Related Experiment Video
Updated: Jun 18, 2025
03:05
Author Spotlight: Advancing Early Detection and Treatment of Gastrointestinal Tumors
Published on: February 16, 2024
1.0K
Explainable machine learning models for early gastric cancer diagnosis
Hongyang Du1, Qingfen Yang2, Aimin Ge2
1Heze Administrative Approval Guarantee Center, 3443 Huanghe East Road, Heze City, 274000, Shandong Province, China.
Scientific Reports
|July 29, 2024
Summary
Explainable machine learning models show promise for early gastric cancer diagnosis. Transparent AI tools can improve accuracy and patient outcomes by increasing clinical trust and identifying key diagnostic biomarkers.
Area of Science:
- Oncology
- Medical Informatics
- Artificial Intelligence
Background:
- Gastric cancer is a major global health issue, particularly prevalent in East Asia.
- Early diagnosis is crucial for improving patient survival rates and treatment efficacy.
- Current diagnostic methods can be enhanced by advanced computational approaches.
Purpose of the Study:
- To evaluate the efficacy of explainable machine learning (ML) models for early gastric cancer detection.
- To assess the role of model transparency in clinical acceptance and diagnostic accuracy.
- To identify critical biomarkers and clinical features for early gastric cancer identification.
Main Methods:
- Utilized and evaluated several ML models, including WeightedEnsemble, CatBoost, and RandomForest.
- Focused on the explainability aspect of the ML models used.
- Performed comprehensive evaluations on diagnostic performance.
Main Results:
- WeightedEnsemble, CatBoost, and RandomForest models demonstrated significant potential in accurately diagnosing early gastric cancer.
- Explainable models enhanced trust and clinical acceptance, contributing to improved diagnostic accuracy.
- Key biomarkers and clinical features vital for early detection were identified.
Conclusions:
- Explainable ML models offer a powerful tool for enhancing early gastric cancer diagnosis.
- Model transparency is essential for clinical integration and improving patient outcomes.
- The developed approach has potential applications in other medical diagnostic fields.
