Online combination algorithm for non-invasive assessment of chronic hepatitis B related liver fibrosis and cirrhosis in resource-limited settings

Abstract

ObjectiveThe use of commercially available noninvasive markers for chronic hepatitis B (CHB) related fibrosis is not widely available in developing countries so clinicians in those countries frequently use free alternatives. We aimed to create an optimized algorithm for selection of patients with the highest probability for presence/absence of significant liver fibrosis and cirrhosis based on the use of multiple free scores. MethodsWe evaluated six free noninvasive markers for CHB related fibrosis against liver biopsy and selected the best thresholds for prediction/exclusion of significant fibrosis and cirrhosis in CHB patients. Algorithm based on four scores and their corresponding thresholds was created. ResultsThe calculator based on developed algorithm can be found at http://www.chb-lfc.com. We evaluated 211 patients in main group and 65 patients in external validation group. We selected four scores for creation of combination algorithm. The algorithm was able to classify 123/211 (58.3%) patients with a 93.5% accuracy of correct classification for prediction of presence/absence of significant fibrosis in main group. In validation group, the algorithm was able to classify 48/65 (73.8%) of patients with 93.8% (45/48) overall accuracy. When used to predict presence/absence of cirrhosis, the algorithm was able to correctly classify 181/211 (85.8%) and 59/65 (90.8%) of patients in main and validation group, respectively, with an overall accuracy of 97.8% and 98.3%, respectively. ConclusionDeveloped algorithm based on routine laboratory tests is a usable, applicable and accurate tool for diagnosis of CHB related fibrosis and cirrhosis, suitable for resource-limited settings where more expensive modalities are unavailable.