Utility of convolutional neural network-based algorithm in medical images for liver fibrosis assessment

To the Editor:Liver fibrosis is the critical stage leading to hepatic dysfunction and might be important in the progress to portal hypertension,biliary cirrhosis,and hepatocellular carcinoma.Therefore,the accurate assessment of liver fibrosis remains a clinical concern for physicians.In recent researches,liver biopsy has been considered the golden standard for the diagnostic and assessment of liver fibrosis for many years.However,due to its invasive practice,sampling variability,and inter-and intra-assessor variability of pathological interpretations,liver biopsy is not well received in the patients and clinical physicians in many circumstances.As a result,multiple non-invasive detections of hepatic fibrosis assessment have been developed as an alternative strategy for liver biopsy as they could offer a convenient operation and acceptable diagnostic accuracy through medical imaging information.Early detection of hepatic fibrosis through clinical imaging could reduce liver failure and prevent its progression.The interpretation and analysis of medical imaging information are usually performed by clinical experts.Benefit from the rapid development of computer-aided diagnosis,especially the improved algorithm of deep learning in artificial intelligence,physicians could extract more accurate assessment for clinical decisions for diagnosis and treatment.[1]