Abstract:Background:Early-stage diagnosis and treatment can improve survival rates of liver cancer patients. Dynamic contrast-enhanced MRI provides the most comprehensive information for differential diagnosis of liver tumors. However, MRI diagnosis is affected by subjective experience, so deep learning may supply a new diagnostic strategy. We used convolutional neural networks (CNNs) to develop a deep learning system (DLS) to classify liver tumors based on enhanced MR images, unenhanced MR images, and clinical data including text and laboratory test results. Methods:Using data from 1,210 patients with liver tumors (N= 31,608 images), we trained CNNs to get seven-way classifiers, binary classifiers, and three-way malignancy-classifiers (Model A-Model G). Models were validated in an external independent extended cohort of 201 patients (N= 6,816 images). The area under receiver operating characteristic (ROC) curve (AUC) were compared across different models. We also compared the sensitivity and specificity of models with the performance of three experienced radiologists. Results:Deep learning achieves a performance on par with three experienced radiologists on classifying liver tumors in seven categories. Using only unenhanced images, CNN performs well in distinguishing malignant from benign liver tumors (AUC, 0.946; 95% CI 0.914-0.979 vs. 0.951; 0.919-0.982,P= 0.664). New CNN combining unenhanced images with clinical data greatly improved the performance of classifying malignancies as hepatocellular carcinoma (AUC, 0.985; 95% CI 0.960-1.000), metastatic tumors (0.998; 0.989-1.000), and other primary malignancies (0.963; 0.896-1.000), and the agreement with pathology was 91.9%.These models mined diagnostic information in unenhanced images and clinical data by deep-neural-network, which were different to previous methods that utilized enhanced images. The sensitivity and specificity of almost every category in these models reached the same high level compared to three experienced radiologists. Conclusion:Trained with data in various acquisition conditions, DLS that integrated these models could be used as an accurate and time-saving assisted-diagnostic strategy for liver tumors in clinical settings, even in the absence of contrast agents. DLS therefore has the potential to avoid contrast-related side effects and reduce economic costs associated with current standard MRI inspection practices for liver tumor patients.

Detection of fibrotic changes in the progression of liver diseases by label‐free multiphoton imaging

Label-free characterization of pathological changes in the portal area of liver fibrosis tissue using multiphoton imaging and quantitative image analysis

Quantitative assessment of hepatic steatosis using label-free multiphoton imaging and customized image processing program

Optical Tissue Diagnostics Based on Label-free, Functional Biomarkers

Recent advances in molecular magnetic resonance imaging of liver fibrosis.

Multiplex Collagen Fingerprinting for the Staging of Hepatic Fibrosis Using High-Precision Fluorescence-Guided SERS Imaging

Nonlinear Optical Microscopy: Use of Second Harmonic Generation and Two-Photon Microscopy for Automated Quantitative Liver Fibrosis Studies

Label-free Imaging of Bile Duct Tissues with Multiphoton Microscopy

The imaging study and progress of liver fibrosis should be emphasized

Serological Diagnosis Of Liver Fibrosis By Quantitative Estimation Of Serum Biomarkers Using Luminex Xmap Assay

Non-linear Optical Microscopy as a Sensitive and Quantitative Technique for Liver Fibrosis Studies

Liver Tumor Detection Via A Multi-Scale Intermediate Multi-Modal Fusion Network on MRI Images.

Deep Learning for Accurate Diagnosis of Liver Tumor Based on Magnetic Resonance Imaging and Clinical Data

Multimodal nonlinear optical imaging of obesity-induced liver steatosis and fibrosis

Dynamic Monitoring of the Fibrosis Disease by a Collagen Targeting Near Infrared Probe.

Dynamic monitoring the fibrosis disease by collagen targeting near infrared probe

Rapid, label-free histopathological diagnosis of liver cancer based on Raman spectroscopy and deep learning

An Automatic Liver Fibrosis Qualitative Analysis Method Based on Hyperspectral Images

Label-free Monitoring of Colonic Cancer Progression Using Multiphoton Microscopy

Label‐Free Detection of Breast Phyllodes Tumors Based on Multiphoton Microscopy

Mapping Organizational Changes of Fiber‐Like Structures in Disease Progression by Multiparametric, Quantitative Imaging