Search

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

Background and aim Vibration Controlled Transient Elastography (VCTE) is a non-invasive test for liver fibrosis and cirrhosis but may be inaccurate in some patients, especially in those with chronic hepatitis B. This study aims at improving the accuracy of VCTE in cirrhosis detection by combining ultrasound and routine blood parameters. Methods Hepatitis B patients with liver biopsies samples ≥20 mm underwent VCTE, ultrasound and blood tests, and were divided into training set ( n = 170) and validation set ( n = 75). Results An algorithm consisting of VCTE, international normalization ratio (INR), ultrasonic hepatic vessel and platelet count (CIR-4) and a VCTE-based cirrhosis six-index score (CIR-6) comprised VCTE, INR, platelet, albumin, ultrasonic hepatic vessel and liver parenchyma were derived. In training set, area under receiver operating characteristics curve of CIR-6 and CIR-4 to detect cirrhosis was 0.946 and 0.945, respectively, which was superior to that of VCTE 0.907. CIR-4 could save more liver biopsies. In validation set, CIR-6 detected cirrhosis with accuracy similar to that in training set. However, the sensitivity of CIR-4 and VCTE in validation set lowered to 0.538 and 0.846, respectively. Conclusions Combining routine markers improve the accuracy of VCTE for cirrhosis detection in hepatitis B patients. CIR-6 may be more valuable.