人工智能检测胃肠道腔内病变的诊断准确性：一项系统评价和荟萃分析。

Diagnostic accuracy of artificial intelligence for detecting gastrointestinal luminal pathologies: A systematic review and meta-analysis.

作者信息

Parkash Om, Siddiqui Asra Tus Saleha, Jiwani Uswa, Rind Fahad, Padhani Zahra Ali, Rizvi Arjumand, Hoodbhoy Zahra, Das Jai K

机构信息

Department of Medicine, Aga Khan University, Karachi, Pakistan.

Center of Excellence in Women and Child Health, Aga Khan University, Karachi, Pakistan.

出版信息

Front Med (Lausanne). 2022 Nov 4;9:1018937. doi: 10.3389/fmed.2022.1018937. eCollection 2022.

DOI:10.3389/fmed.2022.1018937

PMID:36405592

原文链接:https://pmc.ncbi.nlm.nih.gov/articles/PMC9672666/

Abstract

BACKGROUND

Artificial Intelligence (AI) holds considerable promise for diagnostics in the field of gastroenterology. This systematic review and meta-analysis aims to assess the diagnostic accuracy of AI models compared with the gold standard of experts and histopathology for the diagnosis of various gastrointestinal (GI) luminal pathologies including polyps, neoplasms, and inflammatory bowel disease.

METHODS

We searched PubMed, CINAHL, Wiley Cochrane Library, and Web of Science electronic databases to identify studies assessing the diagnostic performance of AI models for GI luminal pathologies. We extracted binary diagnostic accuracy data and constructed contingency tables to derive the outcomes of interest: sensitivity and specificity. We performed a meta-analysis and hierarchical summary receiver operating characteristic curves (HSROC). The risk of bias was assessed using Quality Assessment for Diagnostic Accuracy Studies-2 (QUADAS-2) tool. Subgroup analyses were conducted based on the type of GI luminal disease, AI model, reference standard, and type of data used for analysis. This study is registered with PROSPERO (CRD42021288360).

FINDINGS

We included 73 studies, of which 31 were externally validated and provided sufficient information for inclusion in the meta-analysis. The overall sensitivity of AI for detecting GI luminal pathologies was 91.9% (95% CI: 89.0-94.1) and specificity was 91.7% (95% CI: 87.4-94.7). Deep learning models (sensitivity: 89.8%, specificity: 91.9%) and ensemble methods (sensitivity: 95.4%, specificity: 90.9%) were the most commonly used models in the included studies. Majority of studies ( = 56, 76.7%) had a high risk of selection bias while 74% ( = 54) studies were low risk on reference standard and 67% ( = 49) were low risk for flow and timing bias.

INTERPRETATION

The review suggests high sensitivity and specificity of AI models for the detection of GI luminal pathologies. There is a need for large, multi-center trials in both high income countries and low- and middle- income countries to assess the performance of these AI models in real clinical settings and its impact on diagnosis and prognosis.

SYSTEMATIC REVIEW REGISTRATION

[https://www.crd.york.ac.uk/prospero/display_record.php?RecordID=288360], identifier [CRD42021288360].

摘要

背景

人工智能（AI）在胃肠病学领域的诊断方面具有巨大潜力。本系统评价和荟萃分析旨在评估与专家及组织病理学金标准相比，AI模型对包括息肉、肿瘤和炎症性肠病在内的各种胃肠道（GI）腔内病变的诊断准确性。

方法

我们检索了PubMed、CINAHL、Wiley Cochrane图书馆和Web of Science电子数据库，以识别评估AI模型对GI腔内病变诊断性能的研究。我们提取了二元诊断准确性数据并构建列联表，以得出感兴趣的结果：敏感性和特异性。我们进行了荟萃分析和分层汇总接收器操作特征曲线（HSROC）分析。使用诊断准确性研究质量评估-2（QUADAS-2）工具评估偏倚风险。根据GI腔内疾病类型、AI模型、参考标准和用于分析的数据类型进行亚组分析。本研究已在PROSPERO（CRD42021288360）注册。

结果

我们纳入了73项研究，其中31项经过外部验证，并提供了足够的信息以纳入荟萃分析。AI检测GI腔内病变的总体敏感性为91.9%（95%CI：89.0-94.1），特异性为91.7%（95%CI：87.4-94.7）。深度学习模型（敏感性：89.8%，特异性：91.9%）和集成方法（敏感性：95.4%，特异性：90.9%）是纳入研究中最常用的模型。大多数研究（n = 56，76.7%）存在较高的选择偏倚风险，而74%（n = 54）的研究在参考标准方面风险较低，67%（n = 49）在流程和时间偏倚方面风险较低。