potentially other diseases.

