Abstract
Background
We aimed to formulate a novel predictive nomogram to discriminate liver fibrosis stage in patients with chronic liver disease.
Methods
Nomograms were established based on the results of multivariate analysis. The predictive accuracy of the nomograms was assessed by ROC analysis and calibration. Decision curve analysis (DCA) was used to determine the clinical benefit of the nomograms.
Results
INR, platelets, and N-terminal propeptide type III collagen (PIIINP) were independent predictors for advanced liver fibrosis (≥ S3) and cirrhosis (S4) in patients with chronic liver disease in the training cohort. In the training set, the areas under the ROCs (AUROCs) of nomogram S3S4, APRI, FIB-4, and GPR for stage ≥ S3 were 0.83, 0.71, 0.68, and 0.74, respectively; the AUROCs of nomogram S4, APRI, FIB-4, and GPR for stage S4 were 0.88, 0.74, 0.78, and 0.79, respectively. The calibrations showed optimal agreement between the prediction by the established nomograms and actual observation. In the validation set, the AUROCs of nomogram S3S4, APRI, FIB-4, and GPR for stage ≥ S3 were 0.86, 0.79, 0.78, and 0.81, respectively; the AUROCs of nomogram S4, APRI, FIB-4, and GPR for stage S4 were 0.88, 0.77, 0.81, and 0.83, respectively. Furthermore, the decision curve analysis suggested that the nomograms represent better clinical benefits in both independent cohorts than APRI, FIB-4, and GPR.
Conclusion
The constructed nomograms could be a superior tool for discriminating advanced fibrosis and cirrhosis in chronic liver disease.
Application of non-invasive tools to decide the need for beta-blockers for variceal bleeding prophylaxis in compensated advanced liver disease: A decision curve analysis
