{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,13]],"date-time":"2026-10-13T10:21:28Z","timestamp":1791886888232,"version":"4.3.4"},"reference-count":36,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2025,7,10]],"date-time":"2025-07-10T00:00:00Z","timestamp":1752105600000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"DOI":"10.13039\/501100001809","name":"National Natural Science Foundation of China","doi-asserted-by":"publisher","id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Oncol."],"abstract":"<jats:sec><jats:title>Objectives<\/jats:title><jats:p>Non-contrast MRI(NC-MRI) is an attractive option for liver tumors screening and follow-up. This study aims to develop and validate a deep convolutional neural network for the classification of liver lesions using non-contrast MRI.<\/jats:p><\/jats:sec><jats:sec><jats:title>Methods<\/jats:title><jats:p>A total of 50418 enhanced MRI images from 1959 liver tumor patients across three centers were included. Inception-ResNet V2 was used to generate four models through transfer-learning for three-way lesion classification, which processed T2-weighted, diffusion-weighted (DWI) and multiphasic T1-weighted images. The models were then validated using one independent internal and two external datasets with 5172, 2916, and 1338 images, respectively. The efficacy of non-contrast models (T2,T2+DWI) in differentiating between benign and malignant liver lesions at the patient level was also evaluated and compared with radiologists. The performance of models was evaluated using the area under the receiver operating characteristic curve (AUC),sensitivity and specificity.<\/jats:p><\/jats:sec><jats:sec><jats:title>Results<\/jats:title><jats:p>Similar to multi-sequence and enhanced image-based models, the non-contrast models showed comparable accuracy in classifying liver lesions as benign, primary malignant or metastatic. In the independent internal cohort, the T2+DWI model achieved AUC of 0.91(95% CI,0.888\u20130.932), 0.873(0.848-0.899), and 0.876(0.840-0.911) for three tumour categories, respectively. The sensitivities for distinguishing malignant tumors in three validation sets were 98.1%, 89.7%, and 87.5%%, with specificities over 70% in all three sets.<\/jats:p><\/jats:sec><jats:sec><jats:title>Conclusions<\/jats:title><jats:p>Our deep-learning-based model yielded good applicability in classifying liver lesions in non-contrast MRI. It provides a potential alternative for screening liver tumors with the advantage of reducing costs, scanning time and contrast-agents risks. It is more suitable for benign tumours follow-up, surveillance of HCC and liver metastasis that need periodic repetitive examinations.<\/jats:p><\/jats:sec>","DOI":"10.3389\/fonc.2025.1582322","type":"journal-article","created":{"date-parts":[[2025,7,10]],"date-time":"2025-07-10T06:16:05Z","timestamp":1752128165000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":6,"title":["Deep learning-assisted diagnosis of liver tumors using non-contrast magnetic resonance imaging: a multicenter study"],"prefix":"10.3389","volume":"15","author":[{"given":"Shihui","family":"Zhen","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Peng","family":"Zhang","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Hanxiao","family":"Huang","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Zhiyu","family":"Jiang","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yankai","family":"Jiang","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jihong","family":"Sun","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Liqing","family":"Zhang","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Mei","family":"Ruan","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Qingqing","family":"Chen","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yujun","family":"Wang","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yubo","family":"Tao","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Weizhi","family":"Luo","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Ming","family":"Cheng","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Zhetuo","family":"Qi","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Wei","family":"Lu","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Hai","family":"Lin","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Xiujun","family":"Cai","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2025,7,10]]},"reference":[{"key":"B1","doi-asserted-by":"publisher","first-page":"17","DOI":"10.3322\/caac.21763","article-title":"Cancer statistics, 2023","volume":"73","author":"Siegel","year":"2023","journal-title":"CA Cancer J Clin"},{"key":"B2","doi-asserted-by":"publisher","DOI":"10.1038\/s41571-023-00825-3","article-title":"Global trends in hepatocellular carcinoma epidemiology: implications for screening, prevention and therapy","volume":"20","author":"Singal","year":"2023","journal-title":"Nat Rev Clin Oncol"},{"key":"B3","doi-asserted-by":"publisher","first-page":"364ra153","DOI":"10.1126\/scitranslmed.aag2306","article-title":"Local and transient gene expression primes the liver to resist cancer metastasis","volume":"8","author":"Goodwin","year":"2016","journal-title":"Sci Transl Med"},{"key":"B4","doi-asserted-by":"publisher","DOI":"10.1148\/radiol.2017170706","article-title":"Deep learning with convolutional neural network for differentiation of liver masses at dynamic contrast-enhanced CT: A preliminary study","volume":"286","author":"Yasaka","year":"2018","journal-title":"Radiology"},{"key":"B5","doi-asserted-by":"publisher","DOI":"10.1016\/j.ejrad.2017.08.028","article-title":"Primary benign liver lesions","volume":"95","author":"Grazioli","year":"2017","journal-title":"Eur J Radiol"},{"key":"B6","doi-asserted-by":"publisher","DOI":"10.1038\/ajg.2014.213","article-title":"ACG clinical guideline: the diagnosis and management of focal liver lesions","volume":"109","author":"Marrero","year":"2014","journal-title":"Am J Gastroenterol"},{"key":"B7","doi-asserted-by":"publisher","first-page":"6","DOI":"10.1038\/s41572-020-00240-3","article-title":"Hepatocellular carcinoma","volume":"7","author":"Llovet","year":"2021","journal-title":"Nat Rev Dis Primers"},{"key":"B8","doi-asserted-by":"publisher","DOI":"10.1001\/jamaoncol.2016.3147","article-title":"MRI with liver-specific contrast for surveillance of patients with cirrhosis at high risk of hepatocellular carcinoma","volume":"3","author":"Kim","year":"2017","journal-title":"JAMA Oncol"},{"key":"B9","doi-asserted-by":"publisher","first-page":"1","DOI":"10.3310\/hta17090","article-title":"Contrast-enhanced ultrasound using SonoVue(R) (sulphur hexafluoride microbubbles) compared with contrast-enhanced computed tomography and contrast-enhanced magnetic resonance imaging for the characterisation of focal liver lesions and detection of liver metastases: a systematic review and cost-effectiveness analysis","volume":"17","author":"Westwood","year":"2013","journal-title":"Health Technol Assess"},{"key":"B10","doi-asserted-by":"publisher","DOI":"10.1016\/j.jhep.2019.12.001","article-title":"Non-enhanced magnetic resonance imaging as a surveillance tool for hepatocellular carcinoma: Comparison with ultrasound","volume":"72","author":"Park","year":"2020","journal-title":"J Hepatol"},{"key":"B11","doi-asserted-by":"publisher","DOI":"10.1007\/s00330-020-06754-4","article-title":"Comparison of diagnostic performance of non-contrast MRI and abbreviated MRI using gadoxetic acid in initially diagnosed hepatocellular carcinoma patients: a simulation study of surveillance for hepatocellular carcinomas","volume":"30","author":"Whang","year":"2020","journal-title":"Eur Radiol"},{"key":"B12","doi-asserted-by":"publisher","DOI":"10.1007\/s00261-019-02277-9","article-title":"Diagnostic performance of a minimized protocol of non-contrast MRI for hepatocellular carcinoma surveillance","volume":"45","author":"Kim","year":"2020","journal-title":"Abdom Radiol (NY)"},{"key":"B13","doi-asserted-by":"publisher","DOI":"10.1016\/j.mri.2013.12.021","article-title":"Noncontrast MRI with diffusion-weighted imaging as the sole imaging modality for detecting liver Malignancy in patients with high risk for hepatocellular carcinoma","volume":"32","author":"Kim","year":"2014","journal-title":"Magn Reson Imaging"},{"key":"B14","doi-asserted-by":"publisher","DOI":"10.3348\/kjr.2018.19.4.568","article-title":"The diagnostic performance of liver MRI without intravenous contrast for detecting hepatocellular carcinoma: A case-controlled feasibility study","volume":"19","author":"Han","year":"2018","journal-title":"Korean J Radiol"},{"key":"B15","doi-asserted-by":"publisher","DOI":"10.1007\/s00330-020-07344-0","article-title":"The diagnostic performance of a simulated \u201cshort\u201d gadoxetic acid-enhanced MRI protocol is similar to that of a conventional protocol for the detection of colorectal liver metastases","volume":"31","author":"Ghorra","year":"2021","journal-title":"Eur Radiol"},{"key":"B16","doi-asserted-by":"publisher","DOI":"10.1177\/0284185118788901","article-title":"Non-contrast liver MRI as an alternative to gadoxetic acid-enhanced MRI for liver metastasis from colorectal cancer","volume":"60","author":"Hwang","year":"2019","journal-title":"Acta Radiol"},{"key":"B17","doi-asserted-by":"publisher","DOI":"10.1148\/rg.2019180123","article-title":"Abbreviated MRI protocols for the abdomen","volume":"39","author":"Canellas","year":"2019","journal-title":"Radiographics"},{"key":"B18","doi-asserted-by":"publisher","DOI":"10.1002\/jmri.28059","article-title":"Abbreviated gadoxetic acid-enhanced MRI for the detection of liver metastases in patients with potentially resectable pancreatic ductal adenocarcinoma","volume":"56","author":"Yamaguchi","year":"2022","journal-title":"J Magn Reson Imaging"},{"key":"B19","doi-asserted-by":"publisher","DOI":"10.1001\/jama.2018.13362","article-title":"Gadolinium retention after contrast-enhanced MRI","volume":"320","author":"Levine","year":"2018","journal-title":"JAMA"},{"key":"B20","doi-asserted-by":"publisher","DOI":"10.1002\/hep.24679","article-title":"Magnetic resonance imaging of focal liver lesions: Approach to imaging diagnosis","volume":"54","author":"Fowler","year":"2011","journal-title":"Hepatology"},{"key":"B21","doi-asserted-by":"publisher","first-page":"110840","DOI":"10.1016\/j.ejrad.2023.110840","article-title":"The application of abbreviated MRI protocols in Malignant liver lesions surveillance","volume":"164","author":"Winder","year":"2023","journal-title":"Eur J Radiol"},{"key":"B22","doi-asserted-by":"publisher","DOI":"10.1097\/HEP.0000000000000339","article-title":"Screening of liver cancer with abbreviated MRI","volume":"78","author":"Ronot","year":"2023","journal-title":"Hepatology"},{"key":"B23","doi-asserted-by":"publisher","DOI":"10.1038\/nature14539","article-title":"Deep learning","volume":"521","author":"LeCun","year":"2015","journal-title":"Nature"},{"key":"B24","doi-asserted-by":"publisher","DOI":"10.3322\/caac.21552","article-title":"Artificial intelligence in cancer imaging: Clinical challenges and applications","volume":"69","author":"Bi","year":"2019","journal-title":"CA Cancer J Clin"},{"key":"B25","doi-asserted-by":"publisher","first-page":"24","DOI":"10.1007\/s00261-019-02306-7","article-title":"Deep convolutional neural network applied to the liver imaging reporting and data system (LI-RADS) version 2014 category classification: a pilot study","volume":"45","author":"Yamashita","year":"2020","journal-title":"Abdominal Radiol"},{"key":"B26","doi-asserted-by":"publisher","first-page":"9458","DOI":"10.1038\/s41598-020-65875-4","article-title":"Detection of hepatocellular carcinoma in contrast-enhanced magnetic resonance imaging using deep learning classifier: A multi-center retrospective study","volume":"10","author":"Kim","year":"2020","journal-title":"Sci Reports"},{"key":"B27","doi-asserted-by":"publisher","DOI":"10.3389\/fonc.2020.00680","article-title":"Deep learning for accurate diagnosis of liver tumor based on magnetic resonance imaging and clinical data","volume":"10","author":"Zhen","year":"2020","journal-title":"Front Oncol"},{"key":"B28","doi-asserted-by":"publisher","DOI":"10.1007\/s00330-019-06214-8","article-title":"Deep learning for liver tumor diagnosis part II: convolutional neural network interpretation using radiologic imaging features","volume":"29","author":"Wang","year":"2019","journal-title":"Eur Radiol"},{"key":"B29","doi-asserted-by":"publisher","first-page":"h5527","DOI":"10.1136\/bmj.h5527","article-title":"STARD 2015: an updated list of essential items for reporting diagnostic accuracy studies","volume":"351","author":"Bossuyt","year":"2015","journal-title":"BMJ"},{"key":"B30","doi-asserted-by":"publisher","DOI":"10.1109\/TNNLS.2014.2330900","article-title":"Transfer learning for visual categorization: A survey","volume":"26","author":"Shao","year":"2015","journal-title":"IEEE Trans Neural Networks Learn Syst"},{"key":"B31","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV.2017.74","article-title":"Grad-CAM: visual explanations from deep networks via gradient-based localization","author":"Selvaraju","year":"2017","journal-title":"Proceedings of the IEEE International Conference on Computer Vision (ICCV)."},{"key":"B32","article-title":"Visualizing Data using t-SNE","volume":"9","author":"Hinton","year":"2008","journal-title":"J Mach Learn Res"},{"key":"B33","doi-asserted-by":"publisher","DOI":"10.1109\/JBHI.6221020","article-title":"Extending 2-D convolutional neural networks to 3-D for advancing deep learning cancer classification with application to MRI liver tumor differentiation","volume":"23","author":"Trivizakis","year":"2019","journal-title":"IEEE J BioMed Health Inform"},{"key":"B34","doi-asserted-by":"publisher","DOI":"10.1002\/hep.25681","article-title":"Hepatocellular carcinoma: Detection with diffusion-weighted versus contrast-enhanced magnetic resonance imaging in pretransplant patients","volume":"56","author":"Park","year":"2012","journal-title":"Hepatology"},{"key":"B35","doi-asserted-by":"publisher","first-page":"49","DOI":"10.1186\/s41747-019-0126-1","article-title":"HCC screening: assessment of an abbreviated non-contrast MRI protocol","volume":"3","author":"Chan","year":"2019","journal-title":"Eur Radiol Exp"},{"key":"B36","doi-asserted-by":"publisher","first-page":"20180177","DOI":"10.1259\/bjr.20180177","article-title":"Detection of recurrent hepatocellular carcinoma after surgical resection: Non-contrast liver MR imaging with diffusion-weighted imaging versus gadoxetic acid-enhanced MR imaging","volume":"91","author":"Min","year":"2018","journal-title":"Br J Radiol"}],"container-title":["Frontiers in Oncology"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fonc.2025.1582322\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,7,10]],"date-time":"2025-07-10T06:16:06Z","timestamp":1752128166000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fonc.2025.1582322\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,7,10]]},"references-count":36,"alternative-id":["10.3389\/fonc.2025.1582322"],"URL":"https:\/\/doi.org\/10.3389\/fonc.2025.1582322","relation":{},"ISSN":["2234-943X"],"issn-type":[{"value":"2234-943X","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,7,10]]},"article-number":"1582322"}}