Abstract:Background: Patients with liver cirrhosis are at higher risk of developing hepatocellular carcinoma (HCC). Due to the low accuracy of current methods, new non-invasive strategies for early HCC diagnosis in cirrhotic patients are needed. Methods: A case-control study recruited a total of 2,250 patients with liver cirrhosis (LC), 508 with HCC, and 476 healthy controls (CTRL), from 13 hospitals in 11 provinces of China. Four genomic features of cell-free DNA (cfDNA), including nucleosome footprint, fragmentation, base mismatch and 5-hydroxymethylcytosines were profiled. Training and validation cohorts helped develop the integrated diagnostic model, and a test cohort was used to evaluate its performance. Findings: We established a classifier of each feature for HCC against non-HCC (cirrhosis & healthy control). By integrating these four genomic features, we developed a novel diagnostic model (HIFI, 5-Hydroxymethylcytosines/base mIsmatch/Fragmentation/nucleosome footprInt), and achieved a higher accuracy in differentiating HCCs from liver cirrhosis (AUC=0.997 [0.994-0.999], sensitivity 95.42%, specificity 97.67% in test set) than AFP or PIVKA-II, irrespective of the following demographics and clinical features, age, HBV status, Child-Pugh score, BCLC stage, tumor size, and AFP status. In AFP positive (AFP>20ug/L) liver cirrhosis patients and AFP/PIVKA-II negative (AFP 400μg/L: 100% in liver cirrhosis patients; AFP<20μg/L: 93.9%, PIVKA-II<40mAU/ml: 87.9% in HCC patients). For early-stage HCC (BCLC-0/A, AJCC-I and tumor size<3cm), the accuracies of HIFI method (88.9%/94.4%, 93.3% and 96.7%) significantly outperformed AFP/PIVKA-II (AFP:22.2%/16.7%, 41.3% and 36.7%; PIVKA-II:55.6%/61.1%, 78.7% and 70.0%). Interpretation: HIFI method exhibited an excellent performance for early-stage HCC diagnosis compared with other non-invasive methods, holding great potential of HCC surveillance for cirrhotic patients. Funding Statement: This work was supported by grants from the state Key project for infectious diseases (2018ZX10732202-001), National Research Program of China (2017YFA0505803, 2017YFC0908100), National Natural Science Foundation of China (81790633, 81672860, 81702298, 61922047, 81830054, 91859205 and 81422032), National Natural Science Foundation of Shanghai (17ZR143800). Declaration of Interests: None. Ethics Approval Statement: The study protocol was reviewed and approved by the institutional review board at all participating hospitals. All participants provided written informed consent.

Feasibility of automatic detection of small hepatocellular carcinoma (≤2 cm) in cirrhotic liver based on pattern matching and deep learning

Deep Learning Radiomics for Focal Liver Lesions Diagnosis on Long-Range Contrast-Enhanced Ultrasound and Clinical Factors

Automatic segmentation of hepatocellular carcinoma on dynamic contrast-enhanced MRI based on deep learning

Applicability of Multidimensional Convolutional Neural Networks on Automated Detection of Diverse Focal Liver Lesions in Multiphase CT Images.

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

Deep Learning for Approaching Hepatocellular Carcinoma Ultrasound Screening Dilemma: Identification of -Fetoprotein-Negative Hepatocellular Carcinoma From Focal Liver Lesion Found in High-Risk Patients

Deep Learning for Approaching Hepatocellular Carcinoma Ultrasound Screening Dilemma: Identification of α-Fetoprotein-Negative Hepatocellular Carcinoma From Focal Liver Lesion Found in High-Risk Patients

A Flexible Three-Dimensional Hetero-phase Computed Tomography Hepatocellular Carcinoma (HCC) Detection Algorithm for Generalizable and Practical HCC Screening

Early warning of hepatocellular carcinoma in cirrhotic patients by three-phase CT-based deep learning radiomics model: a retrospective, multicentre, cohort study

Automatic Detection and Classification of Focal Liver Lesions Based on Deep Convolutional Neural Networks: A Preliminary Study

Automatic Volumetric Diagnosis of Hepatocellular Carcinoma Based on Four-Phase CT Scans with Minimum Extra Information.

Deep Learning Combined with Radiologist's Intervention Achieves Accurate Segmentation of Hepatocellular Carcinoma in Dual-Phase Magnetic Resonance Images

Development of a Deep Learning Model to Assist with Diagnosis of Hepatocellular Carcinoma

Development of an AI system for accurately diagnose hepatocellular carcinoma from computed tomography imaging data

Delving Deep into Liver Focal Lesion Detection: A Preliminary Study

Genome-Scale Profiling of Circulating Cell-Free DNA Signatures for Early Detection of Hepatocellular Carcinoma in Patients with Cirrhosis

Development of machine learning-based predictors for early diagnosis of hepatocellular carcinoma

Multi-phase contrast-enhanced magnetic resonance image-based radiomics-combined machine learning reveals microscopic ultra-early hepatocellular carcinoma lesions

Automated machine learning for differentiation of hepatocellular carcinoma from intrahepatic cholangiocarcinoma on multiphasic MRI

Deep learning‐based accurate diagnosis and quantitative evaluation of microvascular invasion in hepatocellular carcinoma on whole‐slide histopathology images

Automated Detection of Primary Liver Cancer Using Different Deep Learning Approaches