Deep learning for detecting early gastric cancer with white-light endoscopy: a systematic review and meta-analysis

Key Points

• Detection accuracy for early gastric cancer using deep learning reached 90%, indicating high diagnostic potential.
• The systematic review included 15 studies and analyzed detection metrics from various datasets on gastric cancer.
• Application of deep learning techniques in white-light endoscopy showed substantial improvement in early diagnosis rates compared to traditional methods.
• These findings highlight the need for clinical incorporation of advanced algorithms to enhance gastric cancer screening protocols.