文章题目:ArtificialIntelligence-assistedSystemImprovesEndoscopicIdentificationofColorectalNeoplasms
研究人员:Shin-eiKudoetal.
研究单位:DigestiveDiseaseCenter,ShowaUniversityNorthernYokohamaHospital,Yokohama,Japan
发表时间:8September
期刊名称:ClinGastroenterolHepatol.
影响因子:7.(Q1)
01
核心亮点
1.EndoBRAIN?showedsufficientdiagnosticabilitiesforbothdiminutiveandsmallcolorectalpolyps.(EndoBRAIN?对微型和小型大肠息肉都有足够的诊断能力。)
2.EndoBRAIN?’sperformancealsomettheclinicalthresholdfor“leaveinsitu”strategieswithoutpathologyinthisstillimage-basedtrial.(在这项基于静态图像的试验中,EndoBRAIN?的表现也达到了“原地不动”策略的临床阈值。)
02
思路与方法
1.TheEndoBRAINsystemwasinitiallytrainedusing69,endocytoscopicimages,takenat-foldmagnification,frompatientswithcolorectalpolypswhounderwentendoscopyat5academiccentersinJapanfromOctoberthroughMarch.(对年10月至年3月在日本5个学术中心接受内镜检查的结直肠息肉患者,使用放大倍的69张内窥镜图像对EndoBRAIN系统进行初步训练。)
2.Thisstudyperformedaretrospective
转载请注明:http://www.xueyauny.com/drbs/6501.html