Review

利用人工智能软件进行CT肺癌筛查结节与癌症检测的系统性准确性评价

Software using artificial intelligence for nodule and cancer detection in CT lung cancer screening: systematic review of test accuracy studies

THORAX

影响因子:7.7

分区:医学1区 Top / 呼吸系统2区

发表日期:2024 Oct 16

作者: Julia Geppert, Asra Asgharzadeh, Anna Brown, Chris Stinton, Emma J Helm, Surangi Jayakody, Daniel Todkill, Daniel Gallacher, Hesam Ghiasvand, Mubarak Patel, Peter Auguste, Alexander Tsertsvadze, Yen-Fu Chen, Amy Grove, Bethany Shinkins, Aileen Clarke, Sian Taylor-Phillips

DOI: 10.1136/thorax-2024-221662

摘要

为了评估人工智能（AI）软件在CT肺癌筛查中的准确性及其影响。对已获得CE标志的基于AI的自动检测和分析肺部结节的软件进行了系统性综述。检索了包括Medline、Embase和Cochrane CENTRAL在内的多个数据库，从2012年到2023年3月。纳入了报告检测准确性或对阅读时间或临床管理影响的原始研究。采用QUADAS-2和QUADAS-C工具评估偏倚风险。进行了叙述性合成。共有11项研究符合条件，评估了六种不同的AI软件，涉及19,770名患者。所有研究偏倚风险均较高，存在多重适用性问题。与未辅助阅读相比，AI辅助阅读速度更快，通常能提高敏感性（检测/分类可行动结节的提升为+5%至+20%；检测/分类恶性结节的提升为+3%至+15%），但特异性略有降低（正确检测/分类无行动结节的下降为-7%至-3%；正确检测/分类无恶性结节的下降为-8%至-6%）。AI辅助倾向于增加结节被归入高风险类别的比例。在假设癌症患病率为0.5%的情况下，这些结果意味着每百万人筛查中将多检测出150至750例癌症，但也会导致59,700至79,600名无癌症的筛查人群接受不必要的CT监测。AI在肺癌筛查中的应用可能提高敏感性，但也增加假阳性和不必要的监测。未来的研究应提升AI辅助阅读的特异性，并通过优化研究设计来减少偏倚和适用性问题。CRD42021298449。

Abstract

To examine the accuracy and impact of artificial intelligence (AI) software assistance in lung cancer screening using CT.A systematic review of CE-marked, AI-based software for automated detection and analysis of nodules in CT lung cancer screening was conducted. Multiple databases including Medline, Embase and Cochrane CENTRAL were searched from 2012 to March 2023. Primary research reporting test accuracy or impact on reading time or clinical management was included. QUADAS-2 and QUADAS-C were used to assess risk of bias. We undertook narrative synthesis.Eleven studies evaluating six different AI-based software and reporting on 19 770 patients were eligible. All were at high risk of bias with multiple applicability concerns. Compared with unaided reading, AI-assisted reading was faster and generally improved sensitivity (+5% to +20% for detecting/categorising actionable nodules; +3% to +15% for detecting/categorising malignant nodules), with lower specificity (-7% to -3% for correctly detecting/categorising people without actionable nodules; -8% to -6% for correctly detecting/categorising people without malignant nodules). AI assistance tended to increase the proportion of nodules allocated to higher risk categories. Assuming 0.5% cancer prevalence, these results would translate into additional 150-750 cancers detected per million people attending screening but lead to an additional 59 700 to 79 600 people attending screening without cancer receiving unnecessary CT surveillance.AI assistance in lung cancer screening may improve sensitivity but increases the number of false-positive results and unnecessary surveillance. Future research needs to increase the specificity of AI-assisted reading and minimise risk of bias and applicability concerns through improved study design.CRD42021298449.