Retrospective Study
Copyright ©The Author(s) 2020. Published by Baishideng Publishing Group Inc. All rights reserved.
World J Gastroenterol. Jul 7, 2020; 26(25): 3660-3672
Published online Jul 7, 2020. doi: 10.3748/wjg.v26.i25.3660
Multiphase convolutional dense network for the classification of focal liver lesions on dynamic contrast-enhanced computed tomography
Su-E Cao, Lin-Qi Zhang, Si-Chi Kuang, Wen-Qi Shi, Bing Hu, Si-Dong Xie, Yi-Nan Chen, Hui Liu, Si-Min Chen, Ting Jiang, Meng Ye, Han-Xi Zhang, Jin Wang
Su-E Cao, Lin-Qi Zhang, Si-Chi Kuang, Wen-Qi Shi, Bing Hu, Si-Dong Xie, Si-Min Chen, Ting Jiang, Han-Xi Zhang, Jin Wang, Department of Radiology, The Third Affiliated Hospital, Sun Yat-Sen University, Guangzhou 510630, Guangdong Province, China
Yi-Nan Chen, Hui Liu, Meng Ye, Department of Scientific and Technological Research, 12 Sigma Technologies, Beijing 100102, China
Author contributions: Cao SE, Zhang LQ, Shi WQ, Chen YN, Liu H, and Ye M contributed to the conception and design of the study; Cao SE, Kuang SC, Shi WQ, Hu B, Jiang T, Chen SM, and Zhang HX collected the patient data, analyzed and interpreted the data; Cao SE wrote original draft and revised the manuscript; Wang J contributed to the conception of the study and provided final approval of the version to be submitted and any revised versions.
Supported by National Natural Science Foundation of China, No. 91959118; Science and Technology Program of Guangzhou, China, No. 201704020016; SKY Radiology Department International Medical Research Foundation of China, No. Z-2014-07-1912-15; and Clinical Research Foundation of the 3rd Affiliated Hospital of Sun Yat-Sen University, No. YHJH201901.
Institutional review board statement: The study was reviewed and approved for publication by our Institutional Reviewer.
Informed consent statement: All study participants or their legal guardian provided informed written consent about personal and medical data collection prior to study enrolment.
Conflict-of-interest statement: All the Authors have no conflict of interest related to the manuscript.
Data sharing statement: The original anonymous dataset is available on request from the corresponding author at wangjin3@mail.sysu.edu.cn.
Open-Access: This article is an open-access article that was selected by an in-house editor and fully peer-reviewed by external reviewers. It is distributed in accordance with the Creative Commons Attribution NonCommercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/
Corresponding author: Jin Wang, MD, Doctor, Professor, Department of Radiology, The Third Affiliated Hospital, Sun Yat-Sen University, No. 600, Tianhe Road, Tianhe District, Guangzhou 510630, Guangdong Province, China. wangjin3@mail.sysu.edu.cn
Received: March 16, 2020
Peer-review started: March 16, 2020
First decision: April 25, 2020
Revised: May 8, 2020
Accepted: June 4, 2020
Article in press: June 4, 2020
Published online: July 7, 2020
Processing time: 112 Days and 1.3 Hours
Abstract
BACKGROUND

The accurate classification of focal liver lesions (FLLs) is essential to properly guide treatment options and predict prognosis. Dynamic contrast-enhanced computed tomography (DCE-CT) is still the cornerstone in the exact classification of FLLs due to its noninvasive nature, high scanning speed, and high-density resolution. Since their recent development, convolutional neural network-based deep learning techniques has been recognized to have high potential for image recognition tasks.

AIM

To develop and evaluate an automated multiphase convolutional dense network (MP-CDN) to classify FLLs on multiphase CT.

METHODS

A total of 517 FLLs scanned on a 320-detector CT scanner using a four-phase DCE-CT imaging protocol (including precontrast phase, arterial phase, portal venous phase, and delayed phase) from 2012 to 2017 were retrospectively enrolled. FLLs were classified into four categories: Category A, hepatocellular carcinoma (HCC); category B, liver metastases; category C, benign non-inflammatory FLLs including hemangiomas, focal nodular hyperplasias and adenomas; and category D, hepatic abscesses. Each category was split into a training set and test set in an approximate 8:2 ratio. An MP-CDN classifier with a sequential input of the four-phase CT images was developed to automatically classify FLLs. The classification performance of the model was evaluated on the test set; the accuracy and specificity were calculated from the confusion matrix, and the area under the receiver operating characteristic curve (AUC) was calculated from the SoftMax probability outputted from the last layer of the MP-CDN.

RESULTS

A total of 410 FLLs were used for training and 107 FLLs were used for testing. The mean classification accuracy of the test set was 81.3% (87/107). The accuracy/specificity of distinguishing each category from the others were 0.916/0.964, 0.925/0.905, 0.860/0.918, and 0.925/0.963 for HCC, metastases, benign non-inflammatory FLLs, and abscesses on the test set, respectively. The AUC (95% confidence interval) for differentiating each category from the others was 0.92 (0.837-0.992), 0.99 (0.967-1.00), 0.88 (0.795-0.955) and 0.96 (0.914-0.996) for HCC, metastases, benign non-inflammatory FLLs, and abscesses on the test set, respectively.

CONCLUSION

MP-CDN accurately classified FLLs detected on four-phase CT as HCC, metastases, benign non-inflammatory FLLs and hepatic abscesses and may assist radiologists in identifying the different types of FLLs.

Keywords: Deep learning; Convolutional neural networks; Focal liver lesions; Classification; Multiphase computed tomography; Dynamic enhancement pattern

Core tip: We developed and evaluated a deep learning-based convolutional neural network (CNN) to classify focal liver lesions (FLLs) on multiphase computed tomography. The most important highlight of the current study is that, to the best of our knowledge, this study is the first to employ four-channel input data to preserve the dynamic enhancement properties. The combination of the lesion's dynamic enhancement pattern with a CNN can imitate the image diagnosis of radiologists and is expected to improve diagnostic accuracy. It was interesting to note that the accuracy and specificity of differentiating each category from others were high. This model may become an efficient tool to assist radiologists in the classification of FLLs.