Digestive and liver disease : official journal of the Italian Society of Gastroenterology and the Italian Association for the Study of the Liver · 2025
Evaluation of the use of convoluted neural network for detecting early gastric cancer and predicting its invasion depth: A systematic review and meta-analysis.
СтатьяСистематический обзорМетаанализНазвание и аннотация: Перевод пока не подготовлен
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Agarwal Samagra, Rajput Mahender Singh, Pandey Shivam, Ramai Daryl, Tabibian James H, Ofosu Andrew, Barakat Monique T, Girotra Mohit, Mahapatra Soumya Jagannath
Аннотация
BACKGROUND AND AIMS: Identification and assessment of early gastric cancer (EGC) has important prognostic implications. We aimed to review the role of artificial intelligence (AI) assisted endoscopy in the diagnosis and depth assessment of EGC.
METHODS: We searched Pubmed, EMBASE, Web of Science, and Cochrane Library database for studies assessing the role of AI in EGC. Included studies were pooled to obtain summary receiver operator characteristics curve (sROC), pooled sensitivity and specificity. Risk of bias was estimated using QUADAS-2 assessment tool and publication bias was assessed using funnel plots.
RESULTS: We included 19 studies, of which 11 evaluated the role of AI in detection of EGC and 10 evaluated depth of invasion. AI-assisted endoscopy was accurate in detecting EGC with area under sROC of 0.95 (95 % CI:0.93-0.98), pooled sensitivity of 84 % (95 % CI 84-85 %) and pooled specificity of 92 % (95 % CI 92-93 %). Similarly, depth of invasion could be assessed with area under sROC of 0.89 (95 % CI:0.86-0.92), pooled sensitivity of 84 % (95 % CI 82-86 %), and pooled specificity of 89 % (95 %CI 87-90 %). No significant experimental or publication bias was found.
CONCLUSIONS: AI assisted endoscopy has a promising role in diagnosis of EGC and deciding its further management.
METHODS: We searched Pubmed, EMBASE, Web of Science, and Cochrane Library database for studies assessing the role of AI in EGC. Included studies were pooled to obtain summary receiver operator characteristics curve (sROC), pooled sensitivity and specificity. Risk of bias was estimated using QUADAS-2 assessment tool and publication bias was assessed using funnel plots.
RESULTS: We included 19 studies, of which 11 evaluated the role of AI in detection of EGC and 10 evaluated depth of invasion. AI-assisted endoscopy was accurate in detecting EGC with area under sROC of 0.95 (95 % CI:0.93-0.98), pooled sensitivity of 84 % (95 % CI 84-85 %) and pooled specificity of 92 % (95 % CI 92-93 %). Similarly, depth of invasion could be assessed with area under sROC of 0.89 (95 % CI:0.86-0.92), pooled sensitivity of 84 % (95 % CI 82-86 %), and pooled specificity of 89 % (95 %CI 87-90 %). No significant experimental or publication bias was found.
CONCLUSIONS: AI assisted endoscopy has a promising role in diagnosis of EGC and deciding its further management.