Diagnosis of focal liver lesions with deep learning-based multichannel analysis of hepatocyte-specific contrast-enhanced magnetic resonance imaging

作     者：Róbert Stollmayer Bettina K Budai Ambrus Tóth IldikóKalina Erika Hartmann Péter Szoldán Viktor Bérczi Pál Maurovich-Horvat Pál N Kaposi 

作者机构：Department of RadiologyMedical Imaging CentreFaculty of MedicineSemmelweis UniversityBudapest 1083Hungary Department of Transplantation and SurgeryFaculty of MedicineSemmelweis UniversityBudapest 1082Hungary MedInnoScan Research and Development Ltd.Budapest 1112Hungary 

出 版 物：《World Journal of Gastroenterology》 (世界胃肠病学杂志（英文版）)

年 卷 期：2021年第27卷第35期

页      面：5978-5988页

核心收录：

学科分类：0831[工学-生物医学工程（可授工学、理学、医学学位）] 100207[医学-影像医学与核医学] 1002[医学-临床医学] 08[工学] 1010[医学-医学技术（可授医学、理学学位）] 100201[医学-内科学(含：心血管病、血液病、呼吸系病、消化系病、内分泌与代谢病、肾病、风湿病、传染病)] 10[医学] 

基　　金：Alfréd Rényi Institute Magyar Tudományos Akadémia, MTA 

主　　题：Artificial intelligence Multi-parametric magnetic resonance imaging Hepatocyte-specific contrast Densely connected convolutional network Hepatocellular carcinoma Focal nodular hyperplasia 

摘      要：BACKGROUND The nature of input data is an essential factor when training neural *** concerning magnetic resonance imaging(MRI)-based diagnosis of liver tumors using deep learning has been rapidly ***,evidence to support the utilization of multi-dimensional and multi-parametric image data is *** to higher information content,three-dimensional input should presumably result in higher classification ***,the differentiation between focal liver lesions(FLLs)can only be plausible with simultaneous analysis of multisequence MRI *** To compare diagnostic efficiency of two-dimensional(2D)and three-dimensional(3D)-densely connected convolutional neural networks(DenseNet)for FLLs on multi-sequence *** We retrospectively collected T2-weighted,gadoxetate disodium-enhanced arterial phase,portal venous phase,and hepatobiliary phase MRI scans from patients with focal nodular hyperplasia(FNH),hepatocellular carcinomas(HCC)or liver metastases(MET).Our search identified 71 FNH,69 HCC and 76 *** volume registration,the same three most representative axial slices from all sequences were combined into four-channel images to train the 2D-DenseNet264 *** bounding boxes were selected on all scans and stacked into 4D volumes to train the 3D-DenseNet264 *** test set consisted of 10-10-10 *** performance of the models was compared using area under the receiver operating characteristic curve(AUROC),specificity,sensitivity,positive predictive values(PPV),negative predictive values(NPV),and f1 *** The average AUC value of the 2D model(0.98)was slightly higher than that of the 3D model(0.94).Mean PPV,sensitivity,NPV,specificity and f1 scores(0.94,0.93,0.97,0.97,and 0.93)of the 2D model were also superior to metrics of the 3D model(0.84,0.83,0.92,0.92,and 0.83).The classification metrics of FNH were 0.91,1.00,1.00,0.95,and 0.95 using the 2D and 0.90,0.90,0.95,0.95,and 0.90 using the 3D ***